New Class of IMP Cyclohydrolases in Methanococcus jannaschii

Abstract

The enzyme responsible for observed IMP cyclohydrolase activity in Methanococcus jannaschii was purified and sequenced: its genetic locus was found to correspond to gene MJ0626. The MJ0626 gene was cloned, and its protein product was expressed in Escherichia coli and shown to catalyze the cyclization of 5-formylamidoimidazole-4-carboxamide ribonucleotide to IMP. The enzyme has no sequence similarity to known enzymes, and its catalytic properties appear distinct from any characterized IMP cyclohydrolase. The purO gene for the enzyme is currently found only in the domain Archaea.

The conversion of 5-aminoimidazole-4-carboxamide ribonucleotide (ZMP or AICAR) to IMP represents the last steps in the de novo biosynthesis of purines. In bacteria and eukaryotes this reaction occurs in two separate steps catalyzed by a single bifunctional enzyme, aminoimidazole carboxamide ribonucleotide transformylase/IMP cyclohydrolase (ATIC), encoded by the purH gene. The first reaction is catalyzed by AICAR transformylase, which resides in the carboxyl terminus of the enzyme, and transfers the formyl group of 10-formyltetrahydrofolate to AICAR with the formation of 5-formaminoimidazole-4-carboxamide ribonucleotide (FAICAR). The product of this reaction is then cyclized by the IMP cyclohydrolase which resides in the amino terminus of the enzyme (9). ATIC is exceptional among known bifunctional enzymes since, up to now, it has been found to function only as a bifunctional enzyme. The two-domain structure of the ATIC enzyme has recently been confirmed by the X-ray crystal structure of the avian enzyme (3).

Recent data have led investigators to question the presence of ATIC in some of the Archaea. In some Archaea (13), folate is not required for this reaction. Genomic DNA sequence comparisons indicate the absence of purH genes in the domain Archaea (5, 10). To identify the gene encoding IMP cyclohydrolase in the Archaea, we have purified the enzyme from Methanococcus jannaschii, obtained partial amino acid sequence data, and determined its genetic locus. The catalytic properties of the enzyme are distinct from any preexisting characterized IMP cyclohydrolase.

The cell extract of M. jannaschii was obtained by the sonication of 1 g (wet weight) of cells suspended in 3 ml of a buffer containing 50 mM TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid], 10 mM MgCl₂, and 20 mM mercaptoethanol (pH 7.0). Biswarup Mukhopadhyay, Department of Microbiology, University of Illinois, Urbana (8), kindly supplied frozen cells of M. jannaschii. After centrifugation (15 min at 16,000 × g), the resulting supernatant was saturated with solid ammonium sulfate at room temperature. The precipitated proteins were removed by centrifugation (15 min at 16,000 × g), and the pellet was washed with the same volume of saturated ammonium sulfate. After centrifugation (15 min at 14,000 × g), the pellet was dissolved in 7 ml of 25 mM TES buffer, pH 7.0, and dialyzed 3 times against this buffer. Half of the resulting material was applied to a MonoQ HR 10/10 column (Pharmacia) equilibrated in this buffer, and the proteins were eluted with a linear gradient in the same buffer to 1 M NaCl at a flow rate of 1 ml/min. The UV absorbance of the eluent was monitored at 280 nm. Fractions (3 ml) were collected and stored frozen. Portions of the fractions (50 μl) were assayed for IMP cyclohydrolase activity by incubation with 0.7 mM FAICAR for 30 min at 50°C. The FAICAR was prepared as described by Flaks et al. (1). After dilution with 950 μl of water, the UV absorbances were read at 223 nm (FAICAR) and 249 nm (IMP), and the change in the absorbance ratio was used to monitor the activity. To verify the formation of IMP, incubation mixtures were diluted with 250 μl of 25 mM TES buffer, pH 9.0, and the total sample was applied to a MonoQ HR 5/5 column equilibrated in the same buffer. The substrates and products were eluted with a linear gradient in the same buffer to 1 M NaCl at a flow rate of 1 ml/min. The elution of the substances was monitored at a UV absorbance of 254 nm. Under these conditions, FAICAR eluted at 15 min and IMP eluted at 19 min.

