Leucyl-tRNA Synthetase Consisting of Two Subunits from Hyperthermophilic Bacteria Aquifex aeolicus *

In a hyperthermophilic bacterium, Aquifex aeolicus, leucyl-tRNA synthetase (LeuRS) consists of two non-identical polypeptide subunits (α and β), different from the canonical LeuRS, which has a single polypeptide chain. By PCR, using genome DNA of A. aeolicus as a template, genes encoding the α and β subunits were amplified and cloned in Escherichia coli. The α subunit could not be expressed stably in vivo, whereas the β subunit was overproduced and purified by a simple procedure. The β subunit was inactive in catalysis but was able to bind tRNALeu. Interestingly, the heterodimer αβ-LeuRS could be overproduced in E. coli cells containing both genes and was purified to 95% homogeneity as a hybrid dimer. The kinetics of A. aeolicus LeuRS in pre-steady and steady states and cross-recognition of LeuRS and tRNALeufrom A. aeolicus and E. coli were studied. Magnesium concentration, pH value, and temperature aminoacylation optima were determined to be 12 mm, 7.8, and 70 °C, respectively. Under optimal conditions, A. aeolicus αβ-LeuRS is stable up to 65 °C.


EXPERIMENTAL PROCEDURES
Materials-L-leucine, dithiothreitol, ATP, nucleotide triphosphate, 5Ј-GMP, tetrasodium pyrophosphate, and inorganic pyrophosphatase were purchased from Sigma. [ 14 C]leucine (300 -400mCi/mmol) and 32 Plabeled tetrasodium pyrophosphate were obtained from Amersham Biosciences. GF/C filters were from Whatman Company. Kinase, ligase, and restriction endonucleases were obtained from Sangon Company, Shanghai Branch. E. coli LeuRS was purified by chromatography on DEAE-Sepharose CL-6B and Blue-Sephacel from the overproducting E. coli strain in our laboratory (17). Its concentration was determined by A 280 of the enzyme solution; 1.62 mg of protein/ml equals 1 optical density unit at 280 nm. E. coli total tRNA containing 50% of tRNA Leu 2 (GAG) was isolated from an overproducing strain constructed in our laboratory (18). T7 RNA polymerase was purified from an overproducing strain in our laboratory (19).
Plasmids-pSML104 was constructed from pACYC184 and pKK-233-2 (20). It contains the p15A replicon from pACYC184, the strong trc promoter, a multicloning site, two sequences for transcription termination (T1 and T2) of the ribosomal operon rrnB from pKK233-2, and the resistance to chloramphenicol and tetracycline. Plasmid pBCP378 contains the trc promoter, resistance to ampicillin, and an NdeI site at its translation start point without altering the amino acid sequence of the synthesized protein, while ptrc and the lacI Q gene confer inducible and controllable expression (21).
Cloning and Expression of the Gene Encoding A. aeolicus tRNA Leu -The A. aeolicus genome contains five tRNA Leu genes. We cloned and expressed its most frequently used isoacceptor, which decodes CUC codons. The tRNA Leu gene with anticodon GAG is located from 5988 to 6072 in the A. aeolicus genome. Its sequence is GCCGGAGTGGCGG-AACTGGCAGACGCGCCGTCTTGAGGGGACGGTGCCCTTTACGGG-CGTGCGGGTTCACTCCCGCCTCCGGCA (anticodon is in bold italic). The cloning of the A. aeolicus tRNA Leu gene was performed as outlined previously (22). Six complementary and overlapping oligonucleotides encoding the gene and its complementary chain were synthesized chemically by Sangon Company, Shanghai Branch. Oligonucleotides were phosphorylated, hybridized, and ligated between the EcoRI and PstI sites of pTrc99B and then introduced by transformation into E. coli strain MT102. Transformants were checked by DNA sequencing, the cells were grown, and the crude tRNA extracts containing the A. aeolicus tRNA Leu were isolated as previously described (22).
