Characterization of the Aspartate Transcarbamoylase fromMethanococcus jannaschii *

The genes from the thermophilic archaeabacteriumMethanococcus jannaschii that code for the putative catalytic and regulatory chains of aspartate transcarbamoylase were expressed at high levels in Escherichia coli. Only theM. jannaschii PyrB (Mj-PyrB) gene product exhibited catalytic activity. A purification protocol was devised for the Mj-PyrB and M. jannaschii PyrI (Mj-PyrI) gene products. Molecular weight measurements of the Mj-PyrB and Mj-PyrI gene products revealed that the Mj-PyrB gene product is a trimer and the Mj-PyrI gene product is a dimer. Preliminary characterization of the aspartate transcarbamoylase from M. jannaschii cell-free extract revealed that the enzyme has a similar molecular weight to that of theE. coli holoenzyme. Kinetic analysis of the M. jannaschii aspartate transcarbamoylase from the cell-free extract indicates that the enzyme exhibited limited homotropic cooperativity and little if any regulatory properties. The purified Mj-catalytic trimer exhibited hyperbolic kinetics, with an activation energy similar to that observed for the E. coli catalytic trimer. Homology models of the Mj-PyrB and Mj-PyrI gene products were constructed based on the three-dimensional structures of the homologous E. coli proteins. The residues known to be critical for catalysis, regulation, and formation of the quaternary structure from the well characterized E. coli aspartate transcarbamoylase were compared.

Organisms from the archaea, prokarya, and eukarya kingdoms all produce aspartate transcarbamoylase, the enzyme that catalyzes the committed step of the pyrimidine biosynthetic pathway, the reaction of carbamoyl phosphate and Laspartate to form N-carbamoyl-L-aspartate and inorganic phosphate (1). There are four major classes or forms of quaternary structures known for aspartate transcarbamoylases. In prokaryotes, aspartate transcarbamoylase is known to exist in three classes. The simplest is class C, a homotrimer of catalytic chains each with a molecular mass of approximately 34 kDa. The aspartate transcarbamoylase from Bacillus subtilis, which lacks both homotropic and heterotropic properties, is an example of this class (2). A second form of aspartate transcarbamoylase, class A, is a dodecamer of six 34-kDa and six 45-kDa polypeptides. Catalytic and regulatory functions of this enzyme are both located on the 34-kDa polypeptides, whereas the function of the 45-kDa polypeptides is unknown. There are several species of Pseudomonas that produce this type of aspartate transcarbamoylase, including Pseudomonas fluorescens (3,4). The third and best characterized class of aspartate transcarbamoylase is class B, comprised of two trimeric catalytic subunits of 34-kDa polypeptides and three dimeric regulatory subunits of 17-kDa polypeptides. The class B form is an allosteric enzyme, exhibiting both homotropic and heterotropic interactions. Escherichia coli, Salmonella typhimurium, Erwinia herbicola, Serratia marcescens, and other members of the family Enterobacteriaceae produce class B aspartate transcarbamoylase (5). In eukaryotes, aspartate transcarbamoylase often is part of a multienzyme complex, such as those found in yeast (6) and hamster (7).
The DNA sequence of the unicellular thermophilic archaeabacterium Methanococcus jannaschii revealed genes (pyrB and pyrI) 1 that are homologous to a class B aspartate transcarbamoylase (8). In the E. coli genome, the pyrB and pyrI genes are separated by only 12 base pairs, whereas in the M. jannaschii genome the two homologous genes are separated by over 200,000 base pairs. Fig. 1 shows a sequence alignment of the E. coli and M. jannaschii pyrB and pyrI genes. The PyrB gene product exhibits 47% identity and 67% similarity, and the PyrI gene product exhibits 35% identity and 52% similarity, suggesting that the PyrB and PyrI gene products of the two species have similar tertiary structures. The M. jannaschii differs from other recently characterized archaeabacteria such as Pyrococcus abyssi (9) and Sulfolobus acidocaldarius (10) that contain an enterobacteria-like pyrBI operon.
In order to study the M. jannaschii aspartate transcarbamoylase, the M. jannaschii pyrB (Mj-pyrB) 2 and M. jannaschii pyrI (Mj-pyrI) genes were inserted into an expression system that yielded substantial amounts of both gene products. A protein purification scheme was developed for the Mj-PyrB and Mj-PyrI gene products, and the quaternary structure of the M. jannaschii aspartate transcarbamoylase found in vivo was analyzed and kinetically characterized.

EXPERIMENTAL PROCEDURES
Materials-Agarose, ATP, CTP, L-aspartate, N-carbamoyl-L-aspartate, 2-mercaptoethanol, isopropyl-␤-thiogalactosidase, potassium di-hydrogen phosphate, sucrose, phenylmethylsulfonyl fluoride, bovine pancreatic deoxyribonuclease I, and uracil were obtained from Sigma. Q-Sepharose Fast Flow and high performance phenyl-Sepharose resin were purchased from Amersham Pharmacia Biotech. Ampicillin and the Sequenase DNA sequencing kit were obtained from United States Biochemical Corp. Restriction endonucleases and T4 DNA ligase were obtained from New England Biolabs.
