Molecular Identification of Human Glutamine- and Ammonia-dependent NAD Synthetases

NAD synthetase catalyzes the final step in the biosynthesis of NAD. In the present study, we obtained cDNAs for two types of human NAD synthetase (referred as NADsyn1 and NADsyn2). Structural analysis revealed in both NADsyn1 and NADsyn2 a domain required for NAD synthesis from ammonia and in only NADsyn1 an additional carbon-nitrogen hydrolase domain shared with enzymes of the nitrilase family that cleave nitriles as well as amides to produce the corresponding acids and ammonia. Consistent with the domain structures, biochemical assays indicated (i) that both NADsyn1 and NADsyn2 have NAD synthetase activity, (ii) that NADsyn1 uses glutamine as well as ammonia as an amide donor, whereas NADsyn2 catalyzes only ammonia-dependent NAD synthesis, and (iii) that mutant NADsyn1 in which Cys-175 corresponding to the catalytic cysteine residue in nitrilases was replaced with Ser does not use glutamine. Kinetic studies suggested that glutamine and ammonia serve as physiological amide donors for NADsyn1 and NADsyn2, respectively. Both synthetases exerted catalytic activity in a multimeric form. In the mouse, NADsyn1 was seen to be abundantly expressed in the small intestine, liver, kidney, and testis but very weakly in the skeletal muscle and heart. In contrast, expression of NADsyn2 was observed in all tissues tested. Therefore, we conclude that humans have two types of NAD synthetase exhibiting different amide donor specificity and tissue distributions. The ammonia-dependent synthetase has not been found in eucaryotes until this study. Our results also indicate that the carbon-nitrogen hydrolase domain is the functional domain of NAD synthetase to make use of glutamine as an amide donor in NAD synthesis. Thus, glutamine-dependent NAD synthetase may be classified as a possible glutamine amidase in the nitrilase family. Our molecular identification of NAD synthetases may prove useful to learn more of mechanisms regulating cellular NAD metabolism.

The coenzyme NAD has a role in the majority of metabolic redox reactions and represents an essential component of metabolic pathways in all living cells. In a number of signaling pathways, NAD also serves as a precursor of potent calciummobilizing agents such as cyclic ADP-ribose and nicotinic acid adenine dinucleotide phosphate (1) and serves as a substrate for post-translational modifications of protein, mono-(2-4) and poly(ADP-ribosyl)ations (5). Depletion of cellular NAD by poly-(ADP-ribosyl)transferase activation in response to DNA damage results in cell death (6). Increased NAD synthesis has been shown to extend life span in yeast (7) and in Caenorhabditis elegans (8) via activation of an NAD-dependent histone deacetylase, silent information regulator 2 (Sir2) (9). The cellular level of NAD may modulate the sensitivity of cells to apoptotic responses through deacetylation of the p53 tumor suppressor by a human homologue of Sir2 (10). Recent publications have demonstrated that fluctuation of the NAD level in cells seems to have significant impact on their physiology. Despite these significant effects of NAD levels on cellular functions, mechanisms regulating cellular contents of NAD through metabolic events remain to be established.
NAD biosynthesis is accomplished through either de novo or salvage pathways (11,12). These two pathways converge at the level of an intermediate nicotinic acid mononucleotide (NaMN), 1 which is then converted into nicotinic acid adenine dinucleotide (NaAD) through the action of NaMN adenylyltransferase and, lastly, into NAD by NAD synthetase (Fig. 1). Although most of the genes involved in both pathways have been identified in procaryotes (13), little is known of those genes, including that of NAD synthetase in eucaryotes, except for nicotinamide mononucleotide adenylyltransferase (14) and quinolinic acid phosphoribosyltransferase (15) genes.
NAD synthetase catalyzes the conversion of NaAD into NAD, and NH 3 or glutamine is used as an amide donor in the following reactions. NaADϩNH 3 ϩATP3 NADϩAMPϩPP i REACTION 1 NaADϩglutamineϩATP3 NADϩglutamateϩAMPϩPP i REACTION 2 In procaryotes, two types of NAD synthetase have been reported, a type catalyzes Reaction 1 and is strictly ammonia-dependent, whereas the other synthetase catalyzes both Reactions 1 and 2 and uses both ammonia and glutamine as amide donors. Bacillus subtilis synthetase, a representative of the former, is a protein of 271 amino acid residues (16), and crystal * 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB091316 (NADsyn1) and AB091317 (NADsyn2).
