Nicotinamide Adenine Dinucleotide Biosynthesis

MICROBIOLOGICAL REVIEWS, Ma. 1980, p. 83-105 0146-0749/80/01-0083/23$02.00/0

Vol. 44, No. 1

Nicotinamide Adenine Dinucleotide Biosynthesis and Pyridine Nucleotide Cycle Metabolism in Microbial Systems JOHN W. FOSTER* AND ALBERT G. MOAT Department of Microbiology, Marshall University School ofMedicine, Huntington, West Virginia 25701 ......... .......... .................. INTRODUCTION ...... BIOSYNTHESIS OF NICOTINAMIDE ADENINE DINUCLEOTIDE ...... .... Anaerobic De Novo Pathways .. ....

83 83 ... 84 84 Dihydroxyacetone phosphate-aspartate pathway Formate-aspartate pathway .............................................. 87 Aerobic Tryptophan Catabolic Pathway 87 Genetics of Anaerobic Nicotinamide Adenine Dinucleotide Biosynthesis .... 89 Regulation of De Novo Biosynthesis of Nicotinamide Adenine Dinucleotide 91 93 PYRIDINE NUCLEOTIDE CYCLE METABOLISM ...................... 93 Biochemistry .............................. Pyridine Nucleotide Cycle Genetics .95 Regulation .96 98 Organisms of Special Interest 98 Haenwphilus ...................9.8...................... ................ .... 98 Mycobacterium ....................... 98 Clostridium butylicum 99 .................. ............ ......... Azotobacter vinelandii ......... 99 Lactobacillus and Leuconostoc 99 EVOLUTIONARY ASPECTS Evolution of Biosynthetic Pathways 99 Evolution of the Pyridine Nucleotide Cycle ........ 100 SUMMARY .100 LITERATURE CITED ..... ............................ ... ... ... 102

phasis on the need for more extensive research into the synthesis, recycling, and regulation of NAD metabolism. No comprehensive review on NAD metabolism has been published since the review by Chaykin in 1967 (16). Meanwhile, a considerable amount of work has been published on this subject. The purpose of this article, then, is to summarize the most recent developments regarding the biochemistry of NAD metabolism as well as the regulation and genetics of this system. Emphasis will be placed on the E. coliSalmonella typhimurium systems, since they are understood best. For the reader's convenience, Tables 1 and 2 present a summary of the enzymes considered in this review.

INTRODUCTION Nicotinamide adenine dinucleotide (NAD) and NAD-phosphate (NADP) are compounds of immeasurable importance in cellular metabolism. They function in numerous anabolic and catabolic reactions and are widely distributed throughout biological systems. The structures of these compounds are presented in Fig. 1 (113). NAD and NADP are known to participate in over 300 enzymatically catalyzed oxidation-reduction reactions. In addition, a number of reactions have been discovered in which NAD serves as a substrate. For example, certain procaryotes, such as Escherichia coli, utilize NAD as a substrate for deoxyribonucleic acid ligase, an essential for deoxyribonucleic acid synthesis, repair, and recombination (68, 80). NAD also serves as a substrate in reactions that produce poly-adenosine 5'-diphosphate-ribose (47, 104). Adenosine 5'-diphosphate ribosylation is proving to be of great importance to both eucaryotic and procaryotic cells (20, 21, 44, 47). Reduced pyridine nucleotide coenzymes also play an important role in the regulation of amphibolic pathways, such as the citric acid cycle and the oxidative pentose pathway (91). Thus, there is a growing awareness of the extent to which cells are dependent upon NAD and an increased em-

BIOSYNTHESIS OF NICOTINAMIDE ADENINE DINUCLEOTIDE Extensive research on the biosynthesis of NAD has been undertaken in mammalian and lower eucaryotic systems. Similar studies in procaryotes have not been pursued as widely. This is surprising in view of the potential significance associated with the evolutionary development of procaryotic systems for NAD biosynthesis. A review of the literature reveals two main biosynthetic pathways to NAD. One pathway involves 83

84

FOSTER AND MOAT

N..,CONH2 O _O_P_O_Co

+

|HH o

HO

NH20H

-0 -p- H HO

OH tLO PO3

FIG. 1. Molecular structures of NAD and NADP. The arrow indicates the point at which the third phosphate group is added by NAD kinase to form NADP.

the aerobic degradation of tryptophan by mammalian cells and a number of lower eucaryotes. Another pathway, found predominantly in procaryotes, is anaerobic and utilizes low-molecular-weight precursors for the synthesis of the pyridine ring structure of NAD. Both pathways lead to the formation of quinolinic acid (QA). Subsequent conversion of QA to NAD occurs via a pathway common to all organisms that have been examined to date.