Sodium dodecyl sulfate, Bio-Prep SE-1000/17, Bio-Prep SE-100/17, gel filtration molecular weight standards, and the Protein Assay Dye were purchased from Bio-Rad. Pefabloc SC, pepstatin, and leupeptin were purchased from Roche Molecular Biochemicals. Carbamoyl phosphate dilithium salt, obtained from Sigma, was purified before use by precipitation from 50% (v/v) ethanol and was stored desiccated at Ϫ20°C (11). Casamino acids, yeast extract, and tryptone were obtained from Difco. Glass beads for isolation of DNA from agarose gels were purchased from Bio 101. Ammonium sulfate, urea, Tris, and electrophoresis-grade acrylamide were purchased from ICN Biomedicals. Antipyrine was obtained from Fisher. Oligonucleotides were purchased from Operon Technologies. The pGEM T-tail vector kit was purchased from Promega.
Strains and Plasmids-The E. coli strains 627239 and 623859 harboring plasmids AMJPK84 and AMJAE67, respectively, were obtained from the American Type Culture Collection (ATCC). The plasmid pET23a was obtained from Novagen, Inc., and the plasmid pSJS1240 was provided by S. Sandler.
Construction of Plasmids for the Expression of the Mj-PyrB and Mj-PyrI Gene Products-DNA encoding the Mj-pyrB and Mj-pyrI genes were derived from plasmids AMJPK84 and AMJAE67, respectively (8). Each plasmid was incubated with EcoRI and BamHI restriction endonucleases, and the corresponding EcoRI-BamHI DNA fragments containing the Mj-pyrB and Mj-pyrI genes were separated by agarose gel electrophoresis followed by isolation from the gel with glass beads. Two sets of oligonucleotide primers were used to amplify the Mj-pyrB and Mj-pyrI gene fragments by PCR (12). One set of primers introduced a unique NdeI restriction site overlapping the 5Ј initiation codon and the other a unique SacI restriction site proximal to the 3Ј end of the termination codon. The PCR products were separated by agarose gel electrophoresis followed by isolation of the appropriate fragment from the gel with glass beads. Each fragment was separately mixed with linear p-GEM-T and treated with T4 DNA ligase for 16 h at 4°C. After selection, plasmids pGEM*pyrB and pGEM*pyrI harboring the Mj-pyrB and Mj-pyrI genes, respectively, were obtained.
In order to express the recombinant genes under the control of the E. coli aspartate transcarbamoylase pyrBI promoter (13), the Mj-pyrB and Mj-pyrI genes were transferred into plasmid pEK164 (14). In order to construct these plasmids, pGEM*pyrB, pGEM*pyrI, and pEK164 were digested with the restriction endonucleases NdeI and SacI. The DNA fragments containing the Mj-pyrB gene, Mj-pyrI gene, and pEK164 vector were isolated after agarose gel electrophoresis using glass beads. The isolated Mj-pyrB and Mj-pyrI gene fragments were individually mixed with the isolated vector fragment and treated with DNA ligase for 16 h at 4°C. The resulting plasmids, pEK400 and pEK401, were confirmed by restriction analysis. DNA sequence analysis was then used to confirm that the PCR-amplified M. jannaschii pyrB and pyrI genes had exactly the same sequence as reported (8).
The Mj-pyrB and Mj-pyrI genes were also inserted into the expression vector pET23a, so an alternate expression system, employing the T7 promoter (15), could be tested. Plasmids pEK400, pEK401, and pET23a were digested with NdeI and SacI endonucleases to remove the Mj-pyrB and Mj-pyrI genes. These DNA fragments were separated by agarose gel electrophoresis, isolated using glass beads, and then were individually mixed with NdeI-and SacI-digested pET23a. The mixtures were then treated with T4 DNA ligase for 16 h at 4°C. The resulting plasmids, pEK406 and pEK407, contained the M. jannaschii pyrB and pyrI genes respectively, under control of the T7 promoter.
Expression of the Mj-PyrB and Mj-PyrI Gene Products-Protein expression was performed with the plasmid/strain combination pEK406/ EK1594 and pEK407/EK1594 for the Mj-PyrB and Mj-PyrI gene products, respectively. EK1594 is a version of the E. coli aspartate transcarbamoylase expression strain EK1104 (16) that has an inducible gene for T7 polymerase on the chromosome along with a deletion in the pyrBI region of the chromosome (17). In addition, the plasmid pSJS1240 was co-transformed into these plasmid/strain combinations to test the importance of low copy number tRNAs for high level expression of these proteins (18).