Transcription and Labeling of tRNA in Vitro-Plasmids carrying A. aeolicus and E. coli tRNA Leu genes were prepared by procedures described previously (20). T7 transcripts were generated in a reaction mixture containing 40 mM Tris-HCl (pH 8.0), 5 mM dithiothreitol, 10 mM MgCl 2, 2 mM nucleotide triphosphates, 10 mM 5Ј-GMP, BstNI-digested template DNA (50 g/ml), 1 unit/ml of inorganic pyrophosphatase, and 2 mg/ml of pure T7 RNA polymerase for 4 h. The transcripts were purified by 20% (W/V) denaturing PAGE and annealed by heating for 2 min in 2 mM MgCl 2 at 75°C and allowed to slow cool to 30°C. The accepting activities of A. aeolicus and E. coli tRNA Leu transcripts were 760 and 650 pmol/A 260 , respectively. The tRNA Leu transcripts were incubated with calf intestinal phosphatase for 3 h at 37°C to remove the 5-terminal phosphate groups. This was followed by a phenol:chloroform extraction. Then, the labeling of the T7 transcripts was performed with T 4 polynucleotide kinase for 1 h at 37°C in the presence of [␥-32 P]ATP. The labeled tRNA Leu transcripts were extracted with phenol, precipitated with ethanol and suspended in an appropriate volume of water.
Cloning the A. aeolicus LeuRS ␣ and ␤ Subunit Genes-The genomic DNA from A. aeolicus was a gift from Dr. R Huber (University of Regensburg, Germany). The genes encoding ␣ and ␤ subunits of LeuRS were amplified by PCR. The ␣ subunit gene with NdeI and HindIII site at its ends was inserted into pBCP378 using the same sites. The gene of ␤ subunit with NdeI and SacI sites at its ends was ligated into pBCP378 or pSML104 using the same sites. E. coli TG1 was transformed with the recombinant plasmids.
Purification of A. aeolicus ␣␤-LeuRS and Its ␤ Subunit-The recombinant vectors pBCP378-lrsa, pBCP378-lrsb were used to overexpress the individual subunits. To overexpress the heterodimeric form ␣␤-LeuRS, both pBCP378-lrsa and PSML104-lrsb were co-transformed into the E. coli TG1 cells. Transformants harboring the recombinant plasmids were grown in 25 ml of LB with 100 g/ml of ampicillin and 20 g/ml of chloramphenicol at 37°C to stationary phase. Cells were harvested by centrifugation, resuspended in 4 ml of disruption buffer (100 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 and 1 mM EDTA), and sonicated for 6 ϫ 20 s at 15 W with a high-intensity ultrasonic processor (375W model). The crude extract was cleared of cellular debris by centrifugation at 12,000 rpm for 40 min and then analyzed by SDS-PAGE. Clarified extract was heated to 75°C for 1 h and centrifuged at 12,000 rpm for 40 min.
For purification, 4 liters of cells were treated as described above. The resulting supernatant was loaded onto a (3 ϫ 18 cm) DEAE-Sepharose CL-6B column equilibrated with lysis buffer and eluted with a potassium phosphate gradient from 50 to 500 mM. The active fractions were pooled, dialyzed against 10 mM potassium phosphate buffer, and applied to a (3.5 ϫ 18 cm) Blue-Sepharose CL-6B column. Elution was performed with a linear gradient of potassium phosphate buffer (from 50 to 500 mM). The fractions containing A. aeolicus ␣␤-LeuRS or its ␤ subunit were pooled and concentrated by dialysis against 10 mM potassium phosphate buffer, pH 6.8, containing 50% glycerol.
Binding Assay of ␤ Subunit of A. aeolicus LeuRS with tRNA-Protein-tRNA interactions were examined by a gel-shift assay. Increasing concentrations of the ␤ subunit were mixed with 0.05-0.1 M 5Ј-[ 32 P]tRNA Leu in a 20-l volume containing 100 mM Tris-HCl buffer, pH 6.8, 30 mM KCl, 12 mM MgCl 2 , 0.1 mM EDTA, and 1 unit/l RNasin at 20°C for 30 min. A 20% sucrose solution (3 l) containing tracer dyes was added immediately before loading. The final mixture was then loaded on a 1-mm thick, 8% native polyacrylamide gel in the reaction buffer for 3 h at 100 V at 15°C. After electrophoresis the gel was dried and exposed.
Determination of Molecular Mass-The molecular mass of ␣␤-LeuRS from A. aeolicus was determined by mass spectrometry analysis using an electrospray ionization-Quadrupole Ion Traps-mass spectrometry apparatus. The molecular mass of the ␤ subunit and its complex with tRNA Leu was determined by HPLC on a Superdex 200 HR 10/30 column.
Circular Dichroism (CD) Spectroscopy-Protein samples at a concentration of 0.6 mg/ml (for A. aeolicus ␣␤-LeuRS) or 0.2 mg/ml (for ␤ subunit) were analyzed on a Jasco J-715 spectropolarimeter with nitrogen purge at room temperature. A 1-mm path-length cuvette was used, and spectra were accumulated over five scans. Estimation of the secondary structure by CD spectrum was calculated according to the method from Yang (23).