Anaerobic De Novo Pathways Relatively few procaryotic species are capable of using tryptophan for the formation of the pyridine ring of NAD. Xanthomonas pruni can convert tryptophan to NAD (16, 117), and the enzymes in the pathway have been well documented. Some members of the actinomycete group also appear to utilize tryptophan for NAD biosynthesis (70). In view of the importance of NAD and NADP in procaryotic metabolism, it is surprising that investigations as to the route of synthesis of these essential cofactors have been so late in developing. Dihydroxyacetone phosphate-aspartate pathway. The first attempt to elucidate a procaryotic pathway to NAD was reported by Ortega and Brown in 1960 (81). They implicated glycerol and a dicarboxylic acid as precursors in the synthesis of the pyridine ring of NAD by E. coli. Subsequently, Andreoli and his co-workers demonstrated that QA was a key intermediate in this de novo pathway (4). QA is now recognized as a precursor involved in all known bio-

MICROBIOL. REV.

synthetic pathways to NAD (Fig. 2). Subsequent work by Chandler et al. (15) established L-aspartic acid as the dicarboxylic acid precursor. Suzuki et al., in 1973, established that the threecarbon precursor is dihydroxyacetone phosphate (DHAP) and not glycerol (107). L-Aspartic acid and DHAP undergo a condensation reaction to form an intermediate which is cyclized to form the pyridine ring of QA. Labeling studies have shown that C-3 of DHAP condenses with C-3 of aspartate (116). Similar studies with Mycobacterium tuberculosis have unequivocally established that the aspartate nitrogen and carbon are incorporated intact into the pyridine ring (42). The condensation between aspartate and DHAP is a two-enzyme step, collectively termed the QA synthetase system, for which an intermediate has not yet been isolated in pure form (Fig. 3). Mutants defective in this de novo pathway have been isolated from E. coli and S. typhimurium, and the relevant genes were designated nadA, nadB, nadC, and nadR (see Genetics of Anaerobic Nicotinamide Adenine Dinucleotide Biosynthesis). Chen and Tritz (17) have isolated from nadC mutants of E. coli a metabolite that is capable of supporting the growth of both nadA and nadB mutants but is incapable of supporting the growth of nadC mutants (QA phosphoribosyltransferase [QAPRTase] deficient). This metabolite has not been characterized completely, but it is doubtful that it is the hypothesized intermediate represented as [X] in Fig. 3 due to the lack of differentiation in the support of growth of nadA and nadB mutants. Tritz theorizes that this compound is either the immediate precursor of QA which cyclizes nonenzymatically or an intermediate in an alternate pathway leading to QA. Kerr and Tritz (61) have also shown that some, but not all, Nad- mutants in each class (nadA, nadB, nadC, and nadR) can grow in a vitamin-free Casamino Acids medium without nicotinic acid (NA). The remote possibility exists that these "dichotomistic" mutants are capable of utilizing an alternate pathway when stimulated by certain amino acids. The compound isolated by Chen and Tritz (17) could be an intermediate in this alternate pathway. More recently, these investigators have presented evidence which suggests that up to six intermediates may be involved in the formation of QA from L-aspartate and DHAP. Experiments were conducted in which the QA synthetase system was provided with either [14C]aspartate or [14C]fructose-1,6-diphosphate (as a source of DHAP) and the reaction mixtures were chromatographed. Autoradiograms of each chromatogram revealed a common spot, providing evi-

85

NAD BIOSYNTHESIS AND PNC METABOLISM

VOL. 44, 1980

TABLE 1. Enzymes involved in anaerobic NAD biosynthesis and PNC metabolisma Map position Reaction name loGenetic S E| C no.'uru Name no. | Alternate E. coli EC Name