‡ To whom correspondence should be addressed. Tel.: 81-853-20-2121; Fax: 81-853-20-2120; E-mail: mikat@shimane-med.ac.jp. structure analysis of the synthetase revealed residues responsible for binding sites for ATP and NaAD (17). Although one of the latter enzymes, Mycobacterium tuberculosis synthetase of 738 amino acids, has been also reported (18), the structural basis underlying the potential to use glutamine or ammonia of two types of synthetase has not been determined. In eucaryotes, although NAD synthetase activity of the latter type has been reported (19,20), molecular cloning and characterization have not been done. In the case of the strictly ammonia-dependent NAD synthetase, the counterpart of B. subtilis enzyme, even the presence of the activity has not been demonstrated in eucaryotic organisms.
We now report molecular identification of two human NAD synthetases, a synthetase that can use not only ammonia but also glutamine and the other synthetase with strictly ammoniadependent activity (referred as NADsyn1 and NADsyn2, respectively). To our knowledge, this is the first report demonstrating the presence of the strictly ammonia-dependent NAD synthetase in eucaryotes. We also describe the structural basis underlying the potential of NADsyn1 to use glutamine as an amide donor as well as the distinct distribution of NADsyn1 and NADsyn2 in animal tissues.

EXPERIMENTAL PROCEDURES
Materials-[␣-32 P]dCTP (6000 Ci/mmol) was purchased from Amersham Biosciences. NaAD, AMP and inorganic pyrophosphatase were from Sigma. ATP, NAD, L-glutamine, and ammonium chloride were from Oriental Yeast (Tokyo, Japan), Roche Molecular Biochemicals (Basel, Switzerland), Nacalai Tesque (Kyoto, Japan), and Wako Pure Chemical Industries (Osaka, Japan), respectively. COS-7 cells and a human promyelocytic leukemia cell line HL60 were obtained from Riken Cell Bank (Tsukuba Science City, Japan). Human glioma cell line LN229 and human hepatocyte cell lines HepG2 and Huh7 were from American Type Culture Collection (Manassas, VA).
Expression of NADsyn1 and NADsyn2 in COS-7 Cells-To express NADsyn1 and NADsyn2 as C-terminal-His 6 -tagged proteins in COS-7 cells, a His 6 tag sequence followed by a TGA termination codon was introduced into the pcDNA3 vector (Invitrogen) between XbaI and ApaI cloning sites to obtain the pcDNA3His 6 vector. Segments of human NADsyn1 cDNA were PCR-amplified from fetal human brain cDNA (Clontech, Palo Alto, CA) using two sets of primers, 5Ј-ATG GGC CGG AAG GTG ACC-3Ј (sense) and 5Ј-CAG ACC TGG CAG CAC ATG-3Ј (antisense) and 5Ј-AAG CCT TGG ACC TGC CTG-3Ј (sense) and 5Ј-GAA GGA ACC GGC CTC A-3Ј (antisense). The PCR products were gel-purified, combined, and amplified using primers (the underlined regions correspond to cloning sites) 5Ј-AAG CTT GGT ACC ATG GGC CGG AAG GTG ACC-3Ј (sense) and 5Ј-AAG CTT TCT AGA GTC CAC GCC GTC CAG GGA-3Ј (antisense), yielding full-length NADsyn1 cDNA. Human NADsyn2 cDNA was amplified from LN229 total RNA treated with deoxyribonuclease I (Nippon Gene, Tokyo, Japan) by reverse transcription-PCR using primers 5Ј-AAG CTT GGA TCC ATG CAA GCC GTA CAG CGC-3Ј (sense) and 5Ј-AAG CTT TCT AGA GGG TGC AAA CGG CAT CAC-3Ј (antisense). Amplified cDNAs were digested with KpnI and XbaI for NADsyn1 and BamHI and XbaI for NADsyn2 and ligated, respectively, into KpnI-and XbaI-digested and BamHI-and