Contribution of the Bacterial Endosymbiont to the Biosynthesis of Pyrimidine Nucleotides in the Deep-sea Tube Worm Riftia pachyptila *

The deep-sea tube worm Riftia pachyptila (Vestimentifera) from hydrothermal vents lives in an intimate symbiosis with a sulfur-oxidizing bacterium. That involves specific interactions and obligatory metabolic exchanges between the two organisms. In this work, we analyzed the contribution of the two partners to the biosynthesis of pyrimidine nucleotides through both the “de novo” and “salvage” pathways. The first three enzymes of the de novo pathway, carbamyl-phosphate synthetase, aspartate transcarbamylase, and dihydroorotase, were present only in the trophosome, the symbiont-containing tissue. The study of these enzymes in terms of their catalytic and regulatory properties in both the trophosome and the isolated symbiotic bacteria provided a clear indication of the microbial origin of these enzymes. In contrast, the succeeding enzymes of this de novo pathway, dihydroorotate dehydrogenase and orotate phosphoribosyltransferase, were present in all body parts of the worm. This finding indicates that the animal is fully dependent on the symbiont for the de novo biosynthesis of pyrimidines. In addition, it suggests that the synthesis of pyrimidines in other tissues is possible from the intermediary metabolites provided by the trophosomal tissue and from nucleic acid degradation products since the enzymes of the salvage pathway appear to be present in all tissues of the worm. Analysis of these salvage pathway enzymes in the trophosome strongly suggested that these enzymes belong to the worm. In accordance with this conclusion, none of these enzyme activities was found in the isolated bacteria. The enzymes involved in the production of the precursors of carbamyl phosphate and nitrogen assimilation, glutamine synthetase and nitrate reductase, were also investigated, and it appears that these two enzymes are present in the bacteria.

The environment of deep-sea hydrothermal vents is very dynamic, with rapidly changing physical and chemical param-eters. Hot water containing high amounts of dissolved minerals merges from the vent and mixes with the ambient deep-sea water, generating an extreme environment in terms of hydrostatic pressure, high temperature, and chemical toxicity. In addition, this environment totally lacks photosynthesis for animal nutrition; and yet, in these hostile ecosystems, dense communities of microorganisms and animals are found (1,2).
Riftia pachyptila is a large tube worm found only in the close vicinity of these deep-sea hydrothermal vents in the Pacific Ocean (3). The anatomical organization of R. pachyptila is shown in Fig. 1 (4,5). The plume is the only part of the worm that comes in free contact with the venting waters; it has a large surface area that is highly vascularized and allows an efficient exchange of metabolites between the environment and the animal. The vestimentum is a muscle that the animal uses to position itself in the tube. Within the large sac formed by the body wall and terminated by the opisthosome are two of the major tissues of the worm: one, the coelomic fluid, bathes the other, the trophosome, which is the symbiont-harboring tissue (3).
R. pachyptila does not possess a digestive tract. The trophosome, which represents 15-30% of the animal mass, is densely colonized by a chemoautotrophic endosymbiotic bacterium (6 -8). This tissue is composed of numerous lobules of ϳ0.15-mm diameter that are vascularized by the circulatory system of the worm. The lobules consist of an outer single layer of cells and the bacteriocytes, the cells containing the bacterial symbiont. The bacteria are typically surrounded by a host vacuolar membrane that includes one or more of these bacteria (3,9,10). The trophosome is richly irrigated by small capillaries (2-3-m diameter), and the maximal diffusion distance from the symbionts to the blood is ϳ10 m (3,6). This organization allows facile metabolic exchanges between the animal and the symbiont (11,12). The endosymbiotic bacteria and the worm constitute a highly integrated system: the bacteria produce metabolic energy from the oxidation of hydrogen sulfide and provide organic compounds to the worm; in return, the worm provides the bacteria with CO 2 , O 2 , H 2 S, NH 3 , and minerals. Thus, this particular nutritional organization involves specific metabolic exchanges between the two organisms.
