Vol. 136, No. 1
JOURNAL OF BACTERIOLOGY, Oct. 1978, P. 136-141 0021-9193/78/0136-0136$02.00/0 Copyright i) 1978 American Society for Microbiology
Printed in U.S.A.
Evidence for an Intermediate in Quinolinate Biosynthesis in Escherichia colit FRANK D. WICKS, SHIGEKI SAKAKIBARA,4 AND R. K. GHOLSON* Department ofBiochemistry, Oklahoma State University, Stillwater, Oklahoma 74074 Received for publication 8 May 1978
Evidence for the formation of an unstable intermediate in the synthesis of quinolinate from aspartate and dihydroxyacetone phosphate by Escherichia coli was obtained using toluenized cells of nadA and nadB mutants of this organism and partially purified A and B proteins in dialysis and membrane cone experiments. The results of these experiments indicate that the nadB gene product forms an unstable compound from aspartate in the presence of flavine adenine dinucleotide, and that this compound is then condensed with dihydroxyacetone phosphate to form quinolinate in a reaction catalyzed by the nadA gene product.
Quinolinic acid (pyridine 2,3-dicarboxylic acid) (QA) is a precursor of the pyridine nucleotides in mammals (8), plants (10), and microorganisms (2). The pathway for the synthesis of QA from tryptophan in mammalian liver and in Neurospora has been clearly elucidated by the efforts of a number of investigators (4). This tryptophan-QA pathway is also present in aerobically grown yeast (1). On the other hand, the details of QA synthesis in plants and most microorganisms have not been reported. Radioactive labeling studies in vivo and in vitro have suggested at least three separate pathways other than the tryptophan pathway for QA synthesis in various microorganisms (15, 17; T. A. Scott and H. Hussey, Biochem. J. 96:9c). In the most thoroughly studied system which is present in aerobically grown Escherichia coli, two protein fractions have been shown to be required for the synthesis of QA from aspartate and dihydroxyacetone phosphate (DHAP) (9,18). One of these, the "B" protein, is coded for by the nadB gene (21) and the other, the "A" protein, by the nadA gene (19). Although these proteins have been partially purified, and the properties of the in vitro QA synthetase system have been studied in some detail, previous efforts to detect the formation of intermediate(s) between aspartate and DHAP and QA in this system were unsuccessful (9, 18). On the other hand, a series of experiments have been reported by other workers which suggest that several intermediates in QA biosynthesis may be formed from aspartate and DHAP in E. coli. Cross-feeding experiments have been reported (12) which indicate that the t Journal article J3462 of the Agricultural Experiment Station, Oklahoma State University, Stillwater, OK 74074. t Present address: Department of Medical Chemiatry, Osaka Medical College, Osaka, Japan.
13E
nadA-coded enzyme produces a compound which can be converted to QA by a mutant containing an intact nadB-coded enzyme. Subsequently, a nadC mutant, which is blocked in the conversion of QA to nicotinic acid mononucleotide (20), was shown to excrete a compound which was not QA, but which stimulated the growth of both nadA and nadB mutants (5). More recently, Chen and Tritz (6) have concluded that five unidentified compounds formed from fructose-1,6-diphosphate and/or aspartate in a cell-free system prepared from a nadC mutant are precursors of QA, since formation of these compounds is repressed by nicotinic acid and inhibited by oxidized nicotinamide adenine dinucleotide. This communication reports evidence that the nadB gene product, in the presence of flavine adenine dinucleotide (FAD), converts aspartate into an unstable compound which is then condensed with DHAP to forn QA in a reaction catalyzed by the nadA gene product. MATERIALS AND METHODS Bacterial strains. All bacteria used were derivatives of E. coli K-12. Strain UTH 4460 contains nadA19 and was originally obtained from R. A. LaValle as his PA-2-18. Strain UTH 4451 contains nadB35 and was originally obtained from R. A. LaValle as P4X-SB16. Strain UTH 4464 contains nadB7 and was originally obtained from T. K. Sundarum as W-3899. These strains are referred to in this paper by their nad allele numbers, which are Coli Genetic Stock Center (Yale University) designations. Growth and harvesting of cells. The minimal medium of Yates and Pardee (23), supplemented with 0.5 ,M nicotinic acid and other necessary growth factors, was used. Cultures were grown in 10-liter tanks in a New Brunswick fermentor to late log phase and were harvested in a Sharples centrifuge. The cell paste
QUINOLINATE BIOSYNTHESIS IN E. COLI
VOL. 136, 1978
washed once with 0.9% saline. Cells from which A and B proteins were isolated were stored at -20°C until used. Preparation of A and B proteins. Frozen cells of nadAI9 and nad35 were suspended in 20 volumes of 0.05 M bicine buffer (pH 8.0), sonically disrupted, and centrifuged as previously described (9). The A protein was purified from the crude extract of nadB35 cells through the 40 to 60% ammonium sulfate step and the B protein was purified through the Sephadex G-200 step, as described by Griffith et al. (9). The proteinreplacing activity of liver, which was used as the B protein source in some experiments, was purified through the Sephadex G-75 step as described by Sakakibara et al. (16). Preparation of toluenized celis. Toluenized cells of nadA19 and nadB7 were prepared following the methods described by Jackson and DeMoss (11). A 0.2-g sample of washed cells was suspended in 1 ml of 0.05 M potassium phosphate buffer (pH 8.0); 100 pl of toluene was added, and the suspension was mixed vigorously at room temperature for 1 min. Toluenized cell suspensions were stored in an ice bath until used (no longer than 30 min). Other Methods. Assay for QA formation during A and B protein purification was carried out as previously described (9). Assay for QA formation in those experiments described here used these methods (9), with some modifications as described in the table legends. Protein was determined by the method of Lowry et al. (13) in crude extracts and spectrophotometrically by the 280/260 absorbance ratio (22) in more purified preparations. The filter cone experiments were carried out using Amicon CF 25 Centriflo membrane cones obtained from Amicon Corp., Lexington, Mass. Various components of the quinolinate synthetase system, in a total volume of 1.0 ml, were placed in a membrane cone, which was inserted in a 50-ml centrifuge tube containing the remaining system components not present in the cone. The tubes were then centrifuged at 3,000 rpm in a SS-34 rotor of a model RC2-B centrifuge (about 1,000 x g) at room temperature for 15 min. Centrifugation forced 80 to 95+% of the liquid contents of the membrane cone into the centrifuge tube. The centrifuge tubes were then incubated at 25°C for a further 15 min. The reaction was stopped, and [14C]quinolinate was determined as previously described (18). The exact contents of membrane cones and tubes in various experiments are given in Table 5. Chemicals. L-[U-14C]aspartate (160 mCi/mmol) was obtained from New England Nuclear. All other chemicals were of reagent grade, obtained from commercial supply houses.
