Metabolic Channeling of Carbamoyl Phosphate, a Thermolabile Intermediate EVIDENCE FOR PHYSICAL INTERACTION BETWEEN CARBAMATE KINASE-LIKE CARBAMOYL-PHOSPHATE SYNTHETASE AND ORNITHINE CARBAMOYLTRANSFERASE FROM THE HYPERTHERMOPHILE PYROCOCCUS FURIOSUS*

Two different approaches provided evidence for a physical interaction between the carbamate kinase-like carbamoyl-phosphate synthetase (CKase) and ornithine carbamoyltransferase (OTCase) from the hyperthermophilic archaeon Pyrococcus furiosus . Affinity electrophoresis indicated that CKase and OTCase associate into a multienzyme cluster. Further evidence for a biologically significant interaction between CKase and OTCase was obtained by co-immunoprecipitation com-bined with formaldehyde cross-linking experiments. These experiments support the hypothesis that CKase and OTCase form an efficient channeling cluster for carbamoyl

form citrulline, which is converted into arginine in two steps. Aspartate carbamoyltransferase (EC 2.1.3.2) catalyzes the condensation of CP and aspartate into carbamoyl aspartate in the committed step of pyrimidine biosynthesis. The decomposition of CP in aqueous solutions at high temperatures (7) leads to the accumulation of cyanate, a powerful and indiscriminate carbamoylating agent. At 100°C, where Pyrococcus furiosus reaches its highest growth rate, the half-life of CP is less than 2 s (8). Thus, because CP is both extremely labile at high temperatures and potentially toxic, it would appear to require metabolic protection in vivo.
Several examples exist that show that enzymes belonging to the same or related biochemical pathways are organized into functional complexes. These enzyme clusters can pass on products very efficiently without accomplishing a diffusion equilibrium with the bulk phase, a phenomenon called channeling (9 -13). Relatively tight channeling could protect chemically labile intermediates and thus play a critical role in the physiology of thermophiles.
Competition experiments between labeled CP produced by carbamate kinase-like carbamoyl-phosphate synthetase (CKase; EC 2.7.2.2) and unlabeled CP added to P. furiosus cell extracts showed a marked preference of OTCase for the utilization of CP synthesized by CKase (8). These findings suggest that CKase and OTCase associate to form a multienzyme complex able to channel this labile intermediate. Similar findings indicated the existence of channeling of CP to both aspartate carbamoyltransferase and OTCase in Thermus ZO5 (14) and in the deep sea hyperthermophilic archaeon Pyrococcus abyssi (15). However, to date no evidence for physical interaction between the synthetase and the carbamoyltransferases had been obtained.
In most prokaryotes, carbamoyl-phosphate synthetase (EC 6.3.4.16) is a heterodimer. A small glutaminase subunit hydrolyzes glutamine and releases ammonia that is transferred to a large subunit, and the latter catalyzes CP synthesis in a complex reaction that also requires bicarbonate and MgATP (16,17). In the archaeon P. furiosus, however, CP is synthesized directly from ammonia (8), and the enzyme from the archaeon is enzymologically and structurally a carbamate kinase (18 -21). A similar enzymatic activity was reported in the related species P. abyssi (22,23). This situation is unprecedented in microbial CP biosynthesis.
Most OTCases are homotrimeric. The OTCase from P. furiosus, however, is built of four trimers disposed in a tetrahedral manner similar to the dodecameric structure of the allosterically regulated OTCase from Pseudomonas aeruginosa (24,25). Thermal stabilization of the P. furiosus OTCase dodecamer results mainly from hydrophobic interactions at the interfaces between the catalytic trimers (24,26).
