A New Subfamily of the Nudix Hydrolase Superfamily Active on 5-Methyl-UTP (Ribo-TTP) and UTP*

A new subfamily of the Nudix hydrolases, identified by conserved amino acids upstream and downstream of the Nudix box, has been characterized. The cloned, expressed, and purified orthologous enzymes have major activities on the non-canonical nucleoside triphosphate 5-methyl-UTP (ribo-TTP) and the canonical nucleotide UTP. In addition to their homologous signature sequences and their similar substrate specificities, the members of the subfamily are inhabitants of or are related to the bacterial rhizosphere. We propose the acronym and mnemonic, utp, for the gene designating this unique UTPase.

The Nudix hydrolases, so named because they catalyze the hydrolysis of Nucleoside diphosphates linked to some other moiety x (1), constitute a superfamily of enzymes with representatives in all three kingdoms. The members of this superfamily can be identified by a highly conserved amino acid signature sequence called the Nudix box, viz.
GXXXXXEXXXXXXXREUXEEXGU SCHEME 1 in which X represents any amino acid and U is a bulky hydrophobic amino acid, usually Ile, Leu, or Val. A current BLAST (2) search of the data banks for polypeptides and expressed sequence tags containing the above sequence, reveals over 1100 open reading frames from more than 250 species ranging from viruses to humans. We have been systematically cloning, expressing, and characterizing members of the superfamily, and without exception, all of the enzymes identified so far hydrolyze nucleoside diphosphate derivatives including nucleotide sugars, dinucleoside polyphosphates, coenzymes, (deoxy)nucleoside triphosphates, and ADP-ribose. Since the same amino acid signature sequence, the Nudix box, represents the nucleotide binding and catalytic site for all these proteins (3)(4)(5), the different specificities toward the respective substrates must lie in a region or in regions peripheral to this area. By aligning amino acid sequences of those enzymes hydrolyzing a similar spectrum of substrates, we have identified certain landmark amino acids outside of the Nudix box, enabling us to classify some of the members of the superfamily into distinct subfamilies (6). This has been of singular value, because it has allowed us to predict the enzymatic activity of unidentified open reading frames involved in physiological processes, merely by observing specific telltale amino acid patterns peculiar to each subfamily.
In this paper, we describe the cloning, expression, purification, and identification of members of a new subfamily of the Nudix hydrolases, highly active on 5-methyl-UTP (ribo-TTP) and UTP, and recognizable by characteristic amino acid sequences outside of the Nudix box.

Materials
The expression plasmid pET-24a(ϩ) (Km r ) and Escherichia coli HMS174 (DE3) were from Novagen (Madison WI); Agrobacterium tumefaciens strain GV3101 was a gift from Judith Bender, The Johns Hopkins School of Public Health; Pseudomonas aeruginosa, strain BB3/ 216/ATCC27853, was a gift from Thomas Cebula of the Food and Drug Administration; and Caulobacter crescentus strain CB15 was a gift of Lucy Shapiro, Stanford University, and Bert Ely, the University of South Carolina. The plasmid, pTrc99A (Amp r ), and Sephadex G-100 were from Amersham Biosciences, and the restriction enzymes, PCR kits, calf intestinal alkaline phosphatase, and yeast inorganic pyrophosphatase were from Stratagene (La Jolla, CA). The canonical nucleotide substrates and common chemicals were from Sigma or Invitrogen, and 5-methyl-UTP was from TriLink Biotechnologies (San Diego, CA). The PCR primers were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA). Protein assay reagent was from Bio-Rad.

Methods
Cloning of A. tumefaciens orf147 and Its Orthologs-Standard cloning technology was used. Briefly, orf147 was amplified directly from one colony of A. tumefaciens strain GV3101 using the DNA PCR. NdeI and BamHI restriction sites were incorporated at the start and end of the gene, respectively, and the amplified DNA was purified, digested with NdeI and BamHI, and ligated into pET24a(ϩ) to place the gene under transcriptional control of the T7 lac promoter. The resultant plasmid, pETORF147 was transformed into E. coli DH5a for storage and into HMS174 (DE3) for protein expression. The orthologous genes from P. aeruginosa and C. crescentus were cloned using similar procedures. For complementation studies, the gene was cut out of pETORF147 with XbaI and HindIII and ligated into the corresponding sites of pTrc99A putting orf147 under control of the trc promoter. The resulting plasmid, pTrcORF147, was transformed into DH5␣.