was
RESULTS
Toluenized cela. The first evidence for the formation of a free intermediate in QA biosynthesis in E. coli obtained in our laboratory was provided by experiments carried out with toluenized cells. Treatment of E. coli with low concentrations of toluene causes the cell membranes to become freely perneable to low-mo-
137
lecular-weight metabolites, while macromolecules are retained within the cells (11). When toluenized nadA (PA-2-18) cells were incubated with all the components of the in vitro QA synthetase system, no detectable [14C]QA was formed (Table 1). When toluenized nadB (W3899) cells were incubated under these same conditions, 1,146 cpm of ['4C]QA was fonned. This result is consistent with our findings that many nadB mutants are slightly leaky and contain low but detectable amounts of B protein (9). However, when toluenized nadA and nadB cells were mixed together and incubated in the reaction mixture (Table 1, experiment 3), five times more [14C]QA was formed. This suggested that a compound formed from the substrate(s) by one mutant is being converted to QA by the other mutant cell type or that the A and B proteins are leaking from the toluenized cells and are catalyzing the extracellular fornation of QA from the substrates. The possibility of leakage was tested by using the supernatant solutions remaining after centrifugation of the 20% suspensions of toleuized cells rather than the cells themselves. Although some [14C]QA was formed under these conditions (experiments 4 and 5, Table 1), this was significantly less than observed when both types of cells were present at the same time, making it unlikely that the results observed in experiment 3 were due to protein leakage. Since the toluenized cells could be rapidly removed from the reaction by centrifugation, it TABLE 1. Quinolinate synthesis in toluenized nadA and nadB cellsa Expt Incubation
Enot Incubat1on 1
no.
. 2 Incubation
Total
dpm of [4C]QA
nadA cells 0 nadB cells 1,146 nadA cells + nadB cells 6,516 nadA cella + superna294 tant from nadB cells 5 nadB cells + superna2,148 tant from nadA cells 6 nadB cells nadA cells + superna1,596 tant from incubation 1 7 nadA cells nadB cells + superna7,248 tant from incubation 1 a Incubations were carried out at 25°C in a total volume of 0.5 ml containing L-[U-"CJaspartate, 250 nmol (IluCi); DHAP, 100 nmol; FAD, 10 gtmol; and bicine buffer (pH 8.0), 50 ,umol for 15 min. A 0.1-ml sample of a 20% suspension of toluenized cells from a nadA or nadB mutant was used as indicated. In experiments 4 and 5, 0.1 ml of the supernatant from centrifuged suspensions of toluenized cells was added as control for protein leakage. In experiments 6 and 7, 0.1 ml of toluenized cells was incubated with the complete reaction mixture for 15 min, the cells were removed by centrifugation, and the supernatant was added to 0.1 ml of a 20% suspension of toluenized nadA or nadB cells and incubated for an additional 15 min. 1 2 3 4
-
138
J. BACTERIOL.
WICKS, SAKAKIBARA, AND GHOLSON
was possible to test which mutant cell type was forming the intermediate by incubating each cell type in a complete reaction medium for 15 min, centrifuging out the cells, and transferring the medium to another tube containing the other cell type for an additional 15-min period. The results of experiments 6 and 7 (Table 1) show that much more ["4C]QA was formed when the incubation medium from nadA cells was transferred to nadB cells for further incubation than when the reverse sequence was followed. These results are consistent with the formation of a diffusible intermediate by the nadB gene product (B protein), which is then converted to QA by the nadA gene product (A protein). These experiments with toluenized cells could be subject to error because of the leakiness of the nadB mutation and the possibility of the incomplete removal of cells by centrifugation. Experiments with partially purified A and B proteins detailed below, however, suggest these same conclusions. Dialysis experiments. Further support for the occurrence of a diffusible intermediate in QA biosynthesis was provided by experiments in which partially purified A and B proteins were incubated on opposite sides of a dialysis membrane. In these experiments B protein prepared from calf liver was used, since enzyme from this source appeared to have consistently higher activity than E. coli B protein (16). All substrate components of the quinolinate synthetase system were placed in both cells of a dialysis apparatus with protein A on one side of the membrane and protein B on the other side. After incubation for 30- and 60-min intervals, the QA content on both sides of the membrane was determined. ['4C]QA was found to be present on both sides of the membrane but with larger amounts in the chamber containing A protein (Table 2). In control experiments with the same enzyme preparations, no ['4C]QA was formed when either A or B protein was incubated alone in a test tube with theother components of the QA synthetase system. These results suggest that one protein forms an intermediate which diffues across the membrane and is converted to QA by the other protein (or that one or both proteins may be leaking across the membrane). Since significantly higher amounts of ["C]QA were found in the chamber containing A protein, these experiments again suggest that the A protein catalyzes the final step in QA synthesis. Another set of experiments demonstrated that very little or no protein leakage across the membrane occurs and also that the intermediate formed is unstable. Table 3 shows the results of experiments in which either the A or the B
TABLE 2. Demonstration of the formation of a dialyzable intermediate in a cell-free systema Expt chamber Chamber con- Incubation Totaf dIpm tents time (min) no. no. QA A + all sub30 5,920 strates II B + all sub1,730 strates III A + all sub2 60 7,210 strates IV B + all sub3,183 strates °Incubations were carried out for the times indicated at 25'C in a plastic dialysis cell with compartments separated by Union Carbide dialysis membrane. Each chamber of the apparatUS contained 250 nmol of L-[U-U4C]aspartate (1 ,Ci), 370 nmol of DHAP, and 50 pmol of bicine buffer (pH 9.0) in a total volume of 0.5 ml. Where indicated, E. coli A protein purified through the Sephadex step (120 pg of protein) or calf liver B protein purified through the Sephadex step (242 pg of protein) was added. No FAD was added to these reactions, since calf liver B protein at this stage of purification is active in the absence of exogenous FAD. At the end of the incubation, the contents of each cell were removed to a conical centrifuge tube, the reaction was stopped with perchloric acid, and ["4C]QA formation was determined as previously described (18). I
1
TABLE
3. Dialysis experiments, test for stability of intermediate' Total dpm
Dialysis Preincubation Incubation of Expt chamber ai Incubation no. conditions conditions QAafter no.