XbaI-digested pcDNA3His 6 . The vector used to express the mutant NADsyn1, in which Cys-175 was replaced with serine (C175S-NADsyn1), was made using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) using the pcDNA3His 6 plasmid vector carrying wild-type NADsyn1 and the oligonucleotide primers 5Ј-TT GGA AGT GAG ATC agt GAG GAG CTC TGG-3Ј and 5Ј-CCA GAG CTC CTC act GAT CTC ACT TCC AA-3Ј, where the altered codon is indicated in lowercase italics. The resultant expression plasmids, purified using a Qiagen plasmid kit (Hilden, Germany), were transfected into COS-7 cells (4.5-6.5 ϫ 10 5 cells/100-mm dish) using an activateddendrimer PolyFect Transfection Reagent (Qiagen) according to the manufacturer's instructions. Forty-eight hours after transfection, the COS-7 cells were washed twice with phosphate-buffered saline and collected by scraping. After the cells were lysed by sonication and centrifugation, recombinant NAD synthetases were purified from the supernatants with His-Bind Resin (Novagen, Madison, WI) according to the manufacturer's protocol. A cDNA fragment of mouse glutamine-dependent NAD synthetase was amplified from Balb/c mouse kidney total RNA treated with deoxyribonuclease I by reverse transcription-PCR using primers based on the sequence of mouse homologue of human NADsyn1 (see Fig. 7),5Ј-ATG GGC CGG AAA GTG ACC-3Ј (sense) and 5Ј-CAG ACC AGG CAG CAC ATG-3Ј (antisense). Sequences of expression plasmids or PCR fragments were confirmed by entire sequencing in both directions.
5Ј-Rapid Amplification of cDNA Ends-Adaptor-ligated double strand cDNA was synthesized using Marathon cDNA amplification kit (Clontech) from poly(A) ϩ RNA isolated from HL60 cells with QuickPrep micro mRNA purification kit (Amersham Biosciences). The 5Ј parts of the NADsyn1 and NADsyn2 cDNAs were amplified with Advantage 2 polymerase mix (Clontech) using sense adaptor primer 5Ј-CCA TCC TAA TAC GAC TCA CTA TAG GGC-3Ј and gene-specific antisense primers 5Ј-CAG CGC AGC TCG CGG TAG TTG CCT TC-3Ј (NADsyn1) and 5Ј-CCC AGC ACG AAG TCC CGG GAG ACT GC-3Ј (NADsyn2). After nested PCR, the products were subcloned into pcDNA3, and positive clones were isolated and sequenced.
Enzyme Assays-Unless otherwise stated, NAD synthetase activity of the recombinant protein was based on fluorometric measurements of the NAD formed, as described below. Recombinant NADsyn1, the mutant NADsyn1, and NADsyn2 were incubated with glutamine or NH 4 Cl as indicated in the reaction mixture (50 l) containing 50 mM HEPES (pH 8.8), 2 mM ATP, 1 mM NaAD, 56 mM KCl, 5 mM MgCl 2 , and 10 g of bovine serum albumin. For the assay of NADsyn1 activity, 2 mM dithiothreitol was included in the reaction mixture. When glutamine was used as a substrate for NADsyn1, 50 mM Tris-Cl Ϫ (pH 7.5) was included instead of 50 mM HEPES (pH 8.8). The reactions were terminated by adding 0.4 ml of 7 N NaOH and then incubated at 37°C for 30 min to obtain the fluorescent product (21). The fluorescence was measured using 380 nm for excitation and 460 nm for emission by Fluoroskan Ascent FL (Labsystems, Helsinki, Finland). The fluorescence intensity of standard NAD solutions at known concentrations was used to calculate the amount of NAD. NAD synthetase activity was calculated by subtracting the NAD content of enzyme-deficient blanks from the NAD content of the complete reaction mixture.