Typically, a 50-ml overnight culture of M9 media (19) supplemented with 0.5% casamino acids, 100 g/ml ampicillin, and 100 g/ml spectinomycin (when the pSJS1240 plasmid was used) was inoculated with the appropriate plasmid/strain combination and grown overnight at 37°C. A 2% inoculum was used to seed each of the four 4-liter flasks containing 2 liters of M9 media supplemented as indicated above (8 liters total). The cells were grown at 37°C, 200 rpm to an A 560 of 0.6 (normally 2.5 h after inoculation). Protein expression was then induced by the addition of solid isopropyl-␤-thiogalactosidase to a final concentration of 0.4 mM. The cell culture was grown for 4 additional h before harvesting.
Purification of the M. jannaschii PyrB Gene Product -The cell culture was centrifuged at 4500 ϫ g for 20 min, and the cell pellet was resuspended in 64 ml of ice-cold 0.1 M Tris-Cl buffer, pH 9.2. After sonication to lyse the cells, the mixture was centrifuged at 31,000 ϫ g for 20 min. In order to increase protein yield, the cell pellet was resuspended in 32 ml of ice-cold 0.1 M Tris-Cl buffer, pH 9.2, and again centrifuged as described above.
The cell-free supernatant and the supernatant from the pellet wash were combined and brought to 30% saturation of ammonium sulfate. Aliquots of 5 ml were transferred into 18 ϫ 150-mm test tubes and placed in a metal rack. This rack was then immersed in a 90°C water bath and heated for 15 min after the mixture reached 90°C (about 3 min). During the incubation at 90°C, the tubes were gently shaken at 5-min intervals. This procedure was used to ensure that each aliquot The PyrB gene products exhibit 47% identity and 67% similarity, whereas the PyrI gene products exhibit 35% identity and 52% similarity. reached 90°C rapidly. After heating, the aliquots were recombined and centrifuged at 31,000 ϫ g for 15 min. The supernatant was retained and dialyzed 2 times in 4 liters of 0.05 M Tris acetate buffer, 2 mM 2-mercaptoethanol, pH 8.3.
The dialyzed supernatant was loaded onto a Q-Sepharose Fast Flow anion exchange column (1.8 ϫ 18 cm) pre-equilibrated with 300 ml of column buffer, 0.05 M Tris acetate, 2 mM 2-mercaptoethanol, pH 8.3. The protein was eluted from the column using 300 ml of column buffer followed by a 600-ml linear gradient of column buffer to column buffer containing 0.5 M NaCl. The column fractions containing the Mj-PyrB gene product, as determined by A 280 measurements and by SDS-PAGE, were pooled and dialyzed into a buffer solution of 40 mM KH 2 PO 4 , 2 mM 2-mercaptoethanol, and 1.3 M ammonium sulfate, pH 7.0.
The final purification step involved the use of hydrophobic interaction chromatography employing a phenyl-Sepharose column (1 ϫ 28 cm) that was pre-equilibrated with a buffer solution of 40 mM KH 2 PO 4 , 2 mM 2-mercaptoethanol, and 1.3 M ammonium sulfate, pH 7.0. The protein was loaded onto the column and washed with the same buffer solution to remove any unbound proteins. The protein was then eluted using a 110-ml linear gradient of 40 mM KH 2 PO 4 , 2 mM 2-mercaptoethanol, 1.3 M ammonium sulfate, pH 7.0, to 40 mM KH 2 PO 4 , 2 mM 2-mercaptoethanol, pH 7.0. Fractions containing greater than 98% pure protein, as determined by SDS-PAGE, were pooled and concentrated.
Purification of the M. jannaschii PyrI Gene Product -A 2-liter cell culture was centrifuged at 4,500 ϫ g for 20 min, and the pellet was resuspended in 16 ml of ice-cold 0.1 M Tris-Cl, 0.1 mM zinc acetate buffer, pH 9.2. After sonication to lyse the cells, the mixture was centrifuged at 31,000 ϫ g for 20 min. The supernatant was brought to 65% saturation of ammonium sulfate, after which it was centrifuged at 31,000 ϫ g for 15 min. The pellet was dissolved and dialyzed 2 times in 4 liters of 40 mM KH 2 PO 4 , 2-mercaptoethanol, 0.1 mM zinc acetate buffer, pH 7.0. The dialyzed supernatant was again brought to 65% saturation of ammonium sulfate and centrifuged at 31,000 ϫ g for 15 min. The pellet was dissolved and dialyzed 2 times in 4 liters of 0.05 M Tris acetate, 0.1 mM zinc acetate, 2-mercaptoethanol, pH 8.3. The dialyzed supernatant was then loaded onto a Source-Q anion exchange column (1 ϫ 6.5 cm) that was pre-equilibrated with a buffer solution of 0.05 M Tris acetate, 0.1 mM zinc acetate, 2-mercaptoethanol, pH 8.3. The column was then eluted with the same buffer solution to remove any unbound proteins. The Mj-PyrI gene product was then eluted using a 32.5-ml linear gradient of 0.05 M Tris acetate, 0.1 mM zinc acetate, 2-mercaptoethanol, pH 8.3, to 0.5 M NaCl, 0.05 M Tris acetate, 0.1 mM zinc acetate, 2-mercaptoethanol, pH 8.3. Fractions containing greater than 98% pure protein, as determined by SDS-PAGE, were pooled.