1
2
mcubation 6,280
II
A + all substrates Substrates
B added
300
I
only B + all sub-
A added
9,700
I
B added
strates Substrates
A added 660 only 'The concentrations of substrates, buffer, and proteins were the same as described in Table 2. Dialysis cells were preincubated for 60 min with all substrates present in both chambers, but with either A or B protein alone in one chamber (total volume, 0.5 ml). Then the contents of each chamber were removed to a test tube, the missing protein was added to both mixtures in a volume of 0.2 ml, and incubation was continued for another 60 min. Reactions were stopped, and [14C]QA was determined as in Table 2. II
protein was incubated for 60 min in only one chamber of a dialysis cell. The contents of both chambers were removed to test tubes, the other protein (A or B) not present in the incubation was added, and incubation continued for another 60 min. In contrast to the results of the previous experiment, significant amounts of ['4C]QA were formed only when both proteins were present in the reaction mixture. However, there was a slight [140]QA formation in the incubation in which medium from chamber II, on the opposite side of the dialysis membrane from chamber I
VOL. 136, 1978
QUINOLINATE BIOSYNTHESIS IN E. COLI
which contained protein B incubated with all substrates, was added to the A protein. These data eliminate the possibility that the results in Table 2 were due to leakage of protein across the membrane. In the 60-min incubation shown in Table 2, sufficient intermediate crossed the membrane to fonn >7,000 cpm of [14C]QA on one side and >3,000 cpm on the other; however, in the experiments shown in Table 3 only enough active intermediate was present in the dialysate (chamber II) to form 660 cpm of [14C]QA. These results suggest that the internediate forned has a half-life of considerably less than 60 min.
A further set of experiments was carried out in an attempt to determine the substrate(s) required for intermediate formation. The basis of this experiment is that if the substrate for intermediate formation and its enzyme are on the same side of the membrane, then only diffusion of the intermediate across the membrane will be required for QA formation. If, on the other hand, the substrate for intermediate fornation and its enzyme are separated by the membrane, two diffusion events may be required before QA can be formed. Considerably more [14C]QA was formed when the incubation began with [14C]aspartate in the compartment with B protein than when aspartate and A protein were placed in one compartment and DHAP and B protein in the other (Table 4). Addition of DHAP along with aspartate to the chamber containing B protein (experiment 4) did not result in any increase in [14C]QA formation. These results are consistent with the formation of a compound from aspartate by the B protein, which is then condensed with DHAP and converted to QA by the A protein. Filter cone experiments. The occurrence of a diffusible internediate was further supported by experiments using Amicon membrane cones. In these experiments, A or B protein and some or all of the other components of the quinolinate synthetase system were placed in an Amicon filter cone, which was then placed in the mouth of a centrifuge tube containing the other components to complete the synthetase system. The tube was then centrifuged, and the small molecules were filtered through the cone into the reaction system below while the proteins were (in theory) retained in the cone (Table 5). In these experiments, B protein from E. coli was used because its apparent molecular weight of 85,000 as determined by gel filtration (18) is about twice that of the B protein activity prepared from beef liver (16). Experiment 1 (Table 5) was carried out as a control to show that small molecules pass through the cones while proteins
139
TABLE 4. Dialysis experiments, test for substrate requirement for intermediate formationa Expt no.
DialysisToadp
chamber Chamber contents Titnaldpm no.inQ 1 I A + [14C]aspartate 1,350 II B + DHAP 350 2 I A + DHAP 5,000 II B + [14C]aspartate 2,030 I 3 A + DHAP 5,010 B + ['4C]aspartate II 1,580 4 I A + DHAP 4,850 II B + [14C]aspartate 2,070 + DHAP aThe concentrations of substrates, buffer, and proteins were the same as described in Table 2. The dialysis cells were incubated for 30 min at 25°C, and the reactions were stopped and ['4C]QA was determined as described in Table 2.
do not. Unfortunately, as can be seen from experimnent l(b), significant amounts of protein A activity passed through the cone on centrifugation. The apparent molecular weight of A as determined by gel filtration is about 35,000 (18). In experiment 2 all the components of the QA synthetase system were incubated with A or B protein in the cone, and the reaction mixture was centrifuged down into a tube containing the other protein. When B was present in the cone [experiment 2(b)], 3,000 cpm of [14C]QA was formed. When A was present in the cone [experiment 2(a)], 1,130 cpm were detected. These results again suggest formation of an intermediate by the B protein. The activity observed in experiment 2(a) is probably due to passage of protein A activity through the filter cone (compare with experiment 1). Experiments 3 and 4 show that the intermediate produced by the B protein is formed from aspartate and that FAD is required for its fornation. DISCUSSION The evidence presented suggests that the B protein can forn an intermediate from aspartate which is then condensed with DHAP by the A protein to form QA. It remains to be established whether this compound is normally a free intermediate or like indole in the reaction catalyzed byE. coli tryptophan synthetase (7), an enzymebound intermediate which only appears free when the A and B proteins are artificially separated. Since the B protein required FAD, the intermediate forned could be a dehydrogenation product of aspartate, and therefore unstable. Imino-aspartate should hydrolyze to form oxaloacetate. The other two possible dehydrogenation products, amino-fumarate and amino-maleate, would isomerize to imino-aspartate and therefore also to oxaloacetate, but might have
140
J. BACTERIOL.
WICKS, SAKAKIBARA, AND GHOLSON
TABLE 5. Requirements for intermediate synthesis demonstrated using Amicon membrane Conesa Total dpm of ['4C]QA Components below cone Contents of cone Expt no. 9,970 A, B, DHAP, FAD ['4C]asp l(a) 1,490 ['4C]asp, B, DHAP, FAD A (b) 290 ['4C]asp, A, DHAP, FAD B (c)
2(a) (b) 3(a) (b)
[14C]asp, DHAP, FAD [14C]asp, DHAP, FAD
A, ['4C]asp, DHAP, FAD B, ['4C]asp, DHAP, FAD
B, A,
B, DHAP, FAD B, ['4C]asp, FAD
A, ['4C]asp, FAD A, DHAP, FAD
1,130 3,000 690 2,140
0 A, DHAP, FAD B, ['4C]asp 4(a) 6,130 A, DHAP B, ['4C]asp, FAD (b) a Amicon CF-25 membrane cones, containing reaction mixtures of the indicated composition in a total volume of 1.0 ml, were placed in 50-ml centrifuge tubes, containing the indicated components in a total volume of 1.0 ml, and were centrifuged for 15 min at 3,000 rpm in a SS-34 head of an RC2-B centrifuge (about 1,000 x g) at room temperature. This centrifugation forced 80 to 99+% of the liquid contents of the cone into the centrifuge tube. The centrifuge tubes were incubated at 25°C for a further 15 min. The reaction was then stopped, and ['4C]QA was determined as described in Table 2. All cones and tubes contained 100 ,umol of bicine (pH 8.0) and the following amounts of the other components as indicated: L-[U-"C]aspartate (["4C]asp), 0.25 ,anol (1 uCi); DHAP, 500 nmol; FAD, 20 ,umol; A protein, 0.4 ml of Sephadex fraction (0.2 mg of protein); B protein, 0.4 ml of Sephadex fraction (0.4 mg of protein); and water to make 1.0 ml.