Growth and Expression of HMS174:pETorf147, ORF147, and Its Orthologs-One colony of the expression strain was inoculated into 40 ml of LB medium containing 30 g/ml kanamycin, incubated at 37°C on a rotary shaker overnight, transferred to 2 liters of the same medium, and grown to an A 600 of 0.8. The culture was derepressed by the addition of isopropyl-␤-D-thiogalactopyranoside to a concentration of 0.5 mM and grown for an additional 3 h, after which the cells were harvested by centrifugation and washed in buffered isotonic saline solution.
Purification of the Enzymes-The washed cells were frozen at Ϫ80°C overnight and resuspended in 2.5 volumes of buffer A (50 mM Tris-Cl, pH 7.5, 1 mM EDTA, 0.1 mM dithiothreitol). The supernatant (fraction I) containing the expressed protein was collected by centrifugation, and 10% streptomycin sulfate in buffer A was added to a final concentration of 1.5% to precipitate nucleic acids. These were removed by centrifugation, and the supernatant (fraction II) was brought to 60% saturation * This work was supported by National Institutes of Heath Grant GM18649. This is publication 1526 from the McCollum-Pratt Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
with ammonium sulfate. The precipitate was collected and dissolved in a minimal volume of buffer A (fraction III), and chromatographed on a Sephadex G-100 gel filtration column equilibrated and eluted with buffer A plus 200 mM sodium chloride. The fractions were located by A 280 absorbance and identified by their migration on an SDS-PAGE gel. Those containing the purified protein were pooled, assayed, and stored at Ϫ80°C (fraction IV). All of the orthologs of ORF147, including those from A. tumefaciens, P. aeruginosa, and C. crescentus were purified by similar procedures.
Enzyme Assays-The standard reaction contained in 50 l: 50 mM Tris-Cl, pH 9.0, 5 mM Mg 2ϩ , 2 mM substrate, 0.5 unit of yeast inorganic pyrophosphatase for substrates such as (deoxy)nucleoside triphosphates and their derivatives, or 4 units of alkaline phosphatase for all other substrates and 0.1-2 milliunits of purified enzyme. The solution was incubated at 37°C for 15 min, stopped by the addition of 250 l of 4 mM EDTA, and the liberated inorganic orthophosphate was assayed by the colorimetric procedure of Fiske and SubbaRow (7) as modified by Ames and Dubin (8). A unit of enzyme hydrolyzes 1 mol of substrate per min.
Complementation Test for the mutT Mutator Phenotype-Gene orf147 was cloned into pTrc99A and transformed into E. coli strain SB3 lacking a functional mutT gene. The mutation frequencies were calculated from the number of colonies resistant to nalidixic acid or streptomycin as described in O'Handley et al. (9).

RESULTS AND DISCUSSION
Recognition of the Putative Subfamily-Our customary procedure for identifying the activity of a new open reading frame containing the Nudix box involves BLAST searches (10) and Clustal alignments (11) of the amino acid sequence of the unknown polypeptide. When this was done for an open reading frame from A. tumefaciens, we identified a group of polypeptides with conserved sequences upstream and downstream of the Nudix box (Fig. 1). Since these regions were different from any landmark amino acids described previously (6,12), we hypothesized that the associated polypeptides represented a new subfamily of Nudix hydrolases. Accordingly, we investigated this possibility by isolating the expressed proteins and identifying the enzymatic activities of some of the members of the group.
Gene Cloning, Expression, and Protein Purification-Cloning of the genes from single colonies of the individual organisms proceeded pro forma; the inserts were sequenced and were as described in the data banks. All three of the orthologous genes from A. tumefaciens, P. aeruginosa, and C. crescentus expressed proteins in soluble form when introduced into E. coli but with different levels of expression. As with some of the other Nudix hydrolases (6, 9, 13-15), the expressed proteins were extractable into buffer merely by freezing and thawing the cells. This left the bulk of the cellular proteins behind, thereby consider-ably simplifying the purification. The procedure described under ''Methods'' led to essentially homogeneous proteins (Fig. 2).
Identification of Enzyme Activity and Substrate Specificity-A unifying property of the Nudix hydrolases is their activity on nucleoside diphosphate derivatives, although these substrates may vary widely for different members of the superfamily. Starting with the purified enzyme expressed from the A. tumefaciens gene, we examined a large number of nucleoside diphosphate derivatives as potential substrates, and we found that UTP was hydrolyzed at the highest rate of all those naturally occurring metabolites tested (Table I). However, we noted an apparent inconsistency in our substrate survey when we observed that dTTP was hydrolyzed at a measurable rate, ϳ18% of UTP, whereas dUTP was not a significant substrate. This seemed surprising since it appeared to us that UTP resembled dUTP more closely than it did dTTP, and so it called our attention to the methyl group as a potentially important feature involved in substrate recognition. Accordingly, we tested 5-methyl-UTP and found it is hydrolyzed at three times the rate of UTP. This is shown in Table I  to the purified orthologous proteins from P. aeruginosa and C. crescentus. The enzymes from P. aeruginosa and A. tumefaciens have similar specificities, but the active site of the C. crescentus enzyme is more accommodating and has approximately equal activities on UTP and the other pyrimidine ribonucleoside triphosphate, CTP.