Molecular Weight Determination of the Mj-pyrB and Mj-PyrI Gene Products-The molecular weights of the Mj-PyrB and Mj-PyrI gene products were determined by gel filtration using a set of molecular weight standards. A Bio-Prep SE-1000/17 column (0.8 ϫ 30 cm) was used for the Mj-PyrB gene product and a Bio-Prep SE-100/17 column (0.8 ϫ 30 cm) was used for the Mj-PyrI gene product. The SE-1000/17 column was equilibrated with 50 mM Tris acetate buffer, 2 mM 2-mercaptoethanol, pH 8.3 buffer, and the same buffer plus 0.1 mM zinc acetate was used for the Mj-PyrI gene product. Gel filtration molecular weight standards were injected, and the proteins were eluted using the appropriate buffer. Once the column was re-equilibrated with the appropriate column buffer, either 200 l of a 8 mg/ml solution containing pure Mj-PyrB gene product or 200 l of a 4 mg/ml solution containing partially purified Mj-PyrI gene product was injected onto the column and eluted with the appropriate column buffer at a flow rate of 0.5 ml/min. Elution of the proteins was monitored by A 280 , by SDS-PAGE (20), and by nondenaturing PAGE (21,22).
Molecular Weight Determination of the M. jannaschii Aspartate Transcarbamoylase by Sucrose Density Gradient Sedimentation-The molecular weight of M. jannaschii aspartate transcarbamoylase was determined using sucrose density gradient sedimentation of a M. jannaschii cell-free extract. Approximately 1-2 g of M. jannaschii cells, kindly provided by Prof. Mary Roberts, Boston College, were resuspended in 2 ml of 0.1 M Tris-Cl buffer, pH 8.0, 1.5 mM EDTA containing 50 M phenylmethylsulfonyl fluoride, 0.1 mM Pefabloc SC, 1 g/ml pepstatin, and 1 g/ml leupeptin and then sonicated to lyse the cells. The resulting highly viscous solution was treated by slowly adding bovine pancreatic deoxyribonuclease I (10 g/ml in 10 mM MgCl 2 ) while stirring for 20 min at 4°C. After centrifugation at 22,000 ϫ g for 40 min, the supernatant was concentrated in an Amicon Centricon YM-10. This crude cell extract was layered on top of a 4-ml 6 -30% preformed sucrose gradient in 0.05 M Tris acetate buffer, pH 8.3. A standard was also prepared by layering 10 l of 1 mg/ml purified E. coli catalytic subunit and 10 l of 1 mg/ml purified E. coli holoenzyme on top of another identical sucrose gradient. By using a Beckman SW 55 Ti rotor, these two tubes were spun at 80,000 ϫ g for 14 h in a Beckman L-70 ultracentrifuge. Immediately after centrifugation, the gradients were fractionated using a Brandel model BR-184-1 fractionator. Fractions of 100 l were collected and were assayed for aspartate transcarbamoylase activity using the colorimetric assay at 37°C.
Thermal Stability of the M. jannaschii Aspartate Transcarbamoylase-The thermal stability of the M. jannaschii and E. coli PyrB gene products were compared. Samples of the M. jannaschii and E. coli PyrB gene products at 0.5 mg/ml were heated at 75°C as a function of time. At each time point an aliquot was rapidly chilled and stored at 4°C followed by determination of enzymatic activity.
Determination of Protein Concentration-The concentrations of the E. coli wild-type holoenzyme, catalytic subunit, and regulatory subunit were determined from absorbance measurements at 280 nm using extinction coefficients of 0.59, 0.72, and 0.32 cm 2 mg Ϫ1 , respectively (23). The concentrations of the Mj-PyrB and Mj-PyrI gene products were determined by the Bio-Rad version of the Bradford dye-binding assay (24).
Determination of Aspartate Transcarbamoylase Activity-Aspartate transcarbamoylase activity was measured by the colorimetric method (25) or by the pH-stat method (26) using a Radiometer-Copenhagen PHM290 titrator and an ABU901 autoburette. Aspartate saturation curves were performed in duplicate at a highly saturating concentration of carbamoyl phosphate, and data points shown in the figures are the average values. Colorimetric assays as a function of temperature were performed in 50 mM Tris acetate buffer, pH 8.3, taking into account the large ⌬pH/°C dependence of Tris buffer. At temperatures of 45°C and higher, assays were carried out for short times (3-5 min) due to the thermal instability of carbamoyl phosphate as well as the accelerated rate of the uncatalyzed reaction. Because of the high rate of the uncatalyzed reaction and the thermal instability of carbamoyl phosphate (t1 ⁄2 ϭ 3.5 min at 55°C), accurate kinetic data could not be obtained at temperatures greater than 55°C. Data analysis of the steady-state kinetics was carried out as described previously (27).