slightly longer half-lives. The results of the dialysis experiments shown in Tables 2 and 3 indicate that the intermediate formed in our system is unstable. The order of proteins established in our experiments (B before A) is the reverse of that suggested by Kerr and Tritz (12) on the basis of cross-feeding experiments carried out in a liquid minimal medium supplemented with Casamino Acids. In those experiments nadA and nadB mutants were inoculated together into the mmnimal medium. An initial period of increase in absorbance lasting 6 to 8 h (presumably due to growth of both mutants) was followed by a lag period of about 6 h, after which increase in absorbance was again observed. Differential plate counts showed that the number of nadA cells increased from 108 to 101"/ml during the second growth phase, while the number of nadB cells reached a maximum of 5 x 10' cells per ml. It was concluded that the nadA cells were being fed by the nadB cells and that the nadA enzyme precedes the nadB enzyme in the pathway (12). However, no feeding between a nadC mutant, which should excrete QA under these conditions (3), and either the nadA or nadB mutant could be demonstrated. Also, the feeding of nadA cells by the nadB cells could only be observed when
Casamino Acids were added to the culture medium (12). A possible alternate explanation for these apparent cross-feeding results may be provided by the observations of Lundquist and Olivera (14) on niacin starvation of several nad mutants. These workers found that when logarithmically growing niacin-requiring cells were
transfered to a niacin-free medium, rapid growth continued for 2 h with a greater than 10-fold increase in numbers. After growth ceased, leakage of niacin (in the form of nicotinic acid ribonucleoside) began. If one assumes that the nadA cells used by Kerr and Tritz (12) were more resistant to niacin starvation than their nadB cells, then the further growth of the nadA celLs they observed might have been at the expense of nicotinic acid ribonucleoside excreted by stationary nadB cells. The discovery that the B protein alone produces a dialyzable product from aspartate should allow us to develop a less cumbersome assay for its activity than the currently used assay for quinolinate synthesis which is coupled to the A protein (18), and should accelerate the purification and characterization of the B protein of E. coli as well as the B protein activity in mammalian liver. ACKNOWLE DGM EMNTS This investigation was supported in part by National Science Foundation grant PCM74-20445 AO0. The very competent technical assistance of Harumi Hagasawa is greatfully acknowledged. LITERATURE CITED 1. Amed, F., and A. G. Moat. 1966. Nicotinic acid biosynthesis in prototrophs and tryptophan auxotrophs of Saccharomyces cerevisiae. J. Biol. Chem. 241:775-780. 2. Andreoli, A. J., M. Ekeda, Y. Nishizuka, and 0. Hayaishi. 1963. Quinolinic acid: a precursor to nicotinamide adenine dinucleotide in Esherichia coli. Biochem. Biophys. Res. Commun. 12:92-97. 3. Chandler, J. LR., and R. K. Gholson. 1972. De novo biosynthesis of nicotinamide adenine dinucleotide in
VOL. 136, 1978
4. 5.
6. 7.
8.
9.
10.
11.
12.
13.
QUINOLINATE BIOSYNTHESIS IN E. COLI
Escherichia coli: excretion of quinolinic acid by mutants lacking quinolinate phosphoribosyl transferase. J. Bacteriol. 111:98-102. Chayklin, S. 1967. Nicotinamide coenzymes. Annu. Rev. Biochem. 36:149-170. Chen, J., and G. J. Tritz. 1975. Isolation of a metabolite capable of differentially supporting the growth of nicotinamide adenine dinucleotide auxotrophs of Escherichia coli. J. Bacteriol. 121:212-218. Chen, J., and G. J. Tritz. 1976. Detection of precursors of quinolinic acid in Escherichia coli. Microbios 16:207-218. Crawford, I. P. 1974. Tryptophan synthetase, p. 223-265. In K. E. Ebner (ed.), Subunit enzymes, vol. 2. Marcel Dekker, Inc., New York. Gholson, R. K., I. Ueda, N. Ogasawara, and L. M. Henderson. 1964. The enzymatic conversion of quinolinate to nicotinic acid mononucleotide in mammalian liver. J. Biol. Chem. 239:1208-1214. Griffith, G. R., J. LR. Chandler, and R. K. Gholson. 1975. Studies on the de novo biosynthesis of NAD in Escherichia coli. VI. The separation of the nadB gene product from the nadA gene product and its purification. Eur. J. Biochem. 54:239-245. Hadwiger, L A., S. E. Badiei, G. R. Waller, and R. K. Gholson. 1963. Quinolinic acid as a precursor of nicotinic acid and its derivatives in plants. Biochem. Biophys. Res. Commun. 13:466-471. Jackson, R. W., and J. A. DeMoss. 1965. Effects of toluene on Escherichia coli. J. Bacteriol. 90:1420-1425. Kerr, T. J., and G. J. Tritz. 1973. Cross-feeding of Escherichia coli mutants defective in the biosynthesis of nicotinamide adenine dinucleotide. J. Bacteriol. 115:982-986. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275.
141
14. Lundquist, R., and B. M. Olivera. 1973. Pyridine nucleotide metabolism in Escherichia coli. II. Niacin starvation. J. Biol. Chem. 248:5137-5143. 15. Ogasawara, N., J. L R. Chandler, R. K. Gholson, R. J. Rosser, and A. J. Andreoli. 1967. Biosynthesis of quinolinic acid in a cell-free system. Biochim. Biophys. Acta 141:199-201. 16. Sakakibara, S., F. D. Wicks, and R. K. Gholson. 1977. Occurrence in mammalian liver of a protein which replaces the B protein of E. coli quinolinate synthetase. Biochem. Biophys. Res. Commun. 76:158-166. 17. Scott, T. A., E. Bellion, and M. Mattey. 1969. The conversion of N-formyl-L-asparatate into nicotinic acid by extracts of Clostridium butylicum. Eur. J. Biochem. 10:318-320. 18. Suzuki, N., J. P. Carlson, G. R. Griffith, and R. K. Gholson. 1973. Studies on the de novo biosynthesis of NAD in Escherichia coli. V. Properties of the quinolinic acid synthetase system. Biochim. Biophys. Acta 340:309-315. 19. Taylor, A. L 1970. Current linkage map of Escherichia coli. Bacteriol. Rev. 34:155-175. 20. Tritz, G. J., T. S. Matney, J. L R. Chandler, and R. K. Gholson. 1970. Chromosomal location of the C gene involved in the biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli K-12. J. Bacteriol.
104:45-59. 21. Tritz, G. J., T. S. Matney, and R. K. Gholson. 1970. Mapping of the nadB locus adjacent to a previously undescribed purine locus in Escherichia coli K-12. J. Bacteriol. 102:377-381. 22. Warburg, O., and W. Chrisdan. 1936. Pyridin, der wasserstoffuibertragende. Bestandteil von Garungpfermenten (Pyridin-Nucleotide). Biochem. Z. 287:291-297. 23. Yates, R. A., and A. B. Pardee. 1956. Pyrimidine biosynthesis in Escherichia coli. J. Biol. Chem. 221:743-756.