When competition experiments (such as those referred to above) indicate shielding from the bulk aqueous phase, the interactions between enzymes should be sufficiently strong to become detectable by techniques designed to monitor the physical association of proteins. Affinity electrophoresis (27) has already been used to demonstrate clustering between consecutive enzymes in the glycolytic pathway and the citric acid cycle (28). This method allows the visualization of a direct interaction between two different proteins or enzymes in conditions that mimic the crowded intracellular environment (29), which is by now acknowledged to be quite different from in vitro conditions. Another widely used method for detecting protein interactions is co-immunoprecipitation. Because it can be performed with cell-free extracts, it permits the detection of multienzyme associations in a system where all cellular components are still present. Moreover, it can be used in combination with cross-linking reactions (30,31). Both techniques thus provide complementary information. In this work, we used both affinity electrophoresis and co-immunoprecipitation with and without a cross-linking agent to demonstrate a physical interaction between P. furiosus CKase and OTCase. We speculate that the interaction between CKase and OTCase is essential for the survival of this extreme thermophilic archaeon, because the direct channeling of CP not only allows for efficient synthesis of proteins but also protects the cell against the thermodegradation of CP into cyanate.

EXPERIMENTAL PROCEDURES
Strains and Culture Conditions-P. furiosus DSM 3638 was grown as previously described in a complex medium based on artificial seawater supplemented with 0.1% yeast extract and 0.5% peptone (8,32). Cells were grown at 95°C in a Braun Biostat U fermentor in 60-liter batch runs, continuously sparging with nitrogen gas at 400 ml min Ϫ1 to maintain anaerobic conditions; agitation was at 200 rpm. Cells were harvested by centrifugation at the end of the exponential growth phase. Saccharomyces cerevisiae strains SS1 (33) and 10W51a (34) were used to overexpress pyrococcal enzymes. The P. furiosus OTCase gene was inserted into a pYEF1 vector (35) and expressed from the galactoseinducible GAL10-CYC1 promoter in the S. cerevisiae strain SS1 (33). The DNA sequence coding for the P. furiosus CKase was inserted into a pYEF2 vector (35) where it is expressed from the galactose-inducible GAL10-CYC1 promoter in S. cerevisiae 10W51a (34). Yeast cells were grown in a 15-liter batch of galactose-containing (1%) medium 164 (36) in a Biolafitte fermentor. Cells were grown at 30°C up to a density of about 4 ϫ 10 7 cells ml Ϫ1 .
Enzyme Purification-P. furiosus CKase was purified from recombinant S. cerevisiae cells as described by Uriarte et al. (21). P. furiosus OTCase was purified from recombinant S. cerevisiae cells as described by Legrain et al. (37) using arginine-Sepharose chromatography for the last purification step. Escherichia coli OTCase was purified as described previously (38). P. furiosus glutamate dehydrogenase was purified directly from the archaeon itself (39). 10 g (wet mass) of frozen Pyrococcus cells were thawed in 10 ml of 100 mM Tris-HCl, pH 8.0, supplemented with 2.5 mg of DNase. Cells were disrupted by ultrasonic oscillation (10 min, 250 watts, 10 kHz). The suspension was then centrifuged for 60 min at 45,000 rpm. The supernatant was applied on a Mono Q HR 10/10 column (Amersham Biosciences FPLC), equilibrated with 50 mM Tris-HCl, pH 8.0. Proteins were eluted with a linear gradient of 0 -0.5 M NaCl in 50 mM Tris-HCl buffer, pH 8.0. Active fractions were pooled and dialyzed against 100 mM NaHCO 3 . No CKase or OTCase activity could be detected in the pooled fractions. All other proteins were from Sigma. It was verified that none of the used protein solutions showed CKase or OTCase activity.
Polyacrylamide Gel Electrophoresis-Electrophoresis under native and denaturing conditions (SDS, ␤-mercaptoethanol) was performed on 8 -25% gradient gels with the Phast System (Amersham Biosciences). Analytical isoelectric focusing (IEF) was performed with IEF pH 3.0 -9.0 gels at 15°C. Protein bands were visualized by staining with Coomassie Brilliant Blue.