Homology Modeling of Mj-PyrB and Mj-PyrI Gene Products-The program Swiss-Model (28,29) was used to generate the homology models of the Mj-PyrB and Mj-PyrI gene products using the catalytic and regulatory chains of the E. coli aspartate transcarbamoylase as the initial model. The program employs ProModII (30) to generate the models and Gromos96 to perform the energy minimization.
The amino acid sequences of the Mj-PyrB and Mj-PyrI gene products were obtained from GenBank TM , accession numbers Q58976 and Q58801, respectively. The homology modeling template was the Protein Data Bank code 1D09, which is the recent 2.1-Å structure of aspartate transcarbamoylase in the presence of the bisubstrate analog N-phosphonacetyl-L-aspartate (31). The A and B chains of code 1D09 were used to model the Mj-PyrB and Mj-PyrI gene products, respectively.
The initial model created using the "First Approach" mode was examined using the Swiss-Protein Data Bank Viewer, and adjustments were made to the sequence alignment to improve the alignment of residues known to be functionally important in the corresponding E. coli protein. The model was then resubmitted to Swiss-Model using the "Optimize" mode. Final analysis of the models was performed in QUANTA (Molecular Simulations, Inc.).

Construction of Plasmids for the Expression of the Mj-pyrB
and Mj-pyrI Genes-Two separate approaches were taken in order to express the Mj-PyrB and Mj-PyrI gene products. The first was to place the M. jannaschii genes under the control of the E. coli pyrBI promoter in an approach similar to that used to express the B. subtilis aspartate transcarbamoylase in E. coli (13). The second was to place the M. jannaschii genes under the control of the bacteriophage T7 promoter using the pET plasmid system of Studier et al. (15).
The DNA encoding the Mj-pyrB and Mj-pyrI genes were isolated from plasmids AMJPK84 and AMJAE67, respectively (8). Since both of the expression systems require an NdeI site at the initiation fMet codon, the required NdeI site as well as an additional restriction site (SacI) after the 3Ј end of the coding region were introduced using PCR (12). The NdeI and SacI sites allowed easy cloning of the Mj-pyrB and Mj-pyrI genes into both the pyrBI and pET expression systems (see "Experimental Procedures" for details).

Expression of the Mj-pyrB and Mj-pyrI Genes Products
Using the E. coli pyrBI Promoter-In order to test the levels of protein expression of the Mj-PyrB and Mj-PyrI gene products under the control of the pyrBI promoter, the appropriate plasmids (pEK400 and pEK401) were transformed into the E. coli overproduction strain EK1104 (16). This system, which depends upon the derepression of the genes of the pyrimidine pathway, did not produce satisfactory levels of the Mj-PyrI and Mj-PyrB gene products.
Expression of the Mj-pyrB and Mj-pyrI Genes Using the pET System-The plasmids pEK406 and pEK407 containing the Mj-pyrB and Mj-pyrI genes inserted into pET23a were transformed into EK1594 (17). This E. coli strain has a chromosomal deletion of the pyrB and pyrI genes and a copy of the T7 DNA polymerase gene inserted into the chromosome under the control of the inducible lac promoter. As seen in Fig. 2, the expression of the Mj-PyrB gene product was poor in this system, whereas substantial amounts of the Mj-PyrI gene product were produced, amounting to approximately 15% of the total cellular protein.
Expression of Mj-PyrB and Mj-PyrI Gene Products Using the pET System in the Presence of pSJS1240 -Kim et al. (18) reported that the expression of archaeal proteins in E. coli is often poor due to differences in codon usage between archaea and E. coli. For example, the codons ATA, AGA, and AGG are found in high abundance in archaeal genes but are rare in E. coli genes. By enhancing the levels of certain low abundance tRNAs (argU, which codes for ATA, and ileX, which codes for AGA and AGG) in E. coli, the level of archaeal protein expression can be enhanced (18). In the case of the Mj-pyrB gene there are 31 of these rare codons, whereas in the Mj-pyrI gene there are only 9.
When the plasmid pSJS1240, encoding the argU and ileX tRNAs, was introduced into the pET Mj-pyrB and Mj-pyrI expression strains, there was a considerable improvement in the expression of the Mj-PyrB gene product, whereas the Mj-PyrI gene product was enhanced little if at all (see Fig. 2). This difference may be directly related to the relative numbers of rare codons found in each of the two genes. The combination of the pET expression system along with the enhancement of the rare tRNA pools thus provides a useful method for overproduc-tion of both the Mj-PyrB and Mj-PyrI gene products in E. coli.
Catalytic Activity of the Mj-PyrB and Mj-PyrI Gene Products-In order to determine whether the Mj-PyrB or Mj-PyrI gene products exhibited aspartate transcarbamoylase activity, crude cell extracts were analyzed using the highly sensitive colorimetric assay (25). Crude cell extracts containing the Mj-PyrB gene product expressed in E. coli exhibited considerable aspartate transcarbamoylase activity (data not shown). In extracts of cells without the plasmid containing the Mj-pyrB gene or with the Mj-pyrI gene, no aspartate transcarbamoylase activity could be detected (data not shown). These data indicate that the Mj-PyrB gene product is sufficient for aspartate transcarbamoylase activity, just as the corresponding PyrB gene product from E. coli is catalytically active even in the absence of the PyrI gene product.