JOURNAL OF BACTERIOLOGY, Oct. 1978, P. 136-141 0021-9193/78/0136-0136$02.00/0 Copyright i) 1978 American Society for Microbiology
Printed in U.S.A.
Evidence for an Intermediate in Quinolinate Biosynthesis in Escherichia colit FRANK D. WICKS, SHIGEKI SAKAKIBARA,4 AND R. K. GHOLSON* Department ofBiochemistry, Oklahoma State University, Stillwater, Oklahoma 74074 Received for publication 8 May 1978
Evidence for the formation of an unstable intermediate in the synthesis of quinolinate from aspartate and dihydroxyacetone phosphate by Escherichia coli was obtained using toluenized cells of nadA and nadB mutants of this organism and partially purified A and B proteins in dialysis and membrane cone experiments. The results of these experiments indicate that the nadB gene product forms an unstable compound from aspartate in the presence of flavine adenine dinucleotide, and that this compound is then condensed with dihydroxyacetone phosphate to form quinolinate in a reaction catalyzed by the nadA gene product.
Quinolinic acid (pyridine 2,3-dicarboxylic acid) (QA) is a precursor of the pyridine nucleotides in mammals (8), plants (10), and microorganisms (2). The pathway for the synthesis of QA from tryptophan in mammalian liver and in Neurospora has been clearly elucidated by the efforts of a number of investigators (4). This tryptophan-QA pathway is also present in aerobically grown yeast (1). On the other hand, the details of QA synthesis in plants and most microorganisms have not been reported. Radioactive labeling studies in vivo and in vitro have suggested at least three separate pathways other than the tryptophan pathway for QA synthesis in various microorganisms (15, 17; T. A. Scott and H. Hussey, Biochem. J. 96:9c). In the most thoroughly studied system which is present in aerobically grown Escherichia coli, two protein fractions have been shown to be required for the synthesis of QA from aspartate and dihydroxyacetone phosphate (DHAP) (9,18). One of these, the "B" protein, is coded for by the nadB gene (21) and the other, the "A" protein, by the nadA gene (19). Although these proteins have been partially purified, and the properties of the in vitro QA synthetase system have been studied in some detail, previous efforts to detect the formation of intermediate(s) between aspartate and DHAP and QA in this system were unsuccessful (9, 18). On the other hand, a series of experiments have been reported by other workers which suggest that several intermediates in QA biosynthesis may be formed from aspartate and DHAP in E. coli. Cross-feeding experiments have been reported (12) which indicate that the t Journal article J3462 of the Agricultural Experiment Station, Oklahoma State University, Stillwater, OK 74074. t Present address: Department of Medical Chemiatry, Osaka Medical College, Osaka, Japan.
13E
nadA-coded enzyme produces a compound which can be converted to QA by a mutant containing an intact nadB-coded enzyme. Subsequently, a nadC mutant, which is blocked in the conversion of QA to nicotinic acid mononucleotide (20), was shown to excrete a compound which was not QA, but which stimulated the growth of both nadA and nadB mutants (5). More recently, Chen and Tritz (6) have concluded that five unidentified compounds formed from fructose-1,6-diphosphate and/or aspartate in a cell-free system prepared from a nadC mutant are precursors of QA, since formation of these compounds is repressed by nicotinic acid and inhibited by oxidized nicotinamide adenine dinucleotide. This communication reports evidence that the nadB gene product, in the presence of flavine adenine dinucleotide (FAD), converts aspartate into an unstable compound which is then condensed with DHAP to forn QA in a reaction catalyzed by the nadA gene product. MATERIALS AND METHODS Bacterial strains. All bacteria used were derivatives of E. coli K-12. Strain UTH 4460 contains nadA19 and was originally obtained from R. A. LaValle as his PA-2-18. Strain UTH 4451 contains nadB35 and was originally obtained from R. A. LaValle as P4X-SB16. Strain UTH 4464 contains nadB7 and was originally obtained from T. K. Sundarum as W-3899. These strains are referred to in this paper by their nad allele numbers, which are Coli Genetic Stock Center (Yale University) designations. Growth and harvesting of cells. The minimal medium of Yates and Pardee (23), supplemented with 0.5 ,M nicotinic acid and other necessary growth factors, was used. Cultures were grown in 10-liter tanks in a New Brunswick fermentor to late log phase and were harvested in a Sharples centrifuge. The cell paste
QUINOLINATE BIOSYNTHESIS IN E. COLI
VOL. 136, 1978
washed once with 0.9% saline. Cells from which A and B proteins were isolated were stored at -20°C until used. Preparation of A and B proteins. Frozen cells of nadAI9 and nad35 were suspended in 20 volumes of 0.05 M bicine buffer (pH 8.0), sonically disrupted, and centrifuged as previously described (9). The A protein was purified from the crude extract of nadB35 cells through the 40 to 60% ammonium sulfate step and the B protein was purified through the Sephadex G-200 step, as described by Griffith et al. (9). The proteinreplacing activity of liver, which was used as the B protein source in some experiments, was purified through the Sephadex G-75 step as described by Sakakibara et al. (16). Preparation of toluenized celis. Toluenized cells of nadA19 and nadB7 were prepared following the methods described by Jackson and DeMoss (11). A 0.2-g sample of washed cells was suspended in 1 ml of 0.05 M potassium phosphate buffer (pH 8.0); 100 pl of toluene was added, and the suspension was mixed vigorously at room temperature for 1 min. Toluenized cell suspensions were stored in an ice bath until used (no longer than 30 min). Other Methods. Assay for QA formation during A and B protein purification was carried out as previously described (9). Assay for QA formation in those experiments described here used these methods (9), with some modifications as described in the table legends. Protein was determined by the method of Lowry et al. (13) in crude extracts and spectrophotometrically by the 280/260 absorbance ratio (22) in more purified preparations. The filter cone experiments were carried out using Amicon CF 25 Centriflo membrane cones obtained from Amicon Corp., Lexington, Mass. Various components of the quinolinate synthetase system, in a total volume of 1.0 ml, were placed in a membrane cone, which was inserted in a 50-ml centrifuge tube containing the remaining system components not present in the cone. The tubes were then centrifuged at 3,000 rpm in a SS-34 rotor of a model RC2-B centrifuge (about 1,000 x g) at room temperature for 15 min. Centrifugation forced 80 to 95+% of the liquid contents of the membrane cone into the centrifuge tube. The centrifuge tubes were then incubated at 25°C for a further 15 min. The reaction was stopped, and [14C]quinolinate was determined as previously described (18). The exact contents of membrane cones and tubes in various experiments are given in Table 5. Chemicals. L-[U-14C]aspartate (160 mCi/mmol) was obtained from New England Nuclear. All other chemicals were of reagent grade, obtained from commercial supply houses.