Enzyme nomenclatureb

o QA + L-Asp + DHAP + FADH3P04 + 2H20 + FADH2

QA synthetase

QAPRTase

Mg2+ NAMN + PPi + 2.4.2.19 QA + PRPPC02

NAMN

Mg2+. deNAD + PPi 2.7.7.18 NAMN + ATP -

Mg2+ NMN + ATP -

adenylyltransferase NAD synthetase NAD (NADH) pyrophosphatase

NAD glycohydrolase

6.3.5.1

3.6.1.22 NAD+H20 mNADP + H20 m-

13.2.2.5

17 58

16 55

nadC

3

2

39

deNAD pyrophosphorylase

deNAD + ATP + glutamine (NH3) Mg2+M- NAD + PPi + AMP

NAD + H20 -Mg

NMN glycohydrolase 3.2.2.00 NMN + H20 phosphate NMN

NAD + PPi

nadA nadB

3.5.1.00 NMN + H20

-*

AMP + NMN , 2',5'-AMP + NMN

2+ I

NAD nucleosidase NAm + ADPribose

Absent

NAm + ribose-5NAMN + NH3

NMN deamidase

pncC

NA + NH3

NAm deamidase

pncA

27

Mg2+- NAMN + NAMN pncB PPi + ADP + P pyrophosphorylase

25

amidohydrolase NAm

3.5.1.19 NAm + H20 -

amidohydrolase NAPRTase

2.4.2.11 NA + PRPP + ATP

NAD kinase

2.7.1.23 NAD +

NADP phosphatase

Mg2+ 3.1.2.00 NADP + H20 -. NAD + H3P04

DNA ligase

6.5.1.2

NADH kinase

ATPMg2+

NAD + nicked DNA NADH + ATP

-e

NADP + ADP

-e

lg

, NMN + AMP + DNA

, NADPH +

51

Absent ADP

NAmPRTase

2.4.2.12 NAm + PRPP + ATP -NMN + PP + ADP + Pi

NMN pyrophosphorylase

2.7.7.1 NMN + PRPP + ATP - NAD + PPi NAD + ADP + Pi pyrophosphorylase adenylyltransferase a Abbreviations: PRPP, phosphoribosyl pyrophosphate; PPi, inorganic pyrophosphate; AMP, adenosine 5'-monophosphate; ADP, adenosine 5'-diphosphate; Pi, inorganic phosphate; Asp, aspartate; DHAP, dihydroxyacetone phosphate; QA, quinolinic acid; NA, nicotinic acid; NAm, nicotinamide; NAMN, nicotinic acid mononucleotide; NMN, nicotinamide mononucleotide;

NMN

deNAD, desamido-nicotinamide adenine dinucleotide or nicotinic acid adenine dinucleotide; NAD, nicotinamide adenine

dinucleotide; NADP, nicotinamide adenine dinucleotide phosphate; NADPH, reduced nicotinamide adenine dinucleotide phosphate; QAPRTase, quinolinic acid phosphoribosyltransferase; NAPRTase, nicotinic acid phosphoribosyltransferase; DNA, deoxyribonucleic acid. 'Recommended names and Enzyme Commission (EC) numbers are taken from reference 53a.

dence for the formation of a compound (or com- tion with partially purified material revealed pounds) labeled by both aspartate and fructose- that the reaction mixture contained six labeled 1,6-diphosphate (18). Subsequent experimenta- spots when ["4C]aspartate was used, whereas

86

FOSTER AND MOAT

MICROBIOL. REV.

TABLE 2. Enzymes involved in the aerobic tryptophan catabolic pathway to NAD Enzyme nomenclaturea

Alternate name(s)

Reaction

EC no.