Purification of the Mj-PyrB Gene Product-As seen in Fig. 3, lane B, the Mj-PyrB gene product is highly overexpressed in the presence of enhanced tRNA pools. A purification procedure for the Mj-PyrB gene product was developed taking advantage of the expected thermal stability of the protein. After the cells were broken open by sonication and centrifuged to remove cell debris, ammonium sulfate was added to 30% saturation followed by a 15-min, 90°C heat treatment. Negligible loss of enzymatic activity and a 2-fold purification were observed (Fig.  3, lane C). The heat step was followed by two chromatography steps (Fig. 3, lanes D and E), the first employing anion exchange chromatography (Q-Sepharose Fast Flow) and the second employing hydrophobic interaction chromatography (phenyl-Sepharose). The overall protein purification procedure resulted in a 5.9-fold purification, negligible loss of enzyme activity, and greater than 98% pure Mj-PyrB gene product (see Table I and Fig. 3, lane E).
Thermal Stability of the M. jannaschii PyrB Gene Product-Given the fact that the optimal growth temperature of M. jannaschii is 85°C as compared with 37°C for E. coli, a comparison of the thermal stabilities of the PyrB gene products was carried out. As seen in Fig. 4, the M. jannaschii PyrB gene product is substantially more stable than the E. coli catalytic subunit. At 75°C half of the initial catalytic activity of the E. coli catalytic subunit was lost in less than 1 min. On the other hand, the M. jannaschii PyrB gene product retained 75% of its activity after 60 min at the same temperature. The heat stability of the M. jannaschii PyrB gene product is very sensitive to the conditions used. In the heat step of the purification (90°C) almost no reduction in catalytic activity was observed; however, the thermal stability of the enzyme was greatly enhanced by the addition of ammonium sulfate.
Quaternary Structure of the Mj-PyrB and Mj-PyrI Gene Prod- ucts-The genomic organizations of the Mj-pyrB and Mj-pyrI and the E. coli pyrB and pyrI genes differ substantially. In E. coli, the two genes are contiguous, separated by only 12 base pairs. In contrast, the Mj-pyrB and Mj-pyrI genes are separated by 200,000 base pairs of intervening sequence. Furthermore, the genomic organization of the Mj-pyrB and Mj-pyrI genes suggest that they are regulated by separate promoters, whereas the E. coli pyrB and pyrI genes are regulated as a single pyrBI operon. The striking differences in genomic organization between these two organisms are difficult to assess in terms of physiological or evolutionary significance. Indeed, it has not been determined whether the Mj-PyrB and Mj-PyrI gene products are expressed differentially or concerted. The high degree of conservation between the M. jannaschii and E. coli pyrB and pyrI sequences, especially between residues characterized as important for quaternary structure formation in E. coli aspartate transcarbamoylase, led us to expect the quaternary structure of M. jannaschii aspartate transcarbamoylase would be similar to that found in E. coli.
In order to determine whether the Mj-PyrB and Mj-PyrI gene products exist as monomers or higher order species, a molecular weight analysis was performed by gel filtration. The molecular masses of the Mj-PyrB and Mj-PyrI gene products were determined against a set of standards and found to be 129 and 32 kDa, respectively. The monomer molecular masses of the Mj-PyrB and Mj-PyrI gene products were determined by SDS-PAGE analysis to be 37 and 17 kDa, respectively. These values compared favorably with the theoretical molecular masses of 33.7 and 17 kDa as calculated from the primary amino acid sequences. These data indicate that the Mj-PyrB gene product associates as a trimer and that the Mj-PyrI gene product asso-ciates as a dimer. The trimeric structure of the Mj-PyrB gene product and the dimeric structure of the Mj-PyrI gene product is consistent with structures of other class B aspartate transcarbamoylases.
Quaternary Structure of M. jannaschii Aspartate Transcarbamoylase-The above data suggest that the M. jannaschii aspartate transcarbamoylase is organized in the class B form. To test this hypothesis, the quaternary structure of the M. jannaschii aspartate transcarbamoylase that exists in vivo was determined directly from an M. jannaschii cell-free extract, using sucrose density gradient sedimentation.
After centrifugation through the sucrose density gradient, the M. jannaschii cell-free extract was fractionated, and each fraction was assayed for aspartate transcarbamoylase activity. As standards, a mixture of the catalytic subunit and the holoenzyme of E. coli aspartate transcarbamoylase were centrifuged through an identical sucrose gradient, fractionated, and assayed for activity. As seen in Fig. 5, the peak aspartate transcarbamoylase activity in the M. jannaschii crude cell-free extract sedimented at a position almost identical to that of the E. coli holoenzyme. These results are consistent with the notion that the Mj-PyrB and Mj-PyrI gene products associate to form a quaternary structure similar to that observed for the class B aspartate transcarbamoylases, such as the E. coli holoenzyme.