In some cases NAD synthetase activity was determined by HPLC analysis. After the NAD synthetase reactions had been terminated by a 10-fold dilution with 0.1% trifluoroacetic acid, the reaction products (NAD, NaAD, and AMP) were separated on a reversed phase Cosmosil 5C-18MS column (4.6 ϫ 150 mm, Nacalai Tesque) with 0.1% trifluoroacetic acid as the mobile phase and detected by measuring absorbance at 254 nm.
Kinetic parameters for NAD synthetase reaction were determined as follows by analysis of a Lineweaver-Burk plot of the initial rates of NAD synthesis. In the reaction mixture at fixed concentrations of two of the three substrates concentrations of the third substrate varied from 0.1 to 2 mM for NaAD, from 0.05 to 1 mM for ATP, from 1 to 40 mM for glutamine, and from 1.5 to 30 mM (NADsyn1 and C175S-NADsyn1) or from 0.01 to 0.2 mM (NADsyn2) for NH 4 Cl. In the reaction catalyzed by NADsyn1, K m values for NaAD and ATP were determined using glutamine as an amide donor, whereas in reactions catalyzed by NADsyn2 and C175S-NADsyn1, values were determined using NH 4 Cl. Amounts of NAD formed were determined by the fluorometric method.
Determination of PP i -NAD synthetase reactions (50 l) were terminated by adding 8 l of 10% trifluoroacetic acid. After standing on ice for 15 min and then neutralizing by 1 M Tris-Cl Ϫ (pH 9.0), the reaction mixtures were incubated with or without 2.5 milliunits/l of pyrophosphatase at 25°C for 30 min. The reactions were terminated by 10% perchloric acid, and the inorganic phosphate formed was determined as described by Yu and Dietrich (19).
Determination of Molecular Masses of Catalytically Active NAD Synthetases-Purified recombinant NADsyn1 and NADsyn2 were electrophoresed on 7% and 12.5% non-denaturing polyacrylamide gels, respectively, in Tris-glycine buffer (pH 8.3) (25 mM Tris, 192 mM glycine) in the presence of 0.2% 2-mercaptoethanol. Gels were sliced into 2-3-mm pieces then incubated in the reaction mixture (100 l) containing either 20 mM glutamine for NADsyn1 or 1 mM NH 4 Cl for NADsyn2 at 37°C for 2 h. NAD formed in the reaction mixture was determined using the fluorometric method, as described above.
Northern Blot Analysis-Total RNAs (20 g) prepared from Balb/c mouse tissues and human cell lines were fractionated on a 1% agaroseformaldehyde gel, transferred to Hybond-Nϩ nylon membrane (Amersham Biosciences), and UV-cross-linked. The blot was prehybridized at 43°C for 5 h in hybridization solution containing 5ϫ SSPE (1ϫ SSPE: 0.15 M NaCl, 8.65 mM sodium dihydrogen phosphate, 1.25 mM EDTA), 50% formamide, 5ϫ Denhardt's solution (1ϫ Denhardt's solution: 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin), 0.5% SDS, and 100 g/ml heat-denatured salmon sperm DNA. The blots were then hybridized at 43°C for 18 h in the same solution containing heat-denatured cDNA probes (corresponding to amino acids 1-377 of mouse glutamine-dependent NAD synthetase in Fig. 7 or the entire coding region of human NADsyn2 cDNA) labeled by PCR with [␣-32 P]dCTP and an antisense primer. The membranes were once washed in 2 ϫ SSC (1ϫ SSC: 0.15 M NaCl, 0.015 M sodium citrate), 0.1% SDS at 25°C for 15 min, then twice with 0.1ϫ SSC, 0.1% SDS at 65°C for hybridization with the mouse glutamine-dependent NAD synthetase cDNA or 0.5ϫ SSC, 0.1% SDS at 50°C for with NADsyn2 cDNA and exposed to x-ray film at Ϫ80°C with an intensifying screen. Under these conditions, NADsyn2 probe did not detect the 3.1-kb message (see Fig. 6B). RNA integrity and loading were assessed using a glyceraldehyde-3-phosphate dehydrogenase probe.