Enzyme Activity-CKase activity was determined using a coupled assay with partially purified E. coli OTCase at 60°C (40). E. coli OTCase is stable at this temperature (41). The reaction mixture contained in a final volume of 2 ml: 100 mM Tris-HCl, pH 7.2, 200 mM NH 4 Cl, 3 mM MgCl 2 , 3 mM ATP⅐Na 2 , 6 mM L-ornithine, 40 mM NaHCO 3 , and 200 units of E. coli OTCase. After a 20-min incubation the reaction was stopped, and the citrulline product was determined by the method of Prescott and Jones (42). OTCase activity was determined by measuring the formation of citrulline at 60°C (37). The reaction mixture contained 200 mM Tris-HCl, pH 7.2, 2 mM L-ornithine, and 10 mM carbamoyl phosphate. After 5 min the reaction was stopped, and the formed citrulline was determined. Glutamate dehydrogenase activity was measured at 40°C by following the appearance of NADPH at 340 nm (39). The reaction mixture contained in a final volume of 1.0 ml of 100 mM phosphate buffer, pH 7.5, 2 mM NADP, and 10 mM L-glutamate ( ϭ 6.22 mol/OD⅐min). One enzyme unit is the amount of enzyme that converts 1 mol of substrate to product per h.
Protein Concentration-Protein concentrations were measured by the method of Lowry et al. (44) or by UV absorption compared with a bovine serum albumin standard.
Rabbit Antibodies against CKase and OTCase-Antibodies against pure CKase and OTCase were obtained from the Laboratory of Hormonology, Centre d'Economie Rurale in Marloie, Belgium. The antibodies, raised in rabbits, were tested by the Ouchterlony double immunodiffusion technique and by dot-blot (Bio-Rad). It was shown that the raised antibodies were specific for their respective antigens. The rabbit serum was cleared with bruto yeast extract.
Affinity Electrophoresis-All proteins were covalently immobilized on Sepharose 4B (Amersham Biosciences) activated with CNBr (0.4g of CNBr/g of Sepharose 4B) as described previously (29,45). Proteins to be coupled were diluted to 2 mg/ml in 100 mM NaHCO 3 buffer, pH 8.2, and 1 g of Sepharose per 10 mg of protein was used. After coupling, the gels carrying the proteins were washed with 100 mM NaHCO 3 and stored at 4°C. Affinity electrophoresis was performed in 92 mM Veronal buffer, pH 8.5 (or 10 mM Tris-HCl, pH 8.5), as described previously (27,28). Strips at the bottom of the 1% agarose layer in Veronal buffer, pH 8.5, were replaced by a mixture of 1% agarose (0.2 ml/cm 2 ) with protein coupled to Sepharose. The amount of Sepharose with immobilized protein corresponded to 0.45 mg of protein/cm 2 . Under the strips with immobilized protein, enzyme at a concentration of 6 mg/ml was loaded (5 l/hole). Electrophoresis was performed at 120 V during about 20 h for mobile CKase or OTCase, using a LBK 2117 Multiphor II flatbed electrophoresis unit equipped with a cooling plate that was maintained at 22°C. The mobile enzyme was stained directly in the gel after electrophoresis. For OTCase a histochemical and an immunochemical staining procedure was used; for CKase only the immunochemical staining could be used. For histochemical staining, gels were soaked in an OTCase reaction mixture (200 mM Tris-HCl, pH 8.0, 10 mM CP, and 4 mM L-ornithine) at 37°C over 30 min. The gel was thoroughly washed with distilled water. Formed orthophosphate was finally precipitated with 3 mM Pb(NO 3 ) 2 dissolved in 50 mM D,L-histidine at pH 7.0. For immunochemical staining, agarose gels were soaked in 10 mM phosphate-buffered saline (PBS), pH 7.3, containing the respective antiserum in 1:500 dilution. After a 3-h incubation at 25°C, the gels were washed five times for 15 min each with PBST (PBS with 0.05% Tween 20) and once for 15 min with PBS. The gel was further incubated for 1 h in a PBS solution containing 0.5 mg/ml protein A-alkaline phosphatase conjugate (Sigma). The gels were again washed (five times for 15 min each with PBST and once for 15 min with PBS). A colored precipitate was formed on the addition of BCIP/NBT (165 g/ml 5-bromo-4-chloro-3-indolyl phosphate⅐Na 2 toluidine and 330 g/ml nitro blue tetrazolium in 100 mM Tris-HCl, pH 9.5, 50 mM MgCl 2 , and 100 mM NaCl).