In further support of this notion, preliminary experiments have demonstrated that mixing the purified Mj-PyrB gene product and partially purified Mj-PyrI gene product results in the formation of a quaternary structure similar in mass to the E. coli holoenzyme. However, the in vitro reconstituted M. jannaschii holoenzyme was extremely difficult to manipulate due to very limited solubility of the newly formed holoenzyme. Future experiments are planned to optimize conditions for the in vitro reconstitution of the M. jannaschii holoenzyme. Nevertheless, in vivo and in vitro studies indicate that the Mj-PyrB and Mj-PyrI gene products assemble to form an aspartate transcarbamoylase of the class B type.
Homotropic and Heterotropic Interactions of the M. jannaschii Aspartate Transcarbamoylase-Preliminary experiments were performed on the M. jannaschii aspartate transcarbamoylase partially purified by sucrose density gradient centrifugation from M. jannaschii cell extracts. Aspartate saturation curves at 37°C revealed that at saturating concentrations of carbamoyl phosphate homotropic cooperativity with respect to aspartate was observed, although the value of the Hill coefficient was less than that observed for the E. coli enzyme. On the other hand, at subsaturating concentrations of a A unit of activity is defined as 1 mol of carbamoyl aspartate formed per min. Activity measurements were carried out by the pH-stat method (25) at 25°C with 30 mM aspartate and 4.8 mM carbamoyl phosphate. aspartate, ATP and CTP did not influence the activity of the enzyme (data not shown). Because of the crude enzyme preparation used, we consider these results tentative until a more highly purified preparation of the M. jannaschii aspartate transcarbamoylase can be tested. The extent of the homotropic and heterotropic interactions must also be evaluated at higher temperatures closer to the optimal growth temperature of M. jannaschii.
A Model of the Three-dimensional Structure of the Mj-PyrB and Mj-PyrI Gene Products-As seen in Fig. 1, the PyrB gene products from M. jannaschii and E. coli are extremely homologous, exhibiting 47% identity and 67% similarity, suggesting that they arose from a common ancestral gene. Residues determined to be critical for catalytic activity in the E. coli enzyme (32), including Ser 52 , 3 Arg 54 , Thr 55 , Arg 105 , His 134 , Gln 137 , Arg 167 , Arg 229 , Gln 231 , Ser 80 , and Lys 84 are all conserved in the Mj-PyrB gene product. As Ser 80 and Lys 84 are donated into the active site of one catalytic chain from an adjacent catalytic chain in E. coli, the conservation of these residues suggests that the active site in the Mj-PyrB gene product is also at the interface between chains and is shared.
The conservation of the Mj-PyrI gene product is slightly lower than that observed for the Mj-PyrB gene product with 35% identity and 52% similarity to the E. coli regulatory chain sequence (see Fig. 1). The residues that comprise the zincbinding site, Cys 109 , Cys 114 , Cys 138 , and Cys 141 are all conserved, strongly suggesting that this site is retained in the Mj-PyrI gene product. In the nucleotide effector binding site, all residues that interact with nucleotides via side chain interactions are conserved, including Asp 19 (40), His 20 (40), Lys 60 (41), and Lys 94 (42), suggesting that the Mj-PyrI gene product has a nucleotide-binding site.
In the interface between one catalytic chain in the top and one regulatory chain in the bottom of the molecule, the C1-R4 interface, the critical interface-stabilizing interaction between Lys 143 (43) and Asp 236 (44) is also conserved. In addition, many of the stabilizing interactions of the C1-R1 interface are also conserved. The only departure from interface residue conservation is in the R1-R6 interface, between the two regulatory chains of a dimer. Little of this interface is conserved; however, many of these interactions are backbone in nature and therefore do not rely on the nature of the specific side chains. Fig. 6A is a three-dimensional model of the Mj-PyrB gene product derived from the x-ray structural data for one catalytic chain of the E. coli enzyme. Fig. 6B shows a three-dimensional model of the Mj-PyrI gene product derived from the x-ray structural data for one regulatory chain of the E. coli enzyme. The models of both the Mj-PyrB and Mj-PyrI gene products are overall extremely similar to the corresponding proteins from E. coli. Small insertions and deletions are located on surface loops such as the 20-, 80-, and 240-s loop of the Mj-PyrB gene product and the 30-and 130-s loops of the Mj-PyrI gene products. Since the 80-and 240-s loops of the E. coli catalytic chain are important for catalysis and homotropic cooperativity, the observed alterations in these loops may result in alterations in these properties of the M. jannaschii enzyme. The modeling of the PyrB and PyrI gene products as well as the conservation of functionally critical amino acid side chains suggest that the M. jannaschii and E. coli enzymes catalyze the transcarbamoylase 3 The residue numbering is based on the E. coli amino acid sequence. 4 Within the E. coli holoenzyme, the catalytic chains of the top catalytic trimer are numbered C1, C2, and C3, whereas the catalytic chains of the bottom catalytic trimer are numbered C4, C5, and C6, with C4 under C1. The regulatory dimers contain chains R1-R6, R2-R5, and R3-R6. A regulatory chain is in direct contact with the same numbered catalytic chain. reaction by a similar mechanism and have a similar tertiary structure.