RESULTS
cDNA Cloning and Structural Analysis of Human NAD Synthetases-Using the deduced amino acid sequence of B. subtilis NAD synthetase (16) (SWISS-PROT accession number P08164) as a probe, we found two candidate sequences encoding human NAD synthetase, GenBank TM accession numbers AK001493 and HSA236685 in a homology search analysis. Using primers corresponding to the presumed 5Ј-and 3Ј-terminal sequences of the candidates, we carried out reverse transcription-PCR as described under "Experimental Procedures" and obtained two human cDNA clones NADsyn1 and NADsyn2, encoding proteins of 706 and 275 amino acids, respectively. To ensure the correct prediction of the open reading frames, we amplified sequences containing 5Ј upstream regions of the cDNAs using 5Ј-rapid amplification of cDNA ends. In the 5Ј-rapid amplification of cDNA ends products, we confirmed sequences corresponding to the sense primers used for amplification of the cDNAs. In 5Ј upstream region of NADsyn1 cDNA (92 bp) we amplified we did not observe in-frame stop codon but found that the presumed start methionine codon was in a favorable Kozak initiation sequence AGGATGG (22), and the codon was the first in-frame methionine codon. For NADsyn2, we obtained an in-frame stop codon 6 bp upstream of the presumed initiation codon. Thus, we concluded that NADsyn1 and NADsyn2 encode complete coding regions of the respective proteins.
These observations suggest that cDNAs for NADsyn1 and NADsyn2 encode NAD synthetases. We designated the NAD_synthase domain and the P-loop motif as the synthetase domain.
In addition to the synthetase domain, NADsyn1 only had the carbon-nitrogen hydrolase (CN_hydrolase) domain (Pfam accession number PF00795) at the N-terminal half (Fig. 2A). The CN_hydrolase domain, with a cysteine residue essential for nitrilase activity (24,25), is shared with enzymes belonging to the nitrilase family that cleave nitriles as well as amides to produce the corresponding acids and ammonia (Fig. 2C) (26,27). Because B. subtilis synthetase, which lacks the CN_hydrolase domain ( Fig. 2A), utilizes ammonia but not glutamine (16), the presence of the CN_hydrolase domain in NADsyn1 but not in NADsyn2 suggested that the former enzyme uses both glutamine and ammonia in NAD synthesis, whereas the latter is strictly ammonia-dependent. Furthermore, because the critical cysteine residue in the CN_hydrolase domain was also conserved in NADsyn1 at a position of 175 (Fig. 2C), the cysteine (Cys-175) in NADsyn1 was considered to be essential for the use of glutamine. From these sequence analyses, we speculated (i) that both NADsyn1 and NDsyn2 have NAD synthetase activity, (ii) that NADsyn1 uses not only ammonia but also glutamine, whereas NADsyn2 uses only ammonia, and (iii) that the site-directed mutagenesis of Cys-175 in NADsyn1 eliminates glutamine-dependent NAD synthetase activity with the ammonia-dependent activity intact.
Expression and Functional Characterization of Human NAD Synthetases-We next expressed NADsyn1, NADsyn2, and a mutant NADsyn1 in which Cys-175 was replaced with serine (C175S-NADsyn1) in COS-7 cells as His 6 -tagged recombinant proteins, and we purified these proteins on nickel chelate resin. SDS-PAGE analysis indicated that the purified wild-type (Fig.  3A, inset) and mutant NADsyn1 (data not shown) have a molecular mass of 80 kDa, in accordance with the value calculated from the deduced sequences, 80.3 kDa. The purified recombinant NADsyn2 appeared as a single band with a molecular mass of 34 kDa (Fig. 3B, inset), slightly larger than the value calculated from the deduced sequence, 30.8 kDa. The difference may depend on the low pI of the recombinant protein (pI 5.9).