Co-immunoprecipitation Experiments-Immunoprecipitations were performed from unlabeled sources (46). P. furiosus extracts were prepared by simply suspending P. furiosus cells (freshly grown or from stock at Ϫ80°C) in 10 mM phosphate, pH 7.2, or 50 mM Tris buffer, pH 7.2, with DNase and incubating them for 30 min at room temperature. After centrifugation (12,000 ϫ g for 15 min) the supernatant was used for immunoprecipitation experiments. 500 l of cell-free extracts of P. furiosus (ϳ50 mg/ml total protein) were incubated with 1 l of polyclonal rabbit antibodies against either CKase or OTCase for 1 h on ice. The immune complexes were purified with protein A-Sepharose CL-4B beads (Sigma). The protein A-Sepharose CL-4B beads were suspended in lysis buffer and washed several times with lysis buffer containing 1% bovine serum albumin to eliminate aspecific binding to the Sepharose beads. 100 l of a 10% (v/v) protein A slurry (in 1% bovine serum albumin/lysis buffer) were added and incubated for 1 h. The complexes were precipitated by centrifugation for 15-30 s at 12,000 ϫ g in a standard table top centrifuge. The immunoprecipitate was washed three times with lysis buffer (10 mM phosphate or 50 mM Tris-HCl, pH 7.2) to remove unbound proteins and then dissolved in 50 l of Laemmli sample buffer.
Immunoblotting-Immunoprecipitated complexes were analyzed by SDS-PAGE followed by immunoblotting. Samples were heated at 100°C for 20 min and separated on a 12% polyacrylamide Tris-HCl gel (Bio-Rad) in a Bio-Rad electrophoresis unit at 120 V for about 1 h and 15 min. The separated polypeptide chains were blotted on a nitrocellulose membrane (16 -20 h, 15-20 V, at 4°C). The nitrocellulose membrane was successively washed with 1% bovine serum albumin/TBS for 1 h, washed with TBS for 10 min, incubated with a 500ϫ dilution rabbit anti-CKase or anti-OTCase in TBST, washed six times for 10 min each with TBST, incubated with a 7000ϫ dilution goat anti-rabbit alkaline phosphatase conjugate (Sigma) for 30 min in TBST, washed three times 10 min with TBST, and washed once for 15 min with TBS before color precipitation with BCIP/NBT (Sigma).
Formaldehyde Cross-linking-Cross-linking was performed on P. furiosus extracts with 1% formaldehyde in 10 mM phosphate buffer, pH 7.2, for 30 min at room temperature. After adding Tris-HCl buffer, pH 7.2, to a final concentration of 50 mM, cross-linked samples were further used in the immunoprecipitation experiments.
Binding Assay by Resonant Mirror Detection-To provide an independent and more quantitative determination of the affinity of both pyrococcal enzymes for each other, the IAsys system (Affinity Sensors, Cambridge, UK) was used (47,48). Biotinylated proteins were immobilized onto a cuvette with a biotin surface through avidin or streptavidin. Cuvettes with only avidin or streptavidin were negative controls for binding studies.

RESULTS
Affinity Electrophoresis- Fig. 1 compares the migration of free OTCase through a strip containing immobilized CKase and underivatized Sepharose 4B. After migration, histochemical staining (precipitation of orthophosphate) revealed the OTCase in the gel. Compared with the strip containing underivatized Sepharose, the immobilized CKase was effective in binding the migrating OTCase, thus indicating an interaction between these consecutive enzymes of the arginine metabolic pathway.
Similar results were obtained when OTCase was detected immunochemically after electrophoresis through a strip containing immobilized CKase (Fig. 2, lane 2). Reciprocal experiments with CKase migrating through strips containing immobilized OTCase or underivatized Sepharose 4B gave similar indications for an interaction between the two enzymes (not shown).