Name

L-Tryptophan + 02 L-formylkynurenine 3.5.1.9 N-Formyl-L-kynurenine + H20 formate + L-kynurenine 1.14.13.9 L-Kynurenine + NADPH + 02 3-hydroxy-L-kynurenine + NADP+ + H20 3-Hydroxykynureninase 3-Hydroxykynurenine 3-hydroxyanthranilate + alanine 3-Hydroxyanthranilate + 02-s 2-amino3-Hydroxyanthranilate 1.13.11.6 3-Hydroxyanthranilate 3,4-dioxygenase oxygenase 3-carboxymuconate semialdehyde Recommended names and Enzyme Commission (EC) numbers are taken from reference 53a. 1.13.11.11 Tryptophan pyrrolase,

Tryptophan 2,3dioxygenase Kynurenine formamidase Kynurenine 3monooxygenase

tryptophan oxygenase Formylase, formylkynureninase Kynurenine 3-hydroxylase

a

AEROBIC

ANAEROBIC

A-

DHAP + ASPARTATE FORMATE + ASPARTATE

TRYPTOPHAN

QxIA

-

aCOOH NA OA

t%NFCOOH

OOH

3COOH

COOH

_IAM

NA t

(i)

/

NA

~NAN

i3

\

CYONH2 K.Pp NMN

t3CONHI N NAm

Ar

COOH

,~.CONH2

-P-P-R

NAm

deNAD

tONH, (f

(CONHt

Ad --

NMN NAD FIG. 2. Pathways of NAD biosynthesis and the four-, five-, and six-step pyridine nucleotide cycles. X refers to the unknown intermediate(s) in the anaerobic biosynthetic pathways to QA. Abbreviations: R, ribose; P, phosphate; Ad, adenine. four spots observed when [ "C]fructose- 1,6before it condenses with DHAP. Furthermore, diphosphate was used. Interestingly, the four syntheses of all six compounds were inhibited by labeled spots derived from ["4C]fructose-1,6-di- the addition of NAD to the reaction mixture phosphate were identical to four of the labeled (18). Gholson and his co-workers have extensively spots derived from ["4C]aspartate, indicating that two modifications of aspartate may occur examined the enzymes involved with QA biosynwere

VOL. 44, 1980


thesis in E. coli. They have confirmed the requirement for two proteins: (i) the "A" protein, coded for by the nadA gene, and (ii) the "B" protein, the product of the nadB gene (107). They have separated the nadB gene product from the nadA gene product via sodium citrate fractionation and have purified it by using various column chromatographic techniques (41). From the ease of separation of the two protein components, it was concluded that QA synthetase, if it exists as a complex, is not a tightly bound complex. Furthermore, a number of findings were presented which suggested that the nadB protein converts aspartate to an intermediate capable of undergoing condensation with DHAP catalyzed by the nadA protein. The B protein requires flavin adenine dinucleotide for activity. Therefore, the intermediate formed could be an unstable dehydrogenation product of aspartate which might be difficult to isolate and identify. Recently there has been some question as to whether or not the DHAP-aspartate pathway is truly an anaerobic pathway. S. Nasu, F. D. Wicks, and R. K. Gholson (Fed. Proc. 38:644, 1979) have shown that the nadB component of the QA synthetase system exhibits a strict requirement for 02 when assayed in vitro. The suggestion has been made that under anaerobic conditions E. coli, and presumably S. typhimurium, would switch to a true anaerobic pathway, such as the N-formylaspartate pathway (see below). If this were true, however, one would expect nadB mutants to lose their requirement for NA under anaerobic conditions, since an alternate pathway could be used. This has not been demonstrated. Another possibility, however, is that under anaerobic conditions the nadB gene product utilizes an electron acceptor other than oxygen. This hypothesis predicts that a nadB

HICO®§) COOH

87

mutant will retain its Nad- phenotype under anaerobic conditions. Recently, 10 nadB mutants and several nadA mutants of S. typhimurium were tested for the Nad- phenotype under anaerobic conditions. All retained their requirement for NA (J. W. Foster, unpublished data). Thus, for the purpose of this review we will present the DHAP-aspartate pathway as one which can function anaerobically. Formate-aspartate pathway. The grampositive anaerobe Clostridium butylicum utilizes a de novo pathway distinct from that described above. Research with this species has revealed the precursors of QA to be L-aspartate, formate, and acetyl coenzyme A (54, 96, 97). Aspartate and formate are condensed to form N-formylaspartate, which subsequently is condensed with acetyl coenzyme A to form QA (Fig. 4). No further investigations have been performed to determine the nature of the intermediate(s) in the C. butylicum pathway. Aerobic Tryptophan Catabolic Pathway The role of tryptophan as a precursor in eucaryotic NAD biosynthesis was first suggested by nutritional studies in which humans stricken with pellagra, an NA deficiency disease, recovered from this illness after the addition of tryptophan or niacin (NA) to their diets (66). Other studies established tryptophan as a precursor of NAD in many animal and plant systems (23). In 1952, Yanofsky and his colleagues used the mold Neurospora to elucidate some of the intermediate reactions in the synthesis of NAD from tryptophan (84). The enzymatic steps were soon clarified through the use of radioisotope tracer experiments and by the isolation of the various intermediates involved. Figure 5 outlines the degradation of tryptophan and subsequent syn-