A comparison of some of the kinetic properties of the three enzymes is shown in Table II. The combined higher k cat and lower K m for 5-methyl-UTP leads to a 10-fold higher catalytic constant, k cat /K m , over that for UTP for the A. tumefaciens and P. aeruginosa enzymes. These inequalities are not nearly as marked for C. crescentus resulting in less than a 2-fold difference in the values of the catalytic constants for the two substrates, supporting the notion that its catalytic site is more accommodating than those of the other two enzymes. In this respect, it is interesting to note that all known eukaryotic UMP kinases, enzymes involved in the phosphorylation of UMP to UDP, are equally active on UMP and CMP (16), suggesting that there may be similarities in the active sites of these kinases and the C. crescentus enzyme. Work in progress on the x-ray crystal structure of members of the family should be informative in ascertaining the basis of these observations. 1 Products of the Reaction-With one possible exception (17), virtually all of the Nudix hydrolases studied so far catalyze a nucleophilic attack by water on a pyrophosphate linkage. An analysis of the products of UTP hydrolysis with purified A. tumefaciens UTPase indicates that this enzyme also catalyzes an attack on the pyrophosphate linkage. Fig. 3A shows that after a 30-min incubation, the UTP peak decreased, a UMP peak appeared, and there was no apparent formation of UDP. This indicates that there was either a direct conversion of UTP to UMP with the release of inorganic pyrophosphate or that UTP was hydrolyzed to UDP as the rate-limiting step, and the UDP was then rapidly converted to UMP. The second pathway is ruled out for two reasons. First, as shown in Fig. 3B, no inorganic orthophos-phate is formed during the course of the reaction, as would be required for conversion of UTP to UDP, unless inorganic pyrophosphatase is present. Second, UDP itself is not a substrate of the enzyme as shown in Table I. Therefore, Reaction 1 describes the stoichiometry, Five other Nudix hydrolase nucleoside triphosphatases are known with major activities on dGTP (18), dATP (19), dCTP (15), 3Ј-amino-3Ј-dATP (20), and dTTP 2 all forming the respective monophosphates and inorganic pyrophosphate, classifying them in the category EC 3.6.1.19 with the systematic name nucleoside-triphosphate pyrophosphohydrolase. Thus, the UTPase described herein is properly designated UTP pyrophosphohydrolase, and it is the first of these enzymes preferring ribo-to deoxyribonucleotides.
Since, to our knowledge, this gene and enzyme has not been described before, we propose the acronym and mnemonic, utp as its designator.
Other Properties of the Enzyme-The A. tumefaciens UTPase requires Mg 2ϩ for activity. Unlike many of the other Nudix hydrolases, this requirement cannot be met by Mn 2ϩ , Zn 2ϩ , or Co 2ϩ . In respect to pH, however, the A. tumefaciens UTPase, like most of the other members of the superfamily, prefers alkaline conditions and is optimally active at pH 9. The protein elutes from a pre-calibrated gel filtration column as expected for an 18-kDa protein and thus probably exists as a monomer in solution.