All living organisms rely on two metabolic pathways for the production of pyrimidine nucleotides. The "de novo" pathway allows the complete synthesis of these nucleotides, including the synthesis of the pyrimidine ring starting with bicarbonate, glutamine, and ATP. The "salvage" pathway ensures the production of these nucleotides from the pyrimidine nucleosides and nucleotide monophosphates provided by the intracellular degradation of nucleic acids.
Carbamyl-phosphate synthetase (CPSase), 1 aspartate transcarbamylase (ATCase), and dihydroorotase (DHOase) catalyze the first three steps of the de novo pyrimidine biosynthetic pathway. The organization of these three enzymes differs in different living organisms. In bacteria, all the pyrimidine biosynthetic enzymes are independent proteins encoded by genes generally dispersed throughout the genome, whereas in mammals, the first three steps of this pathway are catalyzed by a single multifunctional protein named CAD (13)(14)(15).
As far as the origin of the CPSase substrates in R. pachyptila is concerned, it has been postulated that carbonic anhydrase plays a significant role in CO 2 uptake by converting CO 2 to HCO 3 Ϫ (16). The L-glutamine must be synthesized by glutamine synthetase (GSase) from glutamate and ammonia. On the basis of what is known about the general processes of nitrogen assimilation, ammonia may either be obtained from the environ-ment or result from the conversion of nitrate to nitrite by nitrate reductase (NRase) (17) and its reduction to ammonia (18,19).
The subsequent reactions in the de novo pathway are catalyzed in sequence by ATCase, DHOase, dihydroorotate dehydrogenase (DHODase; a membrane-bound or mitochondrial enzyme), orotate phosphoribosyltransferase (OPRTase), and orotidylate decarboxylase, which provides UMP. This nucleotide is then further phosphorylated to UTP, which is aminated into CTP by CTP synthetase (CTPSase) (20). The salvage pathway involves a series of enzymes able to phosphorylate nucleosides and nucleoside monophosphates such as cytidine-uridine kinase and UMP and CMP kinases, enzymes catalyzing the transfer of ribose phosphate such as uracil phosphoribosyltransferase, and enzymes converting one nucleoside into another such as cytidine deaminase (CDase) (20).
Previously reported results showed that CPSase and ATCase, the two first enzymes of the de novo pathway, are present only in the trophosome (21). Here we provide more information about the distribution of the enzymes of the de novo and salvage pathways along with information about the enzymes providing the substrates for CPSase, GSase, and NRase. Source and Storage of R. pachyptila Samples-Samples of the deepsea tube worm were collected in the East Pacific volcanic range at a depth of 2600 m using the submersible "Nautile" during the campaign HOPE 99. To avoid interference with the subsequent enzyme tests, the specimens were immediately bled and dissected on board, and each isolated organ was frozen in liquid nitrogen as previously described (21).
Purification of the Bacterial Symbiont-After collecting and bleeding the animal, the bacterial symbiont was immediately purified by the method proposed by Distel and Felbeck (22) under the conditions previously described (21).
Protein Extract from Each Organ of R. pachyptila-Protein extracts from the organs were freshly prepared before the enzyme assays. Frozen tissue (ϳ2 g) was suspended in 6 ml of ice-cold extraction buffer (30 mM Tris-HCl, pH 7.8, 10 mM NaCl, 10 mM KCl, 1 mM L-dithiothreitol, 5% (v/v) glycerol, 30% (v/v) ethylene glycol, and 4 M sodium cacodylate), and the following protease inhibitors were added: 30 g/ml phenylmethylsulfonyl fluoride, 0.3 mg/ml EDTA, 0.7 g/ml pepstatin A, and 0.5 g/ml leupeptin. The mixture was homogenized in a Potter homogenizer with a Teflon pestle. The homogenate was further disrupted by sonication three times for 60 s each with a Biosonik III sonicator at 20 kilocycles/s. The homogenate was then centrifuged at 9000 ϫ g for 20 min, and the resulting supernatant was used for enzyme assays.