To test the specificity of the observed interaction between CKase and OTCase, several control proteins were used (Fig. 2). P. furiosus glutamate dehydrogenase (lane 4) does not appear to interact with OTCase migrating through the gel, and neither do aspartate aminotransferase (lane 3), immunoglobulin G (lane 5), or bovine serum albumin (lane 6). Hence, free OTCase does not show affinity for the immobilized control proteins, indicating that the interaction between CKase and OTCase is specific. However, Beeckmans et al. (28) pointed out that proteins with comparable isoelectric points should be used as controls to avoid aspecific interactions due to electrostatic behavior. Indeed, lysozyme (not shown), cytochrome c (not shown), and mitochondrial malate dehydrogenase (lane 8) do show interaction with free migrating OTCase and CKase. All these control enzymes have a high pI (Ͼ8) compared with the pI values of CKase and OTCase, which themselves are similar (5.1 and 5.3, respectively). Free migrating OTCase shows affinity for mitochondrial malate dehydrogenase (lane 8) but not to the cytoplasmic isoenzyme, which has a lower pI (lane 7). This indicates that the interactions between OTCase and enzymes with high pI, structurally totally unrelated, is probably electrostatic and aspecific. However, the interaction between P. furiosus CKase and OTCase appears specific, as these enzymes have very comparable isoelectric points and do not interact with control proteins with similar pI values.
Co-immunoprecipitation-P. furiosus cells (freshly grown or from a frozen stock) disrupt spontaneously without sonication or any other treatment when dissolved in buffer. When concentrated P. furiosus extracts were prepared in this way and kept in a buffer of low ionic strength and low salt concentration, a CKase⅐OTCase complex could be precipitated as displayed in Fig. 3, lanes 1 and 2. Lane 1 shows the co-immunoprecipitation of OTCase (lower panel, Immunoblot with anti-OTCase) when the CKase was precipitated (upper panel, Immunoblot with anti-CKase) with an anti-CKase. Likewise, the immunoprecipitation of OTCase resulted in co-immunoprecipitation of CKase (Fig. 3, lane 2). However, only with highly concentrated extracts (ϳ50 mg/ml protein) could clear co-immunoprecipitation results be obtained. When the P. furiosus cell-free extracts were prepared by sonication, co-immunoprecipitation of the CKase⅐OTCase complex was not observed; however, when these extracts prepared by sonication were incubated at 60°C for a few minutes prior to co-immunoprecipitation, the CKase⅐OTCase complex apparently reformed and could be precipitated (Fig. 3, lanes 5 and 6). The experiments included controls for the total amount of each protein. This was achieved by immunoprecipitating and immunoblotting with the cognate antibody. Comparing the amount of co-immunoprecipitated protein with the total amount of that protein gave an indication about the amount of protein in the complex. Co-immunoprecipitation experiments were tested in several buffers ranging from a 10 mM phosphate buffer, pH 7.2, to a 100 mM Tris-HCl buffer, pH 7.2, without salt or up to 150 mM NaCl. The best immunoprecipitation results were obtained using a buffer with a low ionic strength (50 mM phosphate or Tris-HCl, pH 7.2) and a low salt concentration (0 -100 mM NaCl). When the ionic strength of the buffer (100 mM Tris-HCl, pH 7.2) or the salt concentration was too high (150 mM NaCl), co-immunoprecipitation was not observed. These data suggest that the CKase⅐OTCase complex is easily disrupted in vitro.