To investigate whether NADsyn1, NADsyn2, and C175S-NADsyn1 have the predicted enzymatic activities, we incubated the purified recombinant proteins with glutamine or NH 4 Cl in the presence of NaAD and ATP then determined the NAD synthetase activities of the proteins using a fluorometric method. As shown in Fig. 4A, the recombinant wild-type NADsyn1 exhibited almost the same activity with either glutamine or NH 4 Cl. On the other hand, the recombinant NADsyn2 catalyzed NAD synthesis primarily with NH 4 Cl (Fig.  4C). In marked contrast with the wild-type NADsyn1, the activity of the mutant NADsyn1 (C175S-NADsyn1) was not detected when glutamine was used as a substrate, whereas the activity remained unaltered with NH 4 Cl (Fig. 4B). Kinetic analysis indicated that NADsyn1 shows a lower K m for glutamine (1.44 mM) than for NH 4 Cl (13.1 mM) (Table I). Compared with NADsyn1, NADsyn2 showed a much lower K m for NH 4 Cl (34 M) but a much higher K m for glutamine (103 mM) (Table I). NADsyn1 and NADsyn2 showed essentially the same K m values for ATP and NaAD, in the range of those reported for native synthetases (16,19). Kinetic parameters of the mutant NADsyn1 obtained with NH 4 Cl did not differ from those of the wild-type NADsyn1, which suggests that disappearance of the glutamine dependence was not because of a drastic change in the tertiary structure of NADsyn1. Compared with NADsyn2 (34 M), the mutant synthetase exhibited a much higher K m value for NH 4 Cl (23.9 mM). With the glutamine preparation, which contained up to 0.38% ammonia in itself, the concentration of ammonia in the assay solution in Fig. 4 would be up to 76 M. Thus, under these conditions NADsyn2 but not the mutant synthetase could synthesize NAD without exogenously added ammonia (Fig. 4), as was seen with the B. subtilis synthetase (16).
All these results are consistent with our predictions (i) that both NADsyn1 and NADsyn2 have NAD synthetase activity, (ii) that NADsyn1 utilizes not only ammonia but also glutamine as an amide donor, whereas NADsyn2 is primarily an ammonia-dependent NAD synthetase, and (iii) that Cys-175 in NADsyn1 is essential for the ability to use glutamine as an amide donor. In agreement with our sequence analyses noted above, we therefore conclude that the CN_hydrolase domain in the N-terminal half of NADsyn1, in particular Cys-175, is responsible for utilization of glutamine as an amide donor and, thus, confers glutamine dependence on the synthetase, whereas the synthetase domain in NADsyn1 and NADsyn2 participates solely in NAD synthesis from ammonia. To further confirm the role of the CN_hydrolase domain in using glutamine, we made a chimera consisting of the N-terminal region of NADsyn1 and NADsyn2 as well as a construct containing solely the N-or C-terminal half of NADsyn1. However, because of their insufficient expressions, we could not characterize these constructs (data not shown).
We next examined the stoichiometry of the reactions catalyzed by the recombinant synthetases. The purified recombinant NADsyn1 and NADsyn2 were incubated with glutamine and NH 4 Cl, respectively, in the presence of NaAD and ATP and the reaction products were analyzed by reversed phase HPLC. As shown in Fig. 5, 1 nmol of AMP and PP i was produced per 1 nmol of NAD synthesized during the reaction catalyzed by each synthetase. These results indicate that amidation of NaAD by the recombinant enzymes is associated with ATP cleavage to AMP and PP i , as noted for the native enzyme (28). Omission of either ATP, Mg 2ϩ , NaAD, or amide donors from the reaction mixture resulted in a complete loss of NAD synthesis by each enzyme (data not shown).
To examine whether catalytically active forms of human NADsyn1 and NADsyn2 are multimers as native synthetases (16,19,20), we fractionated the purified recombinant synthetases by non-denaturing PAGE and determined NAD synthetase activity in gel slices. As shown in Fig. 3, activities of NADsyn1 and NADsyn2 had mobilities consistent with proteins of 500 and 70 kDa, respectively, suggesting that NADsyn1 and NADsyn2 may exist as a homohexamer and a homodimer, respectively.