QA synthetase

noB O=C HCH nadA no'N,OO s''c CO HCO COOH A-§ H2COH CH FAD FADH2 pi Hi2c ';kN -COOHPR N HpN COOH

DHAP ASPARTATE QUINOLINATE NAMN FIG. 3. DHAP-L -aspartate pathway for QA biosynthesis. The genetic loci involved are designated nadA, nadB, and nadC. X refers to the unknown intermediate(s) involved in the biosynthesis of QA. Abbreviations: ®, phosphate; FAD, flavin adenine dinucleotide; FADH2, reduced FAD; P, inorganic phosphate; PRPP, phosphoribosylpyrophosphate, R, ribose.

HCOOH +

COOH /COOH CH2

C/COOH

CM2

CH3CO-CoA

COOH ill

I X CH HCO CHHN/ACOOH \ /\N OH HN COOH H2N COOH FORMATE ASPARTATE N-FORMYLASPARTATE QUINOLINATE FIG. 4. Formate-aspartate pathways for QA biosynthesis. CoA, Coenzyme A.

88

MICROBIOL. REV.

FOSTER AND MOAT 2 1 Fe~ NlCOCH2CHCOOH -CO ~-NH CHO CO 0 FORMYLKYNURENINE

~s%fNyCOCH2CHCOOH

rir-iI-CH2CHCOOH

NH2~

~

~

~

N.)NH2NH KYNURENI NE N

H TRYPTOPHAN

02,NADPH 3 1

OCH2C H COOH

OH 3-HYDROXYKYNURENINE PA -ALANINE 4 rs1COOH

L

COOH

C OOH 0"H2N H 6 /2-ACROLEYL-3

o

02

NAPH

NH2

OH

5

AMINO FUMARATE

3-HYDROXYANTHRANILATE

(COOH COOH

QUINOLINATE PRPP

|-C2

7 N-'O (+

T I R- P-P- -ADENINE

NADP+

8

4H2

)I _ ATPP INz GLUTAMINE R®®.R-ADENINE DEAMIDO-NAD

R

10

-®R-ADENINE

NAD+

9 FIG. 5. Aerobic tryptophan catabolic pathway to NAD (adapted from Moat [77]). Enzyme designations are as follows: 1, tryptophan 2,3-dioxygenase; 2, kynurenine formamidase; 3, kynurenine 3-monooxygenase; 4, 3hydroxykynureninase; 5, 3-hydroxy-anthranilate 3,4-dioxygenase; 6, spontaneous reaction; 7, QAPRTase; 8, De-NAD pyrophosphorylase; 9, NAD synthetase; 10, NAD kinase. Abbreviations: PALP, pyridoxal 5'phosphate; PRPP, phosphoribosyl pyrophosphate; R, ribose; (), phosphate.

thesis of NAD. The biosynthesis of NAD from tryptophan in mammalian cells and lower eucaryotes is now well documented as to the nature of the intermediates involved and the enzymes used to catalyze the individual reactions. Most of the references to this earlier work are included in the reviews by Chaykin (16) and Dalgliesh (23) and will not be reviewed here. The pathway of tryptophan catabolism is described as aerobic due to the strict oxygen requirements of the first enzyme, tryptophan ox-

ygenase, and the third enzyme, kynurenine 3hydroxylase (95). Remarkably, tryptophan oxygenase has been very difficult to demonstrate in extracts of several lower eucaryotes known to synthesize NAD from tryptophan. For instance, this enzyme defied detection in Neurospora until 1973, when Casciano and Gaertner finally used a fluorometric assay system which involved a kynurenine formamidase-kynureninase-coupled reaction (11). In microorganisms which do not synthesize NAD from tryptophan but do