Carbamyl-phosphate Synthetase Assay-The activity of CPSase was determined by the radioactive method (23) under the standard conditions described previously by Simon et al. (21).
Aspartate Transcarbamylase Assay-The ATCase activity was measured by the radioactive method (24) using L-[ 14 C]aspartate. The standard conditions were 50 mM Tris-HCl, pH 8.0, 1 mM L-[ 14 C]aspartate (specific activity of 0.3 mCi/mmol), and 10 mM carbamyl phosphate. The reaction mixture was incubated at 37°C for 1 h.
Dihydroorotase Assay-The DHOase activity was measured by an assay adapted from Sander et al. (25). The protein extract (50 l) was incubated for 1 h at 37°C with 30 mM carbamyl aspartate in 0.1 M phosphate buffer, pH 6.5, at a final volume of 1 ml. After incubation, 66 l of 4 M HClO 4 was added to the tubes, which were immersed in an ice bath. The denatured protein was removed by centrifugation. The dihydroorotic acid produced during this incubation was determined after the addition of 1 ml of 1 M NaOH to 1 ml of the extract. The absorbance at 240 nm was recorded, and the data were standardized with reference to authentic dihydroorotate samples.
Dihydroorotate Dehydrogenase Assay-The DHODase activity was determined by an adaptation of the method described by Miller (26  sodium dihydroorotate, 7.5 mM KCN, 0.1% Triton X-100, 50 M dichlorophenol-indophenol, 20 M ubiquinone-30, and 50 -100 l of extract (0.5-1.5 mg of protein). The final volume was adjusted to 2 ml with distilled water. Control cuvettes were prepared with 2 ml of the same mixture without the addition of dihydroorotate.
Orotate Phosphoribosyltransferase Assay-This assay was performed in the presence of orotidylate decarboxylase by the method described by Fox (27). The reactions were carried out in a scintillation vial containing an Eppendorf tube with 1 ml of hyamine hydroxide (ICN) for the absorption of 14 CO 2 release from [carboxyl-14 C]orotate. The reaction mixture (1 ml) contained 50 mM Tris-HCl, pH 8.0, 4 mM MgCl 2 , 1 mM mercaptoethanol, 0.2 mM [carboxyl-14 C]orotic acid, 0.25 mM 5-phosphorylribose 1-pyrophosphate, and 0.4 IU of orotidylate decarboxylase (Sigma). The reaction was initiated by the addition of the protein extract and incubated at 37°C for 1 h. A control reaction was prepared in the absence of 5-phosphorylribose 1-pyrophosphate. The reaction was stopped by the addition of 0.4 ml of 0.5 M H 2 SO 4 , and the release of 14 CO 2 was measured. CTP Synthetase Assay-The CTPSase activity was determined by the assay described by Weinfeld et al. (28).
Cytidine Deaminase Assay-The CDase activity was determined by an adaptation of the method of Steuart and Burke (29). A 50-l aliquot (0.5-1.5 mg of protein) of protein extract was incubated for 1 h at 37°C in the presence of 0.1 mM Tris-HCl, pH 8.0, and 1 mM [2-14 C]cytidine (0.34 mCi/mmol) in a total volume of 300 l.
Uridine Kinase Assay-Uridine kinase catalyzes the phosphorylation of uridine and cytidine to their respective monophosphates (30,31). This activity was assayed using a protocol similar to that of Orengo and Kobayashi (30) and Ives et al. (31). A 50-l sample of extract (0.5-1.5 mg of protein) was incubated for 1 h at 37°C in the presence of 50 mM Tris-HCl, pH 8, 5 mM ATP, 5 mM GTP, 5 mM MgCl 2 , and 1 mM [2-14 C]uridine (0.2 mCi/mmol) in a total volume of 150 l.
Nitrate Reductase Assay-The NRase activity was assayed using a protocol similar to that of Hentschel et al. (33).