The relatively weak interaction between CKase and OTCase could be captured when the cell-free extracts were incubated with the cross-linking agent formaldehyde prior to immunoprecipitation (Fig. 3, lanes 9 and 10). Formaldehyde covalently links enzymes that are in close association. Hence, existing CKase⅐OTCase complexes should undergo fixation and could be easily co-immunoprecipitated, whereas without cross-linking the interaction between CKase and OTCase could be partially disrupted, and the two enzymes would not co-immunoprecipitate as readily. Yeast strains expressing only one of the two pyrococcal enzymes, CKase (Fig. 3, lanes 3, 7, and 11) or OTCase (Fig. 3, lanes 4, 8, and 12), were used in parallel in the The influence of CP precursors (ATP, MgCl 2 , NaHCO 3 , NH 4 Cl) ornithine and citrulline on a complex formation was investigated by adding them in excess to the extracts. However, no influence of the substrates was observed. To enhance complex formation, we added excess free enzyme to the cell-free extracts prior to co-immunoprecipitation (Fig. 4). 0.125 mg of purified enzyme (CKase or OTCase) was added to 0.5 ml of cell-free extract. When anti-OTCase was used to precipitate OTCase, excess free CKase was added in vitro. This resulted in a better co-immunoprecipitation of CKase (Fig. 4, lane 6) compared with the co-immunoprecipitation without excess free enzyme (Fig. 4, lanes 8 and 9). Adding excess free OTCase had no effect on the co-immunoprecipitation result (Fig. 4, lane 7). Lane 2 of Fig. 4 shows a better co-immunoprecipitation of OTCase when excess pure OTCase was added to the cell-free extract compared with the co-immunoprecipitation without excess free OTCase (Fig. 4, lanes 3 and 4), but the effect was much smaller for OTCase than for CKase. As extra control, P. furiosus extract was immunoprecipitated with unrelated rabbit antibodies (Fig. 4, lane 5).
In conclusion, although the interaction between CKase and OTCase seems to be weak, clear co-immunoprecipitation of the CKase⅐OTCase complex was observed when used in combination with cross-linking or the addition of excess free enzyme. Together with the affinity electrophoresis results, evidence is provided for a biologically significant physical interaction between CKase and OTCase from P. furiosus.

DISCUSSION
The channeling of metabolic intermediates through functional complexes of enzymes has been proposed to operate in various metabolic pathways and could be a universal feature in molecu-lar physiology. Whereas strong support for channeling between enzymes of thermophiles has been obtained from kinetic studies, evidence for physical interaction between individual enzymes not associated in a stable, permanent complex has been lacking. Here, we present the first evidence for a physical interaction between two hyperthermophilic enzymes for which kinetic evidence had suggested that these enzymes channel a highly thermolabile and potentially toxic intermediate (8).
In affinity electrophoresis experiments, purified CKase or OTCase was allowed to diffuse electrophoretically through a strip containing immobilized OTCase or CKase, respectively, furiosus OTCase) precipitated with anti-OTCase. Lane 5, a P. furiosus extract, prepared by sonication and incubated for 10 min at 60°C, precipitated with anti-CKase. Lane 6, a P. furiosus extract, prepared by sonication and incubated for 10 min at 60°C, precipitated with anti-OTCase. Lane 9, a P. furiosus extract, cross-linked with 1% formaldehyde, precipitated with anti-CKase. Lane 10, a P. furiosus extract, cross-linked with 1% formaldehyde, precipitated with anti-OTCase. Immunoprecipitation was performed in 10 mM phosphate or 50 mM Tris-HCl, pH 7.2. Samples were heated at 100°C for 20 min in SDS loading buffer and analyzed on immunoblots as described under "Experimental Procedures." FIG. 4. Co-immunoprecipitation with excess free enzyme. Lane 1, a P. furiosus extract plus 0.25 mg/ml purified CKase precipitated with anti-CKase. Lane 2, a P. furiosus extract plus 0.25 mg/ml purified OTCase precipitated with anti-CKase. Lane 3, a P. furiosus extract with 10 mM L-ornithine precipitated with anti-CKase. Lane 4, a P. furiosus extract precipitated with anti-CKase. Lane 5, a P. furiosus extract precipitated with an aspecific rabbit-antibody. Lane 6, a P. furiosus extract plus 0.25 mg/ml purified CKase precipitated with anti-OTCase. Lane 7, a P. furiosus extract plus 0.25 mg/ml purified OTCase precipitated with anti-OTCase. Lane 8, a P. furiosus extract with 10 mM L-ornithine precipitated with anti-OTCase. Lane 9, a P. furiosus extract precipitated with anti-OTCase. Extracts were prepared in 50 mM phosphate buffer, pH 7.2, or 100 mM NaCl and incubated at 50°C for 20 min after adding excess free enzyme prior to immunoprecipitation. Samples were heated at 100°C for 20 min in SDS loading buffer and analyzed on immunoblots as described under "Experimental Procedures." and enzyme localization was performed by assaying enzymatic activity or by immunochemical staining. The resulting association between CKase and OTCase was shown to be specific to the extent that OTCase did not interact with other metabolically unrelated immobilized proteins with a comparable pI. Affinity electrophoresis thus provides direct support for the hypothesis that CKase and OTCase can interact and form aggregates in vivo.