was
RESULTS
Toluenized cela. The first evidence for the formation of a free intermediate in QA biosynthesis in E. coli obtained in our laboratory was provided by experiments carried out with toluenized cells. Treatment of E. coli with low concentrations of toluene causes the cell membranes to become freely perneable to low-mo-
137
lecular-weight metabolites, while macromolecules are retained within the cells (11). When toluenized nadA (PA-2-18) cells were incubated with all the components of the in vitro QA synthetase system, no detectable [14C]QA was formed (Table 1). When toluenized nadB (W3899) cells were incubated under these same conditions, 1,146 cpm of ['4C]QA was fonned. This result is consistent with our findings that many nadB mutants are slightly leaky and contain low but detectable amounts of B protein (9). However, when toluenized nadA and nadB cells were mixed together and incubated in the reaction mixture (Table 1, experiment 3), five times more [14C]QA was formed. This suggested that a compound formed from the substrate(s) by one mutant is being converted to QA by the other mutant cell type or that the A and B proteins are leaking from the toluenized cells and are catalyzing the extracellular fornation of QA from the substrates. The possibility of leakage was tested by using the supernatant solutions remaining after centrifugation of the 20% suspensions of toleuized cells rather than the cells themselves. Although some [14C]QA was formed under these conditions (experiments 4 and 5, Table 1), this was significantly less than observed when both types of cells were present at the same time, making it unlikely that the results observed in experiment 3 were due to protein leakage. Since the toluenized cells could be rapidly removed from the reaction by centrifugation, it TABLE 1. Quinolinate synthesis in toluenized nadA and nadB cellsa Expt Incubation
Enot Incubat1on 1
no.
. 2 Incubation
Total
dpm of [4C]QA
nadA cells 0 nadB cells 1,146 nadA cells + nadB cells 6,516 nadA cella + superna294 tant from nadB cells 5 nadB cells + superna2,148 tant from nadA cells 6 nadB cells nadA cells + superna1,596 tant from incubation 1 7 nadA cells nadB cells + superna7,248 tant from incubation 1 a Incubations were carried out at 25°C in a total volume of 0.5 ml containing L-[U-"CJaspartate, 250 nmol (IluCi); DHAP, 100 nmol; FAD, 10 gtmol; and bicine buffer (pH 8.0), 50 ,umol for 15 min. A 0.1-ml sample of a 20% suspension of toluenized cells from a nadA or nadB mutant was used as indicated. In experiments 4 and 5, 0.1 ml of the supernatant from centrifuged suspensions of toluenized cells was added as control for protein leakage. In experiments 6 and 7, 0.1 ml of toluenized cells was incubated with the complete reaction mixture for 15 min, the cells were removed by centrifugation, and the supernatant was added to 0.1 ml of a 20% suspension of toluenized nadA or nadB cells and incubated for an additional 15 min. 1 2 3 4
-
138
J. BACTERIOL.
WICKS, SAKAKIBARA, AND GHOLSON
was possible to test which mutant cell type was forming the intermediate by incubating each cell type in a complete reaction medium for 15 min, centrifuging out the cells, and transferring the medium to another tube containing the other cell type for an additional 15-min period. The results of experiments 6 and 7 (Table 1) show that much more ["4C]QA was formed when the incubation medium from nadA cells was transferred to nadB cells for further incubation than when the reverse sequence was followed. These results are consistent with the formation of a diffusible intermediate by the nadB gene product (B protein), which is then converted to QA by the nadA gene product (A protein). These experiments with toluenized cells could be subject to error because of the leakiness of the nadB mutation and the possibility of the incomplete removal of cells by centrifugation. Experiments with partially purified A and B proteins detailed below, however, suggest these same conclusions. Dialysis experiments. Further support for the occurrence of a diffusible intermediate in QA biosynthesis was provided by experiments in which partially purified A and B proteins were incubated on opposite sides of a dialysis membrane. In these experiments B protein prepared from calf liver was used, since enzyme from this source appeared to have consistently higher activity than E. coli B protein (16). All substrate components of the quinolinate synthetase system were placed in both cells of a dialysis apparatus with protein A on one side of the membrane and protein B on the other side. After incubation for 30- and 60-min intervals, the QA content on both sides of the membrane was determined. ['4C]QA was found to be present on both sides of the membrane but with larger amounts in the chamber containing A protein (Table 2). In control experiments with the same enzyme preparations, no ['4C]QA was formed when either A or B protein was incubated alone in a test tube with theother components of the QA synthetase system. These results suggest that one protein forms an intermediate which diffues across the membrane and is converted to QA by the other protein (or that one or both proteins may be leaking across the membrane). Since significantly higher amounts of ["C]QA were found in the chamber containing A protein, these experiments again suggest that the A protein catalyzes the final step in QA synthesis. Another set of experiments demonstrated that very little or no protein leakage across the membrane occurs and also that the intermediate formed is unstable. Table 3 shows the results of experiments in which either the A or the B
TABLE 2. Demonstration of the formation of a dialyzable intermediate in a cell-free systema Expt chamber Chamber con- Incubation Totaf dIpm tents time (min) no. no. QA A + all sub30 5,920 strates II B + all sub1,730 strates III A + all sub2 60 7,210 strates IV B + all sub3,183 strates °Incubations were carried out for the times indicated at 25'C in a plastic dialysis cell with compartments separated by Union Carbide dialysis membrane. Each chamber of the apparatUS contained 250 nmol of L-[U-U4C]aspartate (1 ,Ci), 370 nmol of DHAP, and 50 pmol of bicine buffer (pH 9.0) in a total volume of 0.5 ml. Where indicated, E. coli A protein purified through the Sephadex step (120 pg of protein) or calf liver B protein purified through the Sephadex step (242 pg of protein) was added. No FAD was added to these reactions, since calf liver B protein at this stage of purification is active in the absence of exogenous FAD. At the end of the incubation, the contents of each cell were removed to a conical centrifuge tube, the reaction was stopped with perchloric acid, and ["4C]QA formation was determined as previously described (18). I
1
TABLE
3. Dialysis experiments, test for stability of intermediate' Total dpm
Dialysis Preincubation Incubation of Expt chamber ai Incubation no. conditions conditions QAafter no.