VOL. 44, 1980)


actively catabolize tryptophan, tryptophan oxygenase is readily demonstrable as well as inducible by tryptophan or a tryptophan metabolite (82, 88). There are a number of similarities between the pathway from tryptophan to NAD and the aromatic pathway involved with the catabolism of tryptophan (74). For example, the first two enzymatic reactions are identical. Furthermore, kynureninase, of the catabolic pathway, and hydroxykynureninase, of the NAD pathway, have similar substrate specificities. The possibility that the tryptophan-NAD pathway evolved from the tryptophan catabolic pathway is discussed in Evolutionary Aspects. Evidence is accumulating which also implicates the aerobic NAD pathway in the biosynthesis of actinomycin D by Streptomyces parvullus and Streptomyces antibioticus (46, 60). The following microorganisms have been shown to use the tryptophan-NAD pathway: X. pruni (10), Neurospora crassa (69), aerobically grown Saccharomyces cerevisiae (1), and S. antibioticus (70). One very interesting aspect of de novo NAD biosynthetic pathways is that they all lead to the formation of a common intermediate, QA (16, 38, 43, 85, 101). Furthermore, the series of steps involved in the conversion of QA to NAD are identical in all of the species which have been examined (Fig. 2). Preiss and Handler, in 1958 (85, 86), studied the pathway from NA to NAD in human erythrocytes and yeast. They isolated the intermediates and identified the enzymes involved by using a combination of radioisotopic labeling and biochemical techniques. The reaction sequence they discovered is referred to as the Preiss-Handler pathway and occurs as follows: NA -- NAMN -- deNAD -* NAD Andreoli et al. (4) showed that in E. coli, QA, not NA, is the precursor of NA mononucleotide (NAMN) on the de novo pathway to NAD. QA is converted to NAMN by means of the phosphoribosyl pyrophosphate-dependent enzyme QAPRTase. NA is now known to be part of the pyridine nucleotide cycle (PNC) which recycles a variety of degradative products of NAD metabolism back to NAD (34). Biosynthetic routes which lead to the synthesis of NAD do so by supplying QA to the PNC (Fig. 2). Dahmen et al. (22) have studied purified desamido-NAD (deNAD) pyrophosphorylase from brewer's yeast and E. coli. Their studies show that, in addition to catalyzing the NAMN-todeNAD reaction, this enzyme can use nicotinamide mononucleotide (NMN) as a substrate to synthesize NAD in vitro. The yeast enzyme synthesized deNAD 2.2 times faster than it synthesized NAD. The enzyme from E. coli, however,

89

had a rate of deNAD synthesis 17 times faster than that of NAD. They concluded that, at least in E. coli, the reactivity of deNAD pyrophosphorylase with NMN in vivo is not physiologically significant. The final step of NAD biosynthesis in either the de novo pathway or the PNC is the amidation of deNAD to form NAD. The enzyme catalyzing this step is NAD synthetase. Preiss and Handler have shown that amidation of deNAD in yeast involves L-glutamine as the amide donor (86), whereas E. coli preferentially uses free NH3 in this reaction (100).

Genetics of Anaerobic Nicotinamide Adenine Dinucleotide Biosynthesis The preponderance of information regarding the genetics of anaerobic NAD biosynthesis comes from work with E. coli. Mutants blocked in the de novo biosynthetic pathway before the PNC may be isolated as NA-requiring auxotrophs (Fig. 6). Three classes of NA-requiring mutants (relevant loci originally designated nic) were discovered in E. coli K-12 and S. typhimurium LT-2. Originally, classification of these loci was based upon mapping data. The genes were later designated nadA, nadB, and nadC to reflect their roles in the biosynthetic pathway to NAD. The genetic map positions of these loci (Fig. 6) are 16, 55, and 2 min, respectively, in E. coli (7, 111, 112) and 17, 58, and 3 units, respectively, in S. typhimurium (28, 67, 90, 102, 103). Chandler and Gholson (12), in 1972, demonstrated the excretion of QA and the absence of QAPRTase in E. coli nadC mutants (78). Both nadA and nadB mutants failed to excrete QA but did possess QAPRTase activity. Extracts from nadC mutants were capable of synthesizing QA from the precursors DHAP and aspartate, whereas extracts from nadA and nadB mutants lacked this capability (12, 14, 15). Extracts from nadA and nadB mutants were able to complement each other in vitro, resulting in the synthesis of QA (110). Therefore, the nadA and nadB genes encompass what is referred to as the QA synthetase system, and nadC is the structural gene for QAPRTase. Though less extensive, the work by Foster and Moat (28) provides evidence for similar gene designations in S. typhimurium. Kerr and Tritz (61), through liquid matrix cross-feeding experiments with E. coli mutants, determined nadA to be the first enzyme of the QA synthetase system and nadB to be the second. Wicks et al. (115) have performed more detailed experiments which suggest that the functional order of nadA and nadB is reversed. They found that toluene-treated nadB cells incubated with supernatant culture fluid from