From the protein separation on the MonoQ HR 10/10 column, one peak of enzymatic activity was observed eluting at approximately 0.2 M NaCl. The fractions containing the activity were combined and concentrated in a 10-ml concentrator (Amicon) to 1 ml and applied to an SP-Sepharose (1 by 10 cm) column. The column was equilibrated with 25 mM MES [2-(N-morpholino)ethane sulfonic acid] buffer at pH 6.0. Proteins were eluted with a linear gradient of this buffer to 1 M NaCl at a flow rate of 1 ml/min. Fractions of 1 ml were collected and assayed for IMP cyclohydrolase activity by using the standard assay, which measured the increase in absorbance due to IMP formation (9). Most of the activity was found in the very early fractions, indicating that the SP-Sepharose did not significantly bind the IMP cyclohydrolase. The fractions containing activity were combined and concentrated by ultrafiltration with centricons (Amicon-10) to 150 μl. After dilution with 2 M ammonium sulfate-25 mM TES, pH 7.0, the sample was applied to a phenyl-Sepharose (1 by 10 cm) column equilibrated in 1 M ammonium sulfate and 25 mM TES buffer at pH 7.0. The proteins were eluted with a linear gradient of decreasing salt concentration to 25 mM TES. Fractions (1 ml) were collected and assayed for activity. The activity was observed to spread out over a broad range of fractions. Fractions containing most of the activity were combined and concentrated with the Amicon concentrator and applied to a Superose 12 HR 10/30 column (Pharmacia). The column was equilibrated in 50 mM TES and 150 mM NaCl buffer, pH 7.0. The proteins were eluted isocratically with this buffer at a flow rate of 0.5 ml/min. Fractions of 0.25 ml were collected. The enzymatic activity of each fraction and sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS-12% PAGE) analysis were determined. Silver staining revealed only one protein band pattern that corresponded in intensity to the measured IMP cyclohydrolase activity. The probable IMP cyclohydrolase comigrated with the 21.5-kDa soybean trypsin inhibitor standard on the SDS-PAGE gel. The active fractions from the Superose 12 column were combined, precipitated with trichloroacetic acid (final concentration, 5%), and separated using an SDS-12% PAGE gel. The gel was stained with Coomassie blue, and the IMP cyclohydrolase band was sequenced by mass spectrometry at the W. M. Keck Biomedical Mass Spectrometry Laboratory, University of Virginia Health System. Analysis of the sequence data revealed the presence of a single enzyme that corresponded to one hypothetical protein encoded by the MJ0626 gene. Protein sequence coverage was 67% based on amino acid count. The MJ0626 gene was cloned and overexpressed as previously described (2) with 5"-GGCCATATGTATATTGGAAG-3" and 5"-CGGGATCCTTATTTGCCCTTAG-3" as the PCR primers for the MJ0626 gene and pET-19b as the vector for recombinant plasmid PMJ0626. The enzyme was purified by heating the Escherichia coli cell extract at 80°C for 10 min to remove the E. coli proteins followed by (NH₄)₂SO₄ precipitation and Sephadex G-25 chromatography as previously described to purify M. jannaschii proteins (2).

The predicted protein derived from the M. jannaschii MJ0626 gene aligns only with two other archaeal proteins: the Methanobacterium thermoautotrophicum MTH1020 gene and the Halobacterium sp. NRC-1 VNG2371C gene. No homologs are found in bacteria or eukaryotes (BLAST search). The enzyme has no sequence similarity to ATIC. As expected, the enzyme is a hyperthermophilic enzyme which loses none of its activity after heating for 1 h at 80°C in air. Heating in air for 30 min at 90 and 100°C resulted in the loss of 16 and 70% of the activity, respectively. The enzyme showed a broad pH optimum in the range of 6 to 8.5 and had a K_m for FAICAR of 82 μM with a k_cat/K_m ratio of 2.7 × 10⁵ M⁻¹ s⁻¹. The enzyme was not inactivated by incubation with 20 mM iodoacetamide for 30 min or by running the assay in the presence of 2 mM EDTA. XMP (50 μM) or IMP (20 μM) did not significantly inhibit the reaction using 54 μM FAICAR as a substrate. These data are to be compared to those reported for the human enzyme which shows a broad pH optimum from pH 8 to 9, a K_m of 0.87 μM, inactivation by iodoacetamide, and a K_i for XMP of 20 μM (11). These data clearly indicate that the two enzymes are not mechanistically or evolutionally related.

The important question arises as to why these archaea contain a different enzyme for the IMP cyclohydrolase reaction. The answer to this question for the two methanogens is likely to be related to the substitution of tetrahydromethanopterin for tetrahydrofolate as the C₁ carrier in these archaea (12, 14, 15). A bifunctional ATIC containing the formate requiring AICAR transformylase would then have been required to use 10-formyltetrahydromethanopterin as the biochemical formyl donor, a function that it has never been demonstrated to do (6). The solution to this problem for these methanogens was the evolution of the ATP- and formate-dependent enzyme (13). In some way the IMP cyclohydrolase was also required to change in this process. The evolution of both a folate-dependent and folate-independent reaction to carry out the same transformation is already known in E. coli for another purine biosynthetic enzyme, 5-phosphoribosylglycinamide (GAR) transformylase. Here GAR transformylase-N (purN) catalyzed the formation of 5"-phosphoribosyl-N-formylglycinamide (FGAR) from GAR and 10-formyltetrahydrofolate (4), whereas GAR transformylase-T (purT) used ATP and formate (7).

Halobacterium, which used the normal folate cofactors (12), has an enzyme homologous to the formate-dependent AICAR transformylase portion of ATIC but no gene homologous to the IMP cyclohydrolase portion of ATIC. Rather, it has a gene homologous to the purO IMP cyclohydrolase. In fact, in Halobacterium, the AICAR transformylase is attached to purN, the other folate-requiring gene in purine biosynthesis. Archaeoglobus fulgidus, on the other hand, has the IMP cyclohydrolase purH but no AICAR transformylase purH. In total, we find that the distribution of these IMP cyclohydrolase genes and/or domains in the archaea to be quite varied, as is outlined in Fig. 1. These data indicate that additional archaeal IMP cyclohydrolase/AICAR transformylase-related genes are still to be discovered in the Archaea.

FIG. 1.

Occurrence and associations of the IMP cyclohydrolase/AICAR transformylase (purH) and the IMP cyclohydrolase (purO) genes among the following: E. coli (Swiss-Prot accession no. P15639), humans (GenBank accession no. AAH08879.1), Thermoplasma acidophilum (EMBL accession no. CAC11208.1), A. fulgidus (GenBank accession no. AAB89440.1), Halobacterium sp. purO (GenBank accession no. AAG20467.1) and purHN (GenBank accession no. AAG18965.1), M. jannaschii (Swiss-Prot accession no. Q58043), M. thermoautotrophicum (GenBank accession no. AAB85516.1), Pyrococcus horikoshii (GenBank accession no. BA000001), and Aeropyrum pernix (GenBank accession no. BA000002).