Glutamine Synthetase Assay-The GSase activity was measured according to the protocol of Bender et al. (34).
Quantitation of Orotic Acid and Dihydroorotic Acid in Riftia Blood-Assays for orotic acid were conducted using modifications of the method described by Rogers and Porter (35). A volume of 0.1 ml of HCl was added to 2.0 ml of worm blood. After 10 min, the coagulated proteins were removed by centrifugation. The supernatant was filtered with a Microsep filter (50 kDa; Pall Filtron), and the filtrate was used for determination of orotic acid. A volume of 0.25 ml of blood filtrate was mixed with 1.25 ml of 0.4 M sodium citrate buffer, pH 2.5, and then with 0.25 ml of saturated bromine water. After 1 min, 0.5 ml of 5% ascorbic acid was added to the solution. The control reaction involved identical aliquots and the same reagents, except that the ascorbic acid reagent was added before the saturated bromine water. After 2 min, 1.0 ml of 2.5% p-dimethylaminobenzaldehyde dissolved in 1-propanol was added, and the mixture was incubated for 10 min. The absorbance due to the formation of 5-(p-dimethylaminobenzylidine)barbituric acid was determined spectrophotometrically at 480 nm.
The assay for dihydroorotic acid was performed in a manner similar to that described by Rogers and Nicolaisen (36). High molecular mass molecules were removed from worm blood serum by filtration with a Macrosep filter (50 kDa), and the filtrate was used for determination of dihydroorotic acid.
pH Dependence Assay-The buffer system used to determine the pH dependence of the enzyme activities comprised 51 mM diethanolamine, 51 mM N-ethylmorpholine, and 100 mM MES (37). This tri-buffer system covers the entire range of pH used without significant change in ionic strength.
Protein Determination-Protein concentration was determined by the method of Lowry et al. (38) using bovine serum albumin dissolved in extraction buffer as the standard.

Distribution of Enzyme Activities of the de Novo Pyrimidine
Nucleotide Pathway in Different Parts of R. pachyptila-Since it was previously shown that CPSase and ATCase are present only in the trophosome (21), the distribution of the subsequent enzymes of the pyrimidine pathway in different parts of the worm was examined. The results obtained are presented in Table I. Interestingly, it appeared that the third enzyme of this pathway, DHOase, was also present only in the trophosome, the symbiont-harboring tissue. In contrast, the next two enzymes (DHODase and OPRTase) and the last enzyme of the pathway, CTPSase, were present in all organs of the animal. The fact that the first three enzymes were present only in the trophosome raises the question of whether these enzymes belong to the bacteria or to the worm. This point was investigated in two different ways.
Enzyme Activities in the Isolated Symbiotic Bacteria-The same enzyme determinations were made on extracts from bacteria isolated onboard the ship, immediately after collection of the animals. The results of this analysis are given in Table I. In the bacterial extract, all enzyme activities analyzed were detected, except for glutamine-dependent CPSase (CPSase(Gln)). The instability of CPSases is well known; and most probably, this enzyme was inactivated during the isolation and/or storage of the bacterial preparations. It is noteworthy that partial inactivation of this enzyme was also observed during the analysis of trophosomal extracts by DEAE-Sepharose chromatogra-TABLE I Activities of the enzymes of the de novo pyrimidine pathway in different tissues and the isolated bacterial symbiont of R. pachyptila All specific enzyme activities were measured at 37°C. Assays for CPSase, ATCase, DHODase, and OPRTase were performed in 50 mM Tris-HCl, pH 8.0. DHOase activity was determined in 100 mM phosphate buffer, pH 6.5. Assays for CTPSase were performed in 50 mM Tris-HCl, pH 8.5. In the case of the isolated bacteria, 100-l aliquots of extract were used under the same conditions as described for the different R. pachyptila tissues, except in the case of CTPSase activity, which was measured in 50 mM BisTris, pH 7.0. The numbers in parentheses indicate the number of determinations made on different Riftia individuals. ND, not detected. phy (see below). The presence of the ATCase and DHOase activities is consistent with the hypothesis supporting the bacterial origin of these enzymes in the trophosome. The OPRTase, DHODase, and CTPSase activities were also present in the isolated bacteria, with the activity of DHODase being particularly high. Thus, it appears that, in contrast to the worm, the bacteria possess all the enzymes of the de novo pyrimidine pathway.