Complementation Studies-The original Nudix hydrolase, MutT, is also a nucleoside pyrophosphohydrolase, although it has a markedly different specificity from the UTPase described here (1,18,21). Nevertheless, we tested a plasmid containing the UTPase gene for its ability to complement a mutT Ϫ strain of E. coli. Although the mutT Ϫ strain gave rise to 892 Ϯ 170 streptomycin-resistant colonies per 10 9 cells, the strain trans- TABLE II Kinetic analysis of the enzymes Standard assays were as described under "Methods" and K m and V max were determined from a non-linear regression analysis. K cat was calculated from V max , and a unit of activity represents the hydrolysis of 1 mol of substrate/min.  formed with a mutT ϩ gene produced only 10 colonies, the same number as a mutT ϩ strain. However, the mutT Ϫ strain when transformed with the UTPase gene produced 1059 Ϯ 590 colonies. Similar results were seen when nalidixic acid was substituted for streptomycin as the selective antibiotic (data not shown). These results disqualify the UTPase as an ortholog of MutT. Two issues regarding this newly discovered subfamily of Nudix hydrolases merit consideration. 1) What is the nexus linking the members of the group, or what is the commonality of the properties they share? 2) What, if any, is the physiologic relevance of their preference for ribo-TTP? In regard to the first question, this small group of 12 open reading frames (Fig. 1) was culled from over 1100 putative Nudix hydrolases identified by BLAST (10) searches of the data banks. Clustal (11) alignments revealed the ''L(VL)VRK'' motif upstream and the ''AANE'' motif downstream of the Nudix box, and these two amino acid cassettes define the signature sequence of the family and demarcate it as a distinct subfamily of proteins. The conserved "PGGK" tetrad, overlapping the Nudix box, is not diagnostic, since it is seen in many other Nudix hydrolases with dissimilar specificities. A unifying feature of these six diverse bacterial genera is their relationship to plants, either directly as symbionts (Rhizobium galegae), growth-promoting rhizobacteria living in the soil or rhizosphere (Pseudomonas fluorescens, P. aeruginosa, Pseudomonas putida, Corynebacterium efficiens), plant pathogens (A. tumefaciens, Pseudomonas syringae), or indirectly as indicated by comparative genomics. For example, virulence factors in the animal pathogens Brucella melitens and Brucella suis are related to factors involved in endosymbiosis or pathogenesis in plants, and the erythromycin-resistant plasmid containing the Nudix gene in the human pathogen Corynebacterium diphtheriae is related to the chloramphenicol resistance plasmid in the soil bacterium Corynebacterium glutamicum. The free-swimming aquatic bacterium, C. crescentus, although not found in the same ecosystem as the other members of the subfamily, derives much of its nutrients from plant polysaccharides including cellulose, lignins, xylan, glucan, and pectin and has gene clusters for their catabolism. Finally, Mycobacterium smegmatis, usually associated with humans, was originally isolated from soil.
An exegesis of the preference of the enzyme for ribo-TTP is more elusive. We are not aware of any reports of ribo-TTP in the nucleotide pool of any organism, to say nothing of the organisms comprising the subfamily, although we are also not aware of anyone looking for this nucleoside triphosphate. On the other hand, 5-methyluracil is the most common methylated base in tRNA, and therefore ribo-TMP could arise from tRNA turnover. Furthermore, it has been reported that thymidylate synthase, which catalyzes the methylation of dUMP to TMP, can also use UMP as the acceptor with the formation of ribo-TMP (22).
The formation of the corresponding triphosphate would require sequential activities of a nucleotide kinase and nucleoside diphosphokinase. Although not tested directly, studies on the specificity of purified UMP kinase (23) suggest that ribo-TMP could be phosphorylated to ribo-TDP, and the next step, the phosphorylation to ribo-TTP, would almost certainly proceed smoothly because of the exceedingly broad specificity of nucleoside diphosphokinase. Once formed, ribo-TTP could be readily incorporated into RNA since it has been demonstrated that both prokaryotic and eukaryotic RNA polymerases can substitute ribo-TTP for UTP during RNA synthesis (24,25). What effect methyluracils would have on the physiological functions of messenger and/or ribosomal RNA is conjectural. It has been shown that E. coli mutants completely lacking FIG. 3. Product formation by the A. tumefaciens UTPase. A, ion pair high pressure liquid chromatography of aliquots taken at zero time and after a 30-min incubation of a standard reaction mixture minus inorganic pyrophosphatase. B, the standard assay mixture minus pyrophosphatase was scaled up and incubated at 37°C. At the indicated times, 2 aliquots were removed. One (solid circles) was used for the determination of inorganic orthophosphate by the procedure of Fiske and SubbaRow (7). The 2nd aliquot was treated with Norit to remove nucleotide derivatives, and the supernatant was incubated with yeast inorganic pyrophosphatase for an additional 15 min after which inorganic orthophosphate was measured as above.
5-methyluracil in their tRNA show no differences in growth rate, codon recognition, protein synthesis, and macromolecular composition. However, there is a distinct advantage of wildtype cells over those lacking 5-methyluracil in mixed population studies (26).
An interesting parallel exists between the ribo-TTPase described here and Orf135, another Nudix hydrolase from E. coli (15). The latter enzyme also prefers the non-canonical methylated nucleotide, 5-methyl-dCTP instead of the natural substrate, dCTP. As pointed out earlier (1), the Nudix hydrolases may be thought of as surveillance enzymes, hydrolyzing potentially toxic or deleterious compounds arising during metabolism or preventing the accumulation of anabolic intermediates during the vicissitudes of the cell cycle. The enzymes of this newly discovered subfamily conform to this pattern. If ribo-TTP is an undesirable metabolite indigenous to these related organisms, then they have a means to deal with it. On the other hand, the role of the enzymes may be to monitor the excessive accumulation of the natural nucleoside triphosphate, UTP, and thus could be related functionally to the dGTPase (21), dAT-Pase (19), dCTPase (15), and dTTPase 2 Nudix hydrolases regulating the composition of the nucleotide pool.