Comparison of the Amino Acid Sequence of the M. jannaschii and E. coli PyrB and PyrI Gene Products-There are several major differences in the amino acid composition of the M. jannaschii and E. coli PyrB and PyrI gene products. The percent of the large hydrophobic residue Ile and the charged residues Lys and Glu are all substantially higher in the PyrB and PyrI gene products of M. jannaschii as compared with the corresponding E. coli proteins. There is also a large decrease in the percent of Ala residues. A pairwise alignment of the deduced amino acid sequences between the two species indicated that the increased number of Ile residues found in the M. jannaschii PyrB and PyrI gene products largely corresponds to either a Val or Leu replacement in the PyrB and PyrI gene products. The increased incidence of the hydrophobic Ile residues in the M. jannaschii gene products occurred primarily within the hydrophobic core as deduced from the homology model. Overall, there is a 7 and 6% increase in the number of hydrophobic residues in the M. jannaschii PyrB and PyrI gene products, respectively, as compared with the corresponding E. coli proteins.
The increased incidence of charged residues, Lys and Glu, in the M. jannaschii PyrB and PyrI gene products is significant. In fact, there is a 37 and 23% increase in the number of charged residues that comprise the PyrB and PyrI gene products, respectively, as compared with the corresponding E. coli proteins. A pairwise alignment of the deduced amino acid sequences between the two species indicated that the increased number of charged residues found in the M. jannaschii corresponds to the replacement of either Ala or charged residues other than Lys or Glu in the E. coli sequence. The homology model of the M. jannaschii PyrB and PyrI gene products indicated that nearly all of the charged residues are exposed to solvent.
Structural studies of proteins from thermophilic organisms have led to characteristics that may be important in defining protein structural stability. Factors such as an increase in ionic bonds, core protein hydrophobic interactions, and the number of bulkier hydrophobic residues have been characterized in thermostable enzymes. Other factors include a decrease in the number of thermolabile residues such as Trp, Gln, Asn, and Cys and an increase in the number of Pro residues (9, 45). Furthermore, structural studies have demonstrated that pro-tein stabilization is enhanced by electrostatic interactions between an aspartic acid side chain at position N-2 of an ␣-helix with the positive charge at the end of the ␣-helix (45). For instance, aspartic acid substitutions occur at position N-2 of the putative ␣-H2 of the M. jannaschii regulatory chain and ␣-H7 and -8 of the catalytic chain as deduced from amino acid sequence and modeling analysis as compared with the E. coli counterpart. The thermostability of the M. jannaschii PyrB gene product may be partially attributed to these substitutions. Other factors may include the decreased number of Trp, Gln, Asn, and Cys residues and the increased number of proline and bulkier hydrophobic residues.
Steady-state Kinetics of the Mj-PyrB Gene Product-A steady-state kinetic analysis was performed on the purified Mj-catalytic trimer. Initial characterization of the Mj-catalytic trimer was carried out at 25°C. At this temperature, the Mjcatalytic trimer exhibited Michaelis-Menten kinetics with a maximal velocity 4.6-fold lower and a K m value of aspartate 5.8-fold higher than the E. coli catalytic trimer (see Table II). The temperature dependence of the Mj-catalytic trimer enzymatic activity was also investigated. At all temperatures tested, the aspartate saturation curves were hyperbolic. The E. coli catalytic trimer demonstrated a significantly higher max-  imal velocity as compared with that of the Mj-catalytic trimer at all temperatures tested. Based on the maximal observed specific activities, the activation energy of the transcarbamoylase reaction was calculated to be 9.4 Ϯ 0.5 and 8.6 Ϯ 0.5 kcal/mol for the Mj-catalytic trimer and E. coli catalytic trimer, respectively (Fig. 7). The near equivalence of the two activation energies implies that the rate-limiting step for the Mj-catalytic trimer and the E. coli catalytic trimer are likely the same.
Whereas both M. jannaschii and E. coli catalytic trimers demonstrate catalytic conservation and increased maximal velocity as a function of temperature, binding of substrate to the enzymes follows opposite trends. The K m value of aspartate for the Mj-catalytic trimer decreases as a function of temperature and, in contrast, the K m value of aspartate for the E. coli catalytic trimer increases as a function of temperature. Although the activation energies are nearly equivalent for both enzymes, the differences found between K m of aspartate and maximal velocity may have evolved independently of the catalytic mechanism in order to optimize enzymatic function in the cellular environs of each organism.
In this work we have compared the aspartate transcarbamoylases from E. coli and M. jannaschii. These two organisms are evolutionarily distant, yet their aspartate transcarbamoylases are remarkably similar. Further study of the M. jannaschii enzyme should provide insight into the means of its extreme heat stability and further elucidate the mechanisms of the homotropic and heterotropic properties of aspartate transcarbamoylase.