Analyses of Trophosomal CPSase, ATCase, and DHOase Activities by Anion-exchange Chromatography-In eukaryotes, the enzymes that catalyze the first three reactions of the pyrimidine pathway (CPSase(Gln), ATCase, and DHOase) are associated in a multifunctional protein (CAD), whereas in prokaryotes, these three enzymes are separated proteins (39). To determine the origin of these enzymes in the trophosome, extracts were analyzed by DEAE-Sepharose chromatography. The results obtained (Fig. 2) demonstrate that all three activities were distinctly separated. Taken together, these results and those reported above indicate that the first three enzymes of the pathway are of bacterial origin.
Search for Dihydroorotic Acid and Orotic Acid in the Blood of the Worm-The trophosomal location of the three first enzymes and the presence of the subsequent enzymes in all tissues of the worm suggest that the intermediary metabolite dihydroorotate must be delivered to the worm tissues. The high DHODase activity found in the bacteria (Table I) suggests that orotate might also be provided. Consequently, the presence of orotic acid and dihydroorotic acid was examined in the worm blood by colorimetric assays. Orotic acid was found at concentrations of 7.7 Ϯ 1.1 and 12.9 Ϯ 1.9 M in blood samples from two different animals. Dihydroorotic acid was not detected using the dihydroorotic acid assay, whose limit of detection is ϳ5 M (36).
Assignment of the Trophosomal Enzymes on the Basis of Their Catalytic Properties-Since the trophosome can contain enzymes from both the bacteria and the worm, we attempted to distinguish between these two possibilities on the basis of some kinetic properties of these enzymes. It has been shown previously that the properties of the trophosomal ATCase (molecular mass, K m for aspartate, and sensitivity to allosteric effectors) are characteristic of bacterial enzymes (21). As far as OPRTase and CTPSase are concerned, it can be seen in Fig. 3 (A and B) that their pH activity profiles are significantly different in the trophosome and the vestimentum. The activity of OPRTase in the trophosome was virtually constant from pH 6 to 11. In contrast, OPRTase in the vestimentum showed a clear pH optimum of ϳ7.9 (Fig. 3A). The profiles of the enzymes in the trophosome and the vestimentum are clearly different in the acidic pH range. In an analogous fashion, the CTPSase pH activity profiles in the trophosome and the vestimentum are very different (Fig. 3B), showing pH optima of 7.2 and 8.7, respectively. These results suggest the existence of different enzyme forms in the trophosome and the vestimentum, suggesting that the trophosome contains a bacterial CTPSase.
On the basis of size-exclusion chromatography, DHOase appears to have a molecular mass of ϳ110 -120 kDa (data not shown), which is comparable to previously determined values in other bacteria (40,41). The carbamyl aspartate saturation curves shown in Fig. 4 provided a K m value of 2.5 Ϯ 0.2 mM for this enzyme. Similar results were obtained in the case of Pseudomonas putida DHOase (41). These results are consistent with the bacterial origin of the trophosomal DHOase. Furthermore, the pH dependence of the DHOase activity (Fig. 3C) showed a maximum at pH 6.6, a value that is higher than those found in other bacteria, plants, and mammals (40).
The effect of pH on the activity of CPSase(Gln) is shown Fig.  3D. The pH optimum determined is 9.6, which is also significantly higher than the known pH optima of homologous enzymes from Escherichia coli (42)(43)(44) and mammals (45). It is noteworthy that it was found previously that the ATCase present in the trophosome also exhibits an unusually high pH optimum (21).