Tissue Distribution of NAD Synthetases-To evaluate the tissue distribution of NADsyn1 and NADsyn2, Northern blot a NAD synthetase activity was below the limit of detection, and thus, the K m value was not determined. analyses were done with total RNA from various mouse tissues and several human cell lines. As shown in Fig. 6A, a message of 3.1 kb was detected for NADsyn1, with various intensities. The major sites of NADsyn1 gene expression were the small intestine, kidney, liver, and testis, whereas the skeletal muscle, spleen, lung, heart, and brain showed a weak signal. In the liver and small intestine, an additional signal was observed at 2.1 kb. The NADsyn1 gene was also expressed in human glioma (LN229) and promyelocytic leukemia (HL60) cell lines (data not shown). Although an EST data base homology search did not reveal a human clone with a high degree of similarity to NADsyn2, Northern blot analysis clearly showed that an mRNA species of 1.4 kb is expressed in human cells LN229, HL60, and HepG2 and Huh7 (hepatocyte cell lines) (Fig. 6B). In the mouse, all of the tissues tested expressed the 1.4-kb mRNA, probably representing a mouse homologue of NADsyn2. In the lung and in skeletal muscle, the same 1.4-kb signal was observed after long exposure of the blot (data not shown). In mouse brain and kidney and human cell lines an additional 2.6-kb mRNA species was observed.
Sequence Alignment of NADsyn1 Homologues-The deduced amino acid sequences of NADsyn1 and NADsyn2 showed mismatches of three and four amino acids, respectively, compared with those deposited in GenBank TM data base (AK001493 and HSA236685, respectively) (Figs. 7 and 2B). A homology search in a protein data base revealed that NADsyn1 exhibits significant amino acid identity to several eucaryotic putative proteins from the mouse (83%), Saccharomyces cerevisiae (58%), Drosophila melanogaster (53%) and C. elegans (46%) (Fig. 7).
Alignment of the NADsyn1 sequence with these proteins showed that there are highly conserved regions among the five proteins, including ATP and NaAD binding sites in the Cterminal regions and the essential cysteine residues for use of glutamine in the N-terminal halves. The CN_hydrolase domain of NADsyn1 also showed sequence similarity to NAD synthetases from M. tuberculosis (18) and Rhodobacter capsulatus (29) (Fig. 2C), and in particular, the critical cysteine residues were strictly conserved in the two bacterial synthetases. NADsyn2 did not exhibit significant sequence similarity to any other eucaryotic proteins over the entire length. DISCUSSION The present study is the first identification of open reading frames encoding two NAD synthetases, NADsyn1 and NADsyn2, in humans. Heterologous expression of the synthetases indicated that although NADsyn1 utilizes both glutamine and ammonia as amide donors, ammonia may not serve as a physiological amide donor for NADsyn1 in vivo (K m for NH 4 Cl Ͼ10 mM). In marked contrast, NADsyn2 uses ammonia more efficiently than does NADsyn1 (K m for NH 4 Cl ϭ 34 M) but appears to be unable to use glutamine as a physiological amide donor (K m for glutamine Ͼ100 mM). Thus, we conclude that NADsyn1 is a glutamine-dependent NAD synthetase, whereas NADsyn2 is a strictly ammoniadependent synthetase. To our knowledge, this is the first evidence for the presence of an ammonia-dependent NAD synthetase in eucaryotes. Furthermore, by comparing the catalytic activity of NADsyn2, which lacks CN_hydrolase Twenty g of total RNA prepared from the indicated mouse tissues and human cell lines were analyzed using Northern blot hybridization with radiolabeled mouse glutamine-dependent NAD synthetase cDNA (A, top) or human NADsyn2 cDNA (B, top). After washing, the blot was exposed to x-ray film then stripped and reprobed with a radiolabeled glyceraldehyde-3-phosphate dehydrogenase cDNA (bottom) to control for loading. Positions of 28 S and 18 S ribosome RNA are indicated by arrows.
domain, with the mutant NADsyn1, in which Cys-175 corresponding to the catalytic cysteine residue in nitrilases (24,25) was replaced with Ser, we identified the CN_hydrolase domain as the functional domain of NAD synthetase to ab-stract nitrogen from the amide of glutamine and, thus, to use glutamine as an amide donor.