Enzymatic and Kinetic Assay-ATP-PP i exchange and aminoacylation activities of LeuRS were measured at 37°C as described (24). The aminoacylation activity was determined at 37°C in the reaction mixture containing 100 mM Tris-HCl, pH 7.8, 30 mM KCl, 12 mM MgCl 2 , 4 mM ATP, 0.5 mM dithiothreitol, 10 M tRNA Leu , and appropriate amounts of [ 14 C]-labeled amino acids and enzymes. The kinetic constants of enzymes were determined using various concentrations of the relevant substrates. The same assays were carried out at 60°C. Presteady state kinetics was performed with a quenched-flow apparatus according to Ref. 25. The rate of aminoacylation was measured at 40°C in the following reaction mixture: 100 mM Tris-HCl, pH 7.8, 30 mM KCl, 12 mM MgCl 2 , 4 mM ATP, 3.34 M tRNA, 5.67 M [ 14 C]leucine, 0.1 mg/ml bovine serum albumin, 2.5 mM GSH, and 1.66 M A. aeolicus ␣␤-LeuRS.
Active Site Titration-The titration was performed according to method of A. Fersht (26). In the amino acid activation reaction, when the initial concentration of ATP ([ATP] 0 ) was much higher than that of A. aeolicus ␣␤-LeuRS ([E] 0 ), the number of moles of aminoacyl adenylate/mole of A. aeolicus ␣␤-LeuRS, that is the number of active sites (n), may be deduced from the following equation:

RESULTS
Cloning and Expression of the tRNA Leu Gene-The plasmid containing A. aeolicus tRNA Leu (GAG) gene was constructed and used to transform E. coli MT102. In this strain, which overexpresses tRNA Leu , the charge of [ 14 C]leucine on the tRNA reaches 724 pmol/A 260 . As the accepting activity of pure tRNA Leu was considered to be 1600 pmol/A 260 , tRNA Leu should be about 45% of the total tRNA. This crude overexpressed A. aeolicus tRNA Leu was used to assay A. aeolicus LeuRS activity and binding property. The low content of endogeneous E. coli tRNA Leu was considered to be negligible compared with the overexpressed A. aeolicus tRNA Leu .
Cloning and Expression of the A. aeolicus LeuRS ␣ and ␤ Subunit Genes-When compared with the sequence of Thermus thermophilus LeuRS, it appears that, except for 30 residues at the C-terminal end, the ␣ subunit sequence corresponds to the 580 first residues of the T. thermophilus LeuRS (Fig. 1). Except for the first 40 residues, the ␤ subunit sequence is similar to the peptide fragment 636 -878 of T. thermophilus LeuRS. Thus, at the junction of the two peptides, two additional domains of about 30 and 40 residues may form some connecting structure between the two subunits ( Fig. 1). Theoretically, the molecular masses of ␣ and ␤ subunits are 74.0 and 33.5 kDa, respectively. The genes encoding ␣ and ␤ subunits of LeuRS were amplified by PCR. Both DNA fragments were ligated into pBCP378 or pSML104. The resulting recombinant vectors, pBCP378-lrsa, pBCP378-lrsb and PSML104-lrsb were used to transform E. coli TG1. After growth of the different transformants and disruption of the cells, a strong signal corresponding to the ␤ subunit was detected by SDS-PAGE (lane 5, Fig. 2A). However the ␣ subunit was not detected in the crude extract (lane 4, Fig.  2A). The ␤ subunit was purified to 85% homogeneity by heating the crude extract at 75°C for 30 min and centrifugation (lane 6, Fig. 2A).
Overproduction and Purification of A. aeolicus ␣␤-LeuRS-To overexpress the heterodimeric form ␣␤-LeuRS, both pBCP378-lrsa and PSML104-lrsb were co-transformed into the E. coli cells. In these conditions, the ␣ and ␤ subunits were detected in the crude extract (lanes 2 and 3, Fig. 2A). A. aeolicus ␣␤-LeuRS was purified to about 70% homogeneity after heating at 75°C for 30 min followed by centrifugation. Thus, in vivo, in the presence of the ␤ subunit, the ␣ subunit was stable and thermostable.