1
2
mcubation 6,280
II
A + all substrates Substrates
B added
300
I
only B + all sub-
A added
9,700
I
B added
strates Substrates
A added 660 only 'The concentrations of substrates, buffer, and proteins were the same as described in Table 2. Dialysis cells were preincubated for 60 min with all substrates present in both chambers, but with either A or B protein alone in one chamber (total volume, 0.5 ml). Then the contents of each chamber were removed to a test tube, the missing protein was added to both mixtures in a volume of 0.2 ml, and incubation was continued for another 60 min. Reactions were stopped, and [14C]QA was determined as in Table 2. II
protein was incubated for 60 min in only one chamber of a dialysis cell. The contents of both chambers were removed to test tubes, the other protein (A or B) not present in the incubation was added, and incubation continued for another 60 min. In contrast to the results of the previous experiment, significant amounts of ['4C]QA were formed only when both proteins were present in the reaction mixture. However, there was a slight [140]QA formation in the incubation in which medium from chamber II, on the opposite side of the dialysis membrane from chamber I
VOL. 136, 1978
QUINOLINATE BIOSYNTHESIS IN E. COLI
which contained protein B incubated with all substrates, was added to the A protein. These data eliminate the possibility that the results in Table 2 were due to leakage of protein across the membrane. In the 60-min incubation shown in Table 2, sufficient intermediate crossed the membrane to fonn >7,000 cpm of [14C]QA on one side and >3,000 cpm on the other; however, in the experiments shown in Table 3 only enough active intermediate was present in the dialysate (chamber II) to form 660 cpm of [14C]QA. These results suggest that the internediate forned has a half-life of considerably less than 60 min.
A further set of experiments was carried out in an attempt to determine the substrate(s) required for intermediate formation. The basis of this experiment is that if the substrate for intermediate formation and its enzyme are on the same side of the membrane, then only diffusion of the intermediate across the membrane will be required for QA formation. If, on the other hand, the substrate for intermediate fornation and its enzyme are separated by the membrane, two diffusion events may be required before QA can be formed. Considerably more [14C]QA was formed when the incubation began with [14C]aspartate in the compartment with B protein than when aspartate and A protein were placed in one compartment and DHAP and B protein in the other (Table 4). Addition of DHAP along with aspartate to the chamber containing B protein (experiment 4) did not result in any increase in [14C]QA formation. These results are consistent with the formation of a compound from aspartate by the B protein, which is then condensed with DHAP and converted to QA by the A protein. Filter cone experiments. The occurrence of a diffusible internediate was further supported by experiments using Amicon membrane cones. In these experiments, A or B protein and some or all of the other components of the quinolinate synthetase system were placed in an Amicon filter cone, which was then placed in the mouth of a centrifuge tube containing the other components to complete the synthetase system. The tube was then centrifuged, and the small molecules were filtered through the cone into the reaction system below while the proteins were (in theory) retained in the cone (Table 5). In these experiments, B protein from E. coli was used because its apparent molecular weight of 85,000 as determined by gel filtration (18) is about twice that of the B protein activity prepared from beef liver (16). Experiment 1 (Table 5) was carried out as a control to show that small molecules pass through the cones while proteins
139
TABLE 4. Dialysis experiments, test for substrate requirement for intermediate formationa Expt no.
DialysisToadp
chamber Chamber contents Titnaldpm no.inQ 1 I A + [14C]aspartate 1,350 II B + DHAP 350 2 I A + DHAP 5,000 II B + [14C]aspartate 2,030 I 3 A + DHAP 5,010 B + ['4C]aspartate II 1,580 4 I A + DHAP 4,850 II B + [14C]aspartate 2,070 + DHAP aThe concentrations of substrates, buffer, and proteins were the same as described in Table 2. The dialysis cells were incubated for 30 min at 25°C, and the reactions were stopped and ['4C]QA was determined as described in Table 2.
do not. Unfortunately, as can be seen from experimnent l(b), significant amounts of protein A activity passed through the cone on centrifugation. The apparent molecular weight of A as determined by gel filtration is about 35,000 (18). In experiment 2 all the components of the QA synthetase system were incubated with A or B protein in the cone, and the reaction mixture was centrifuged down into a tube containing the other protein. When B was present in the cone [experiment 2(b)], 3,000 cpm of [14C]QA was formed. When A was present in the cone [experiment 2(a)], 1,130 cpm were detected. These results again suggest formation of an intermediate by the B protein. The activity observed in experiment 2(a) is probably due to passage of protein A activity through the filter cone (compare with experiment 1). Experiments 3 and 4 show that the intermediate produced by the B protein is formed from aspartate and that FAD is required for its fornation. DISCUSSION The evidence presented suggests that the B protein can forn an intermediate from aspartate which is then condensed with DHAP by the A protein to form QA. It remains to be established whether this compound is normally a free intermediate or like indole in the reaction catalyzed byE. coli tryptophan synthetase (7), an enzymebound intermediate which only appears free when the A and B proteins are artificially separated. Since the B protein required FAD, the intermediate forned could be a dehydrogenation product of aspartate, and therefore unstable. Imino-aspartate should hydrolyze to form oxaloacetate. The other two possible dehydrogenation products, amino-fumarate and amino-maleate, would isomerize to imino-aspartate and therefore also to oxaloacetate, but might have
140
J. BACTERIOL.
WICKS, SAKAKIBARA, AND GHOLSON
TABLE 5. Requirements for intermediate synthesis demonstrated using Amicon membrane Conesa Total dpm of ['4C]QA Components below cone Contents of cone Expt no. 9,970 A, B, DHAP, FAD ['4C]asp l(a) 1,490 ['4C]asp, B, DHAP, FAD A (b) 290 ['4C]asp, A, DHAP, FAD B (c)
2(a) (b) 3(a) (b)
[14C]asp, DHAP, FAD [14C]asp, DHAP, FAD
A, ['4C]asp, DHAP, FAD B, ['4C]asp, DHAP, FAD
B, A,
B, DHAP, FAD B, ['4C]asp, FAD
A, ['4C]asp, FAD A, DHAP, FAD
1,130 3,000 690 2,140
0 A, DHAP, FAD B, ['4C]asp 4(a) 6,130 A, DHAP B, ['4C]asp, FAD (b) a Amicon CF-25 membrane cones, containing reaction mixtures of the indicated composition in a total volume of 1.0 ml, were placed in 50-ml centrifuge tubes, containing the indicated components in a total volume of 1.0 ml, and were centrifuged for 15 min at 3,000 rpm in a SS-34 head of an RC2-B centrifuge (about 1,000 x g) at room temperature. This centrifugation forced 80 to 99+% of the liquid contents of the cone into the centrifuge tube. The centrifuge tubes were incubated at 25°C for a further 15 min. The reaction was then stopped, and ['4C]QA was determined as described in Table 2. All cones and tubes contained 100 ,umol of bicine (pH 8.0) and the following amounts of the other components as indicated: L-[U-"C]aspartate (["4C]asp), 0.25 ,anol (1 uCi); DHAP, 500 nmol; FAD, 20 ,umol; A protein, 0.4 ml of Sephadex fraction (0.2 mg of protein); B protein, 0.4 ml of Sephadex fraction (0.4 mg of protein); and water to make 1.0 ml.