90

MICROBIOL. REV.

FOSTER AND MOAT

/

S. typhimurium markers

80o

E. co/i

markers

>

9° 1 ch

Xi

20

o~~~~~~temnato

'4~~~~~~

FIG. 6. Genetic linkage maps of E. coli and S. typhimurium, showing the genetic loci involved in NAD biosynthesis and PNC metabolism. Locations of the E. coli markers are shown on the inside of the circle, and those of S. typhimurium are shown on the outside of the circle. The darkened area represents the suggested inverted region.

nadA cells could synthesize ["C]QA from ["4C]aspartate. The reverse situation, in which toluene-treated nadA cells were incubated with supernatant from nadB cells, proved unsuitable for QA synthesis. Thus, nadA cells produced some diffusible compound or compounds which nadB cells could use to synthesize QA. Consequently, the nadB gene product is the first enzyme of the QA synthetase system and the nadA gene product is the second (Fig. 3). Tritz and Chandler (110), while performing in vitro complementation assays between extracts of various E. coli Nad- mutants, discovered a gene involved in the regulation of NAD biosynthesis. The procedure involved mixing extracts from two different Nad- auxotrophs and observing whether QA was synthesized or not. When extracts of nadA mutants were combined with extracts from nadB mutants (nadA and nadB designations were originally based on mapping data), the nadB mutants fell into two separate categories: (i) those that would complement the nadA extracts and (ii) those that would not. The

second group appeared to be regulatory mutants in that they expressed neither the nadA nor the nadB gene product. The gene involved was therefore designated nadR. This gene could not be separated from nadB by transductional analysis, and abortive transduction experiments were never performed (G. J. Tritz, personal communication). Therefore, the status of nadR as a separate cistron remains to be proven. Tritz (109) attempted to examine the nature of the nadR locus by performing merodiploid analysis. Merodiploids with the genotype nadR+/nadR lost their requirement for NA, thus indicating positive regulation by the nadR gene product. The model presented in Fig. 7 attempts to illustrate positive control while maintaining the basic concept that the addition of end product results in the repression of enzyme synthesis. Although one might expect some form of negative control (see Regulation of De Novo Biosynthesis of Nicotinamide Adenine Dinucleotide), this model is feasible. Griffith et al. (41) report, however, that the nadR strain used by Tritz was


VOL. 44, 1980 nodS

nodR

nodA

nad5

nadR

nodA

91

Diffusible product + Corepressor

(probably NAD) Inactive regulotor protein FIG. 7. Postulated mechanism for the positive regulation of nadA and nadB by the nadR gene product. (A) A diffusible product of the nadR gene would serve to promote or induce transcription of the unlinked nadAB operon. (B) However, in the presence of the corepressor, most probably NAD, this diffusible product would be incapable of "turning on" the expression of nadAB. mRNA, Messenger ribonucleic acid.

QA excreted by nadC mutants incubated in a special secretion medium was directly related to the presence or absence of NA in the preparatory growth medium. The accumulated data definitely indicate a repressive type of regulation over the QA synthetase system. The QA synthetase system may be subject to feedback inhibition as well as end product repression. In vitro assays with crude (13) and partially purified (41) preparations indicate that concentrations of oxidized NAD (NAD+) and reduced NAD (NADH) which approximate intracellular