After growth of the transformants in 4 liters of Luria broth medium (16 g of wet cells) about 20 mg of purified A. aeolicus ␣␤-LeuRS was obtained as described under "Experimental Procedures." The purified A. aeolicus ␣␤-LeuRS was 95% homogeneous as determined by SDS-PAGE (lane 2, Fig. 2B). A. aeolicus LeuRS consisting of ␣ and ␤ subunits was active, and the molecular mass of the enzyme was 107.5 kDa as measured by mass spectrometry determination. These data demonstrate that A. aeolicus LeuRS is a hybrid dimer active in its dimeric form.
CD Spectra of A. aeolicus ␣␤-LeuRS and its ␤ Subunit-To understand why the ␣ subunit was unstable, the secondary structures of A. aeolicus ␣␤-LeuRS and its ␤ subunit alone were compared for their CD spectra. The parameters of their secondary structures were estimated, as summarized in Table I (23).
The ␤ subunit contains more ␣ helix and ␤ sheet, less ␤ turn and random than A. aeolicus ␣␤-LeuRS. Thus, the ␤ subunit is more structured and probably more rigid than A. aeolicus ␣␤-LeuRS, probably due to the association with the ␣ subunit, which is probably less structured and less stable than the ␤ subunit as suggested by its inability to be overexpressed in vivo.
Binding of ␤ Subunit with tRNA Leu -Molecular mass of the ␤ subunit was 33.5 kDa as determined by HPLC gel filtration. It means that ␤ subunit is a monomer when expressed independently from the ␣ subunit. However, the ␤ subunit is inactive and cannot bind ATP and leucine as shown by competition assay and gel-shift assay (data not shown). By gel-shift assay we detected the binding of the ␤ subunit to the tRNA Leu from A. aeolicus (Fig. 3A). The binding constant, K d , calculated by screening the area of the bands was about 20 Ϯ 3 M, an unusually high value for an aaRS:tRNA interaction. This is in contrast with the K d value measured for the whole enzyme (0.5 M, data not shown). In general, the K d of aaRS for the cognate tRNA is Ͻ2 M, which means that the binding of the ␤ subunit to tRNA Leu is not very tight. Interestingly, a supershift appears on increasing the concentration of unlabeled tRNA Leu during the competition assay (Fig. 3B). Thus, it seems that the ␤ subunit may form two kinds of complexes, the upper band corresponding to the ␤ subunit with a probable tRNA/enzyme stoichiometry of 2/1. Alternatively, the ␤ subunit can bind labeled E. coli tRNA Leu (data not shown) and tRNA Arg (Fig. 3C) with two forms also. The assay of specificity was extended to tDNA encoding A. aeolicus tRNA Leu , the tDNA competed with the homologous A. aeolicus tRNA Leu , but the large complex did not appear (Fig. 3D).
To confirm the existence of two tRNA binding sites on the ␤ subunit the complex was generated according to the experimental conditions of Fig. 3 and analyzed by gel filtration on HPLC. Two species were separated, with M r of 63.5 and 93.3 kDa, corresponding to the ␤ subunit associated with one or two tRNA Leu respectively. Thus, the isolated ␤ subunit exhibits two tRNA binding sites, one of them being nonspecific at the level of its tRNA recognition specificity.
Active Site Titration-At 60°C, the plot of [ATP] t against t was a straight line (Fig. 4A). The number of active sites was 1.1 Ϯ 0.1 (the value averaged from three independent determinations), which is consistent with that from E. coli LeuRS (27).
Kinetic Constants of A. aeolicus ␣␤-LeuRS-The specific activities and kinetic parameters of A. aeolicus ␣␤-LeuRS in amino acid activation and aminoacylation reactions were assayed at 37°C in order to directly compare them with those of E. coli LeuRS under reaction conditions, which are optimized for that of the E. coli enzyme.
The E. coli and A. aeolicus LeuRS display a remarkable difference in the aminoacylation reaction conditions. For the A. aeolicus enzyme, the magnesium/ATP ratio required for optimal activation and charging reaction is 3 (8 mM excess of free magnesium), an unusually high value for a class I aminoacyl-tRNA synthetase. Compared with the stoichiometric ratio of magnesium/ATP used for the E. coli enzyme (28), this suggested that additional magnesium ions are used by the thermophilic enzyme to stabilize the productive enzyme-substrate complex. In the ATP-pyrophosphate exchange reaction the Michaelis-Menten constant (K m ) values for leucine and ATP were 1.3 and 360 M, respectively, and k cat were 3.5 and 3.3 s Ϫ1 , respectively. Compared with the E. coli LeuRS, these values only reach 2% of the level of activity (Table II, (24)). In the aminoacylation reaction, K m values for leucine, ATP, and tRNA Leu (A. aeolicus) were 6.4, 550, and 0.38 M; k cat values were 0.39 s Ϫ1 (Table II). Except K m for ATP in aminoacylation reaction, all of the constants were lower than those for E. coli LeuRS (Table II) (24). Although the substrates bind A. aeolicus ␣␤-LeuRS tightly, the reaction is slower at 37°C.