Distribution of the Enzyme Activities of the Salvage Pathways in Different Parts of Riftia-The results presented above indicate that the worm is unable to synthesize the pyrimidine nucleotides through the de novo pathway. Thus, it must rely on the salvage pathways. Consequently, we investigated whether the enzymes of such pathways are present in different parts of the worm. The results of these analyses are given in Table II. CDase, uridine kinase, and uracil phosphoribosyltransferase were present in all tissues of the host. However, it is interesting to note that the CDase and uridine kinase activities were much higher in the body wall than in other tissues. Unexpectedly, the isolated bacteria did not exhibit activity for any enzyme of the salvage pathways studied. Complementary analyses were performed to obtain information about the origin of the enzymes of the salvage pathways in the trophosome. For this purpose, extracts from trophosomal and host tissues were analyzed by anion-exchange chromatography, and the pH activity profiles of these enzyme were compared. Fig. 5 shows that the extracts from the trophosome and the vestimentum exhibited a peak of CDase at the same position, suggesting a common host origin for this enzyme. Due to the variable degree of oligomerization of OPRTase and uracil phosphoribosyltransferase in most organisms, an analogous analysis could not be performed in this case. However, the pH activity profiles of CDase, uridine kinase, and uracil phosphoribosyltransferase were determined in both the trophosomal and host tissues, and the results obtained are shown in Fig. 6. For each enzyme, these profiles were identical. Taken together, these finding indicate that the enzymes of salvage pathways present in the trophosome belong to the host.

Enzyme Activities Involved in Nitrogen Assimilation and Synthesis of the Precursors of Carbamyl
Phosphate-The first reaction of the de novo pyrimidine pathway is catalyzed by CPSase using the substrate glutamine, which derives from the process of nitrogen assimilation. To obtain information about the origin of this substrate in R. pachyptila, NRase and GSase activities were measured in different tissues of the worm. Table  III shows the results obtained. NRase activity was detected only in the trophosome, as expected for a bacterial enzyme (46). GSase activity was detected in all tissues tested. The highest activity was found in the plume at the two pH values at which this activity was measured (see below).
To obtain some information about the origin of the GSase activity present in the trophosome, further analysis of this activity was carried out using anion-exchange chromatography, pH dependence of activity, and thermostability. Two peaks of GSase activity were observed when the trophosomal extract was chromatographed on DEAE-Sepharose, one with high and one with low activity (Fig. 7A). The GSase of the vestimentum extract eluted as a single peak that corresponded to the low peak from the symbiont-containing trophosome (Fig.  7B). The major peak of GSase activity present in the trophosome and that in the vestimentum extract were used for the determination of the dependence of their activities on pH. The vestimentum pooled fractions containing GSase activity had a pH optimum of 7.7 (Fig. 8A). This pH optimum and pH activity profile corresponds to that of the vestimentum crude extract (Fig. 8C, not heated). The isolated major peak of GSase in the trophosomal extract has a different pH profile with an optimum of 9.1 (Fig. 8B). The profile of the trophosomal crude extract (Fig. 8D, not heated) appears to be a combination of the pH activity profiles of the main peak of the trophosomal extract and the vestimentum (Fig. 8, A and B).   To further analyze this trophosomal GSase content, the thermostability of these enzymes was investigated. It has been reported that the GSases from prokaryotes are more thermoresistant than those from eukaryotes (47). Consequently, heating is expected to cause inactivation of the GSase of the host, but not of the endosymbiont. To determine the thermal stability of the host and symbiont GSases, the trophosomal and vestimentum protein extracts were heated at 60°C for 20 min. The   FIG. 8. pH dependence of GSase activity. The GSase activity was assayed in pooled fractions obtained after ion-exchange chromatography (see Fig. 7) as well as in crude extracts prior to and after heating for 20 min at 60°C as described under "Experimental