slightly longer half-lives. The results of the dialysis experiments shown in Tables 2 and 3 indicate that the intermediate formed in our system is unstable. The order of proteins established in our experiments (B before A) is the reverse of that suggested by Kerr and Tritz (12) on the basis of cross-feeding experiments carried out in a liquid minimal medium supplemented with Casamino Acids. In those experiments nadA and nadB mutants were inoculated together into the mmnimal medium. An initial period of increase in absorbance lasting 6 to 8 h (presumably due to growth of both mutants) was followed by a lag period of about 6 h, after which increase in absorbance was again observed. Differential plate counts showed that the number of nadA cells increased from 108 to 101"/ml during the second growth phase, while the number of nadB cells reached a maximum of 5 x 10' cells per ml. It was concluded that the nadA cells were being fed by the nadB cells and that the nadA enzyme precedes the nadB enzyme in the pathway (12). However, no feeding between a nadC mutant, which should excrete QA under these conditions (3), and either the nadA or nadB mutant could be demonstrated. Also, the feeding of nadA cells by the nadB cells could only be observed when
Casamino Acids were added to the culture medium (12). A possible alternate explanation for these apparent cross-feeding results may be provided by the observations of Lundquist and Olivera (14) on niacin starvation of several nad mutants. These workers found that when logarithmically growing niacin-requiring cells were
transfered to a niacin-free medium, rapid growth continued for 2 h with a greater than 10-fold increase in numbers. After growth ceased, leakage of niacin (in the form of nicotinic acid ribonucleoside) began. If one assumes that the nadA cells used by Kerr and Tritz (12) were more resistant to niacin starvation than their nadB cells, then the further growth of the nadA celLs they observed might have been at the expense of nicotinic acid ribonucleoside excreted by stationary nadB cells. The discovery that the B protein alone produces a dialyzable product from aspartate should allow us to develop a less cumbersome assay for its activity than the currently used assay for quinolinate synthesis which is coupled to the A protein (18), and should accelerate the purification and characterization of the B protein of E. coli as well as the B protein activity in mammalian liver. ACKNOWLE DGM EMNTS This investigation was supported in part by National Science Foundation grant PCM74-20445 AO0. The very competent technical assistance of Harumi Hagasawa is greatfully acknowledged. LITERATURE CITED 1. Amed, F., and A. G. Moat. 1966. Nicotinic acid biosynthesis in prototrophs and tryptophan auxotrophs of Saccharomyces cerevisiae. J. Biol. Chem. 241:775-780. 2. Andreoli, A. J., M. Ekeda, Y. Nishizuka, and 0. Hayaishi. 1963. Quinolinic acid: a precursor to nicotinamide adenine dinucleotide in Esherichia coli. Biochem. Biophys. Res. Commun. 12:92-97. 3. Chandler, J. LR., and R. K. Gholson. 1972. De novo biosynthesis of nicotinamide adenine dinucleotide in
VOL. 136, 1978
4. 5.
6. 7.
8.
9.
10.
11.
12.
13.
QUINOLINATE BIOSYNTHESIS IN E. COLI
Escherichia coli: excretion of quinolinic acid by mutants lacking quinolinate phosphoribosyl transferase. J. Bacteriol. 111:98-102. Chayklin, S. 1967. Nicotinamide coenzymes. Annu. Rev. Biochem. 36:149-170. Chen, J., and G. J. Tritz. 1975. Isolation of a metabolite capable of differentially supporting the growth of nicotinamide adenine dinucleotide auxotrophs of Escherichia coli. J. Bacteriol. 121:212-218. Chen, J., and G. J. Tritz. 1976. Detection of precursors of quinolinic acid in Escherichia coli. Microbios 16:207-218. Crawford, I. P. 1974. Tryptophan synthetase, p. 223-265. In K. E. Ebner (ed.), Subunit enzymes, vol. 2. Marcel Dekker, Inc., New York. Gholson, R. K., I. Ueda, N. Ogasawara, and L. M. Henderson. 1964. The enzymatic conversion of quinolinate to nicotinic acid mononucleotide in mammalian liver. J. Biol. Chem. 239:1208-1214. Griffith, G. R., J. LR. Chandler, and R. K. Gholson. 1975. Studies on the de novo biosynthesis of NAD in Escherichia coli. VI. The separation of the nadB gene product from the nadA gene product and its purification. Eur. J. Biochem. 54:239-245. Hadwiger, L A., S. E. Badiei, G. R. Waller, and R. K. Gholson. 1963. Quinolinic acid as a precursor of nicotinic acid and its derivatives in plants. Biochem. Biophys. Res. Commun. 13:466-471. Jackson, R. W., and J. A. DeMoss. 1965. Effects of toluene on Escherichia coli. J. Bacteriol. 90:1420-1425. Kerr, T. J., and G. J. Tritz. 1973. Cross-feeding of Escherichia coli mutants defective in the biosynthesis of nicotinamide adenine dinucleotide. J. Bacteriol. 115:982-986. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275.
141
14. Lundquist, R., and B. M. Olivera. 1973. Pyridine nucleotide metabolism in Escherichia coli. II. Niacin starvation. J. Biol. Chem. 248:5137-5143. 15. Ogasawara, N., J. L R. Chandler, R. K. Gholson, R. J. Rosser, and A. J. Andreoli. 1967. Biosynthesis of quinolinic acid in a cell-free system. Biochim. Biophys. Acta 141:199-201. 16. Sakakibara, S., F. D. Wicks, and R. K. Gholson. 1977. Occurrence in mammalian liver of a protein which replaces the B protein of E. coli quinolinate synthetase. Biochem. Biophys. Res. Commun. 76:158-166. 17. Scott, T. A., E. Bellion, and M. Mattey. 1969. The conversion of N-formyl-L-asparatate into nicotinic acid by extracts of Clostridium butylicum. Eur. J. Biochem. 10:318-320. 18. Suzuki, N., J. P. Carlson, G. R. Griffith, and R. K. Gholson. 1973. Studies on the de novo biosynthesis of NAD in Escherichia coli. V. Properties of the quinolinic acid synthetase system. Biochim. Biophys. Acta 340:309-315. 19. Taylor, A. L 1970. Current linkage map of Escherichia coli. Bacteriol. Rev. 34:155-175. 20. Tritz, G. J., T. S. Matney, J. L R. Chandler, and R. K. Gholson. 1970. Chromosomal location of the C gene involved in the biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli K-12. J. Bacteriol.
104:45-59. 21. Tritz, G. J., T. S. Matney, and R. K. Gholson. 1970. Mapping of the nadB locus adjacent to a previously undescribed purine locus in Escherichia coli K-12. J. Bacteriol. 102:377-381. 22. Warburg, O., and W. Chrisdan. 1936. Pyridin, der wasserstoffuibertragende. Bestandteil von Garungpfermenten (Pyridin-Nucleotide). Biochem. Z. 287:291-297. 23. Yates, R. A., and A. B. Pardee. 1956. Pyrimidine biosynthesis in Escherichia coli. J. Biol. Chem. 221:743-756.