The kinetic constants of A. aeolicus ␣␤-LeuRS were also studied at 60°C (Table III). All k cat values increased 3.5-4.8fold as compared with those at 37°C. In amino acid activation, the K m values for both leucine and ATP were identical to those at 37°C. In the aminoacylation reaction, the K m value for leucine was unchanged, whereas the K m value for ATP was slightly increased. At 60°C, the catalytic efficiencies observed for the different substrates are comparable to those of the E. coli LeuRS at 37°C.
The aminoacylation reaction was studied at the pre-steady state level using a quenched flow apparatus (25). The reaction was started by mixing the tRNA to the preformed enzyme: adenylate complex. A two-phase kinetics was detected, with an  initial faster rate (k cat ϭ 7.2 s Ϫ1 ) followed after 52 ms by a slower rate with a k cat value of 0.36 s Ϫ1 , which is similar to the value measured by hand in steady state, 0.39 s Ϫ1 . The Leu-tRNA Leu synthesized during the burst was in a 0.82:1.0 stoichiometric ratio with the synthetase present (Fig. 4B). Thus, the decrease of the rate of tRNA Leu charging after completion of the first catalytic cycle of the synthetase suggests that one or several end products dissociate slowly from the enzyme.
Thermal Properties of A. aeolicus ␣␤-LeuRS-As expected, A. aeolicus ␣␤-LeuRS is a thermostable enzyme. Under the experimental conditions it is stable up to 70°C (Fig. 5). The specific activities of amino acid activation and aminoacylation increased with temperature. At 75°C the two activities were 12-and 5-fold greater than at 37°C. Changes of activity with temperature are shown in Fig. 6. The activation energy was calculated from the slope of the line when log v was plotted against 1/T. The activation energy for the amino acid activation reaction was 14.9 kcal/mol. Below 50°C, the activation energy for the aminoacylation reaction was 14.9 kcal/mol, the same as in amino acid activation. Above 50°C the value was 6.53 kcal/ mol. Above 75°C, evaporation in the test tube prevented accurate determination of activation energy. The aminoacylation rate began to decrease above 70°C, 15°C under its optimum of 85°C (15). This suggests that the composition of the reaction mixture is not optimal for enzyme activity, or that the A. aeolicus tRNA Leu overproduced in E. coli does not possess all the modified bases that should improve its thermal stability.
Cross-recognition of LeuRSs and tRNA Leu from A. aeolicus and E. coli-At 37°C, both enzymes display for their homologous substrates comparable k cat /K m values, with the faster rate for the E. coli enzyme and the best tRNA affinity for the A. aeolicus enzyme (Table II). In the cross-recognition of the tRNAs, the Aquifex enzyme exhibits a 5-fold decrease of the k cat in the aminoacylation of the E. coli tRNA and a 2-fold decrease of the tRNA affinity. The resulting 10-fold decrease of k cat /K m suggests that the heterologous Aquifex ␣␤-LeuRS/ E. coli tRNA Leu complex does not reach an optimal conformation for catalysis. Increasing the temperature increases the k cat to 0.4 s Ϫ1 , still seven times lower than that of E. coli LeuRS at 37°C. DISCUSSION Analysis of the DNA sequence of A. aeolicus LeuRS revealed that this organism presents an unusual split LeuRS in contrast FIG. 5. Determination of the thermal stability of A. aeolicus ␣␤-LeuRS. A. aeolicus ␣␤-LeuRS (40 g/ml) in 50 mM potassium phosphate buffer, pH 6.8, containing 400 g/ml bovine serum albumin was incubated at various temperatures for 10 min. Then the aminoacylation activity was assayed at 37°C after dilution of the reaction mixture with cold 50 mM potassium phosphate buffer, pH 7.5. to the monomeric form found in all other sequenced LeuRS genes (15). However, despite this difference in quaternary structure, extensive sequence homologies are found with other LeuRSs, with a conservation of the HIGH sequence (HMGH), KMSKS sequence (-MSKS) and CP domain presumably involved in editing. The enzyme is split at the level of the active site, between the two halves of the Rossmann fold, and before the second consensus sequence KMSKS. By analogy with the T. thermophilus LeuRS whose structure has been solved by x-ray crystallography, the split should occur in the additional domain preceding the KMSKS sequence (29).