Procedures." A, pooled fractions obtained after ion-exchange chromatography of the vestimentum extract; B, pooled fractions obtained after ion-exchange chromatography of the trophosomal extract; C, vestimentum extract before and after heat treatment at 60°C for 20 min; D, trophosomal extract before and after heat treatment at 60°C for 20 min. The GSase activity measurements were made using 10-l aliquots of crude extracts and 100-l aliquots of chromatography fractions. enzyme activities measured after the heat treatment of the trophosomal and vestimentum protein extracts and their pH dependences are presented in Fig. 8 (C and D). It can be seen that heating at 60°C did not inactivate fractions associated with the higher peak from the trophosomal extract (pH optimum of 9.1), but it strongly inactivated both the vestimentum activity and the trophosomal GSase activity centered at pH 7.7. Taken together, these findings indicate that two GSase activities are present in the trophosome, one from the symbiont and one from the worm. Table III shows that, in the crude extracts, the specific enzyme activity in the vestimentum was higher than that in the trophosome. However, the measurements made after separation by chromatography show a significant decrease in the vestimentum activity and in the corresponding fraction of the trophosomal extract. These results indicate that the GSase of the host is an unstable enzyme and explain the low activity of the smaller peak obtained from the trophosomal extract (Fig. 7). DISCUSSION In this study, we have investigated in R. pachyptila the specific metabolic organization and interdependence between the host and the bacterial symbiont for the biosynthesis of pyrimidines. Our analysis characterized the enzymes involved and their distribution in different tissues of the worm and in the bacteria. Fig. 9 shows the model that summarizes the results of this investigation and that describes the interconnections between the bacteria and its host for the biosynthesis of the pyrimidine nucleotides.
The first three enzymes (CPSase, ATCase, and DHOase) involved in the de novo pyrimidine biosynthetic pathway are present only in the trophosome. Our study has indicated their microbial origin, based on the distribution of enzymes in both the trophosome and the isolated symbiotic bacteria as well as their catalytic and regulatory properties. The apparent molecular mass, K m , and sensitivity to effectors of ATCase and CP-Sase in the trophosome are characteristic of bacterial rather than eukaryotic homologs (21). The complementary results reported here for DHOase (apparent molecular mass and K m ) also demonstrate the bacterial origin of this enzyme. In addition, all three enzyme activities are found associated with distinct separated proteins, a characteristic of prokaryotic enzymes. In eukaryotes, the first three steps of the de novo pyrimidine nucleotide pathway are catalyzed by a single multifunctional protein, CAD (13)(14)(15).
The subsequent enzymes of the de novo pyrimidine pathway (DHODase, OPRTase, and CTPSase) appear to be present in both the bacteria and the worm. The host, R. pachyptila, can synthesize pyrimidine from orotate or dihydroorotate provided by the bacteria.
The absence of the de novo pyrimidine biosynthesis has been previously reported for aerobic protozoan parasites such as Giardia lamblia, Trichomonas vaginalis, and Tritrichomonas fetus (48 -50). These parasites appear to rely solely on salvage pathways to obtain their pyrimidine requirements. The results reported in this work show that all tissues of R. pachyptila possess the enzyme equipment of salvage pathways for the production of their pyrimidines.
Alternatively, the source of nucleotides or nucleosides for the host might result from release of these products by digestion of the symbiont cells in the trophosomal tissue. Indeed, studies of the trophosome demonstrated the presence of bacterial symbionts in various steps of degradation (51). In this way, these degradation metabolites might be transported and incorporated by the salvage pathways into nucleotides for the host cells. In contrast to nucleotides that are unable to cross cell membranes, nucleosides (uridine, cytidine, uracil, and others) can traverse the membranes (20) and could thereby be delivered to any part of the worm. As far as the bacteria are concerned, the de novo pyrimidine biosynthetic pathway is the only route for the biosynthesis of pyrimidines.