Further evidence for a biologically significant interaction between CKase and OTCase was obtained by co-immunoprecipitation experiments. In these experiments, only a fraction of CKase was shown to coprecipitate with OTCase and vice versa, suggesting relatively weak, unstable interactions. We performed immunoprecipitations in buffers with varying ionic strength and with different salt concentrations. The best indications for an association between CKase and OTCase were obtained at lower ionic strength and lower salt concentrations, indicative of relatively weak interactions, at least in vitro. Similar findings were reported by Caspary et al. (49) for the interaction between two proteins involved in RNA splicing. Using the IAsys technology with immobilized CKase or OTCase, we were unable to obtain a more quantitative appreciation of the observed interaction between the two enzymes. The interaction between CKase and OTCase is probably too weak to be detected by this technique.
The yield of co-immunoprecipitate was increased when the extract was pretreated with a cross-linking agent, or when excess pure enzyme was added to the cell extracts. By using a cross-linking agent such as formaldehyde, possible enzymeenzyme interactions can be stabilized prior to immunoprecipitation, increasing the yield of the binding partners in the coimmunoprecipitate. Another approach to increasing the yield of the co-immunoprecipitate is to add excess pure enzyme in vitro, thereby driving complex formation. Adding free CKase to the extract resulted in a much better co-immunoprecipitation of this enzyme. A similar approach was used by Huang et al. (50) to identify proteins associated with the human retinoblastoma susceptibility protein.
When extracts were prepared by sonication, no co-immunoprecipitate was observed. Burns (51) showed previously that gentle sonication of yeast cells stopped both purine and pyrimidine biosynthesis during the sonication period, suggesting that sonication disrupted multienzyme complexes, resulting in the interruption of these biosynthetic processes. However, in P. furiosus extracts prepared by sonication, the CKase⅐OTCase complex is reformed after incubation at a higher temperature. Together with the cross-linking results, this suggests that a very significant fraction of the cognate enzymes associates in vivo. Channeling efficiency appears to increase with the temperature as shown by isotopic competition experiments in P. furiosus and P. abyssi (8,15). Bera et al. (52) recently reported kinetic evidence for the channeling of phosphoribosylamine, a labile intermediate, in purine biosynthesis in the extreme thermophilic bacterium Aquifex aeolicus, and they observed an increase in coupling efficiency with temperature.
The partial channeling of CP has been suggested to occur in pyrimidine biosynthetic complexes from yeast (53,54), Neurospora (55,56), and mammals (57,58) as well as between enzymes of the mammalian urea cycle (43,59). In addition, evidence for CP channeling in Thermus ZO5 (14), P. abyssi (15), and P. furiosus (8) was obtained from kinetic experiments, and we provide evidence for the formation of an enzyme complex directing the channeling of CP in P. furiosus. This suggests that CP channeling is widespread. However, the carbamoyl-phosphate synthetases from bacteria, fungi, and mammals are structurally very different from the CKase from Pyrococcus. It is interesting that during evolution, CP channeling has developed with structurally different enzymes in different organisms.
The evidence presented here for physical interaction between CKase and OTCase from P. furiosus corroborates predictions made from kinetic analysis and provides a basis for future, more quantitative experiments.