Based on the sequence homologies and x-ray structure of the T. thermophilus enzyme, we can postulate that the ␤ subunit of the A. aeolicus enzyme possesses half of the active site, with the CP domain generally involved in editing attached to the Rossmann fold (Fig. 1) (29). The ␤ subunit should contain the second part of the active site and the helix bundle domain found in the class Ia synthetases (Arg-, Met-, Val-and IleRS). This domain is involved in tRNA-anticodon binding as shown by x-ray crystallography (30 -32) and functional studies (33,34). An additional domain located at the C-terminal end of the ␣ subunit, characteristic of the T. thermophilus and A. aeolicus enzymes corresponds to the split site for the A. aeolicus enzyme. In light of the structural data, one may imagine that the ␤ subunit is stably expressed due to the central helix bundle domain that may strongly structure the subunit. On the other hand, the ␣ subunit is unstable apart from the ␤ subunit, perhaps due to its modular structure composed of two large domains, namely half of the Rossmann fold and the whole CP domain (200 residues), connected by long hinge peptides, which may induce flexibility and the observed instability of the A. aeolicus enzyme (29).
The ␤ subunit is small; however, in vivo or in vitro it is stable and seems to be a chaperon of the ␣ subunit in vivo. Although inactive in its monomeric form, the ␤ subunit binds tRNA Leu with low affinity (K d ϭ 20 Ϯ 3 M). Compared with the dimeric form ␣␤ (K d ϭ 0.5 M, K m ϭ 0.38 M), it is a 40-fold difference, which may account for the loss of interactions with the acceptor arm of the tRNA molecule, assuming that the tRNA binds the ␤ subunit in its original mode. A second, unspecific tRNA binding site was characterized on the ␤ subunit, which may be related to the results observed for ValRS (35). Recently it has been shown that LeuRS from different sources can substitute for the splicing function of yeast mitochondrial LeuRS (36). Thus, LeuRSs intrinsically possess the ability to bind diverse RNA substrates, even larger than tRNA. This may explain why two tRNA molecules can bind simultaneously.
The low affinity value exhibited by the ␤ subunit for its tRNA is similar to the values measured for the additional domains found in eukaryotic synthetases, like the C-terminal extension of MetRS related to EMAPII (K d ϭ 15 M) (37), the repeated domains of GluProRS (K d ϭ 2.5-30 M) (38,39) and the p43 component of the mammalian multisynthetase complex (40 M) (40). It has been shown that the additional domains increase the catalytic efficiency of the synthetase without contributing to the specificity. Thus, they can be partially compared with the ␤ subunit of the Aquifex enzyme, which represents a later stage in the evolution at which point the added domain became more specific for the tRNA substrate.
The ␣␤-LeuRS from A. aeolicus is split in the middle of the bipartite Rossmann fold domain in a place where the ancestor two half Rossmann folds were juxtaposed in order to give the actual class I active site. Thus, the split A. aeolicux enzyme seems to have gone back in evolution to a stage when the class I active site was not assembled as the actual bipartite Rossmann fold.
The dimerization of the A. aeolicus enzyme raises other questions concerning the assembly of the subunits. No additional domains involved in assembly are detectable by sequence analysis. Thus, the intramolecular contacts found in the monomeric enzyme should have been transformed into intermolecular contacts after separation of the gene into pieces. Previous work on E. coli IleRS demonstrates that this enzyme, closely related with LeuRS, could be expressed in different combinations of split enzymes (16,41). Thus, it seems that these large monomeric enzymes have the built-in capacity to be synthesized as heterodimers. Similarly, four pairs of fragments corresponding to N-and C-terminal parts of E. coli LeuRS with junctions at different positions in the CP domain were expressed and assembled in vivo. Except the pair with the junction at the level Glu-292-Ala-293, the assembled proteins were active for aminoacylation (42). However, the Glu-292-Ala-293 mutant still supported the ATP-PP i exchange activity (42). In fact, the assembled E. coli LeuRS mimicked a heterodimer LeuRS. We recently tried to split the E. coli LeuRS in order to mimic the ␣ and ␤ subunits of A. aeolicus LeuRS; however, the fragments were not stably expressed in E. coli.