It has not been possible thus far to grow the bacterial symbiont of R. pachyptila in culture; and therefore, a thorough investigation has not been completed. However, it appears that the kinetic and regulatory properties of CPSase, ATCase, and DHOase show strong similarity to those of the homologous enzymes from Pseudomonas species (21,46). In addition, the ribosomal RNA sequence comparison of the 5 S and 16 S ribosomes has also showed that the R. pachyptila symbiont is closely related to Pseudomonas species (52,53). It is interesting to note that another sequenced protein of the R. pachyptila symbiont, the histidine protein kinase, is extremely similar to the homologous enzyme from Pseudomonas aeruginosa (54).
The pH dependence studies reported here show that, unexpectedly, CPSase, ATCase, and DHOase as well as GSase of the bacterial symbiont show pH optima for activity that are significantly higher than those of the homologous enzymes from other bacteria. In this regard, it is noteworthy that a comparable observation was made from the genome sequence of the FIG. 9. Integrated proposed scheme of pyrimidine biosynthetic pathways in the trophosomal cells of R. pachyptila. This model is based on the distribution of the enzymes of the de novo and salvage pathways for the biosynthesis of the pyrimidine nucleotides. It describes the exchanges between the endosymbiont and the trophosomal host cell. The enzymes that were characterized in this work are boxed. Question marks indicate the pathways that are not completely elucidated. The arrows pointing out of the bacteriocyte indicate the exchanges with the other organs of the worm through the circulatory system. The arrows pointing in refer to the entry of the precursors provided by the external environment. PRPP, 5-phosphorylribose 1-pyrophosphate; UracilPRTase, uracil phosphoribosyltransferase; C-UKase, cytidine-uridine kinase. endocellular bacterial symbiont of the aphid Buchera species APS (55). It was shown that the predicted isoelectric points of the products of the open reading frames are, on average, much more basic than those of the homologous proteins from other bacteria. The average pI value of Buchera proteins is 9.6, whereas those of E. coli and Hemophilus influenzae proteins are 7.2 and 7.3, respectively. Thus, the basic character of the proteins might be related to the environment under which the bacterial endosymbionts are living.
The pyrimidine biosynthesis is initiated by the CPSase-catalyzed conversion of HCO 3 Ϫ , ATP, and L-glutamine into carbamyl phosphate. The availability of the substrates for this reaction in R. pachyptila raises several questions. Carbonic anhydrase, which facilitates the transformation of carbon dioxide into bicarbonate ions, was identified in R. pachyptila. This enzyme shows a similarity to enzymes from mammalian sources (17). L-Glutamine plays a key role as a source of nitrogen for the synthesis of nucleotides and is essential for cell proliferation (56,57). L-Glutamine is synthesized from glutamate and ammonia by GSase, which is a primary enzyme for the assimilation of ammonia. Previous analysis of GSase in R. pachyptila showed that GSase is present in the trophosome and the vestimentum (26). However, the existence in the trophosome of only the bacterial enzyme was not entirely established (26). For this reason, we carried out a complementary analysis of GSase activity in other tissues of R. pachyptila, and we now have evidence for the existence of distinct forms of GSase in the host and the symbiont. This analysis has shown that all tissues of R. pachyptila possess GSase. Furthermore, two forms of GSase were distinguished in the trophosome, one possessing host properties and one showing bacterial proprieties. Interestingly, the highest GSase activity was found in the branchial plume, strongly suggesting that GSase indeed has an important role in the fixation of ammonia from the vent environment.
Ammonia, a substrate of GSase, may be supplied to this enzyme either from the environment or from nitrate, converted to nitrite by NRase (17), and further reduced to ammonia (18,19). The NRase activity was previously detected in the trophosomal tissue (18) and in the purified symbiont from R. pachyptila (46). Our complementary results indicate that only the symbiont possesses this activity, which ensures the reduction of nitrate to nitrite in R. pachyptila.
The existence of the salvage pathway in the worm indicates the presence of the catabolic production of nucleosides. These metabolites might be used not only by the salvage pathway enzymes, but also by the enzymes involved in their catabolism. This process could represent a possible source of carbon and nitrogen for the worm. This possibility is presently being investigated in our laboratory.