Identification of the CN_hydrolase domain as the determinant of glutamine dependence means that the homologues of human NADsyn1 found in different species are also glutaminedependent NAD synthetases. The domain occurring in the NAD synthetase from M. tuberculosis, known to catalyze NAD synthesis with glutamine (18), now provides the previously unrecognized structural basis underlying the glutamine dependence of the synthetase. These results suggest that the glutamine-dependent NAD synthetases can be classified as a possible glutamine amidase into the nitrilase family. In the nitrilases, an invariant cysteine residue has been proposed to act as a nucleophile in the catalytic mechanism, where a nitrile carbon is subjected to a nucleophilic attack by sulfhydryl group in the active site of the enzyme (24,25). Because the critical cysteine residue in the nitrilases is conserved in these glutaminedependent NAD synthetases, the cysteine residues in the synthetases probably carry out a nucleophilic attack on a carbonyl carbon of glutamine, abstracting ammonia from glutamine. In the synthetase domains of these synthetases, after adenylation of NaAD in the presence of ATP, the ammonia thus abstracted in the CN_hydrolase domain attacks the adenylated NaAD, resulting in NAD, as proposed (28). X-ray diffraction analysis and a detailed structure-functional analysis of NADsyn1 will give a better understanding of the catalytic mechanism of these glutamine-dependent NAD synthetases.
We showed that NADsyn1 exerts catalytic activity in a multimeric form. NAD synthetases purified from human erythrocytes and yeast, glutamine-dependent and, thus, expected to possess CN_hydrolase domain, are also multimeric enzymes (19,20). It has been reported that nitrilase family members form a multimer, probably by subunit contact through highly hydrophobic regions conserved in the CN_hydrolase domain (26) (Fig. 2C). Thus, it appears that the CN_hydrolase domain, including hydrophobic regions, participates in multimer formation of NADsyn1.
Glutamine-dependent yeast NAD synthetase has been reported to have two components, an 80-kDa ammonia-dependent NAD synthetase subunit and an additional 65-kDa subunit, and the latter has been hypothesized to use glutamine as amide donor (19,30). However, because the yeast homologue of human NADsyn1 with a calculated molecular mass of 80.7 kDa has the CN_hydrolase domain, it seems that the 80-kDa subunit solely represents the yeast synthetase, and it is unlikely that the 65-kDa subunit is required for glutamine-dependent NAD synthetase activity.
The wide variability of NADsyn1 expression revealed by Northern blot analysis may reflect differences in NAD demand among animal tissues. Abundant expression of NADsyn1 was observed in the small intestine, liver, kidney, and testis, whereas skeletal muscle and the heart showed very weak signals. However, nicotinamide mononucleotide adenylyltransferase, catalyzing the formation of the substrate of NAD synthetase NaAD (Fig. 1), has been reported to be expressed mainly in skeletal muscle and the heart (14), thus being inconsistent with NADsyn1 expression. This raises the question of how NAD synthesis occurs in these tissues. NADsyn2 is expressed in skeletal muscle and in the heart. In these tissues, gene expression of glutaminase catalyzing the formation of ammonia from glutamine has also been demonstrated (31). Taken together the finding that NADsyn2 could catalyze NAD synthesis using ammonia as an amide donor, NADsyn2 may largely mediate NAD synthesis in these tissues. Alternatively, based on a somewhat higher affinity of nicotinamide mononucleotide adenylyltransferase for NMN than for NaMN (14), NAD may also be synthesized via direct conversion of NMN to NAD in the tissues. For a better understanding of the regulation of NAD biosynthesis in higher organisms, including humans, further analyses on quinolinic acid phosphoribosyltransferase and nicotinic acid phosphoribosyltransferase expression in animal tissues are under investigation.
In the present study, we identified glutamine-and ammoniadependent human NAD synthetases, NADsyn1 and NADsyn2, respectively, with distinct tissue distribution of the synthetases, and we obtained evidence that the CN_hydrolase domain confers glutamine dependence on the former enzyme. Our results suggest that the glutamine-dependent NAD synthetase is classified as a glutamine amidase into the nitrilase family and that the newly identified metabolic pathway involving the ammonia-dependent NAD synthetase plays a role in NAD biosynthesis. These are important clues to better understand detailed structures fundamental to catalytic activity of the enzyme and to elucidate regulatory mechanisms of cellular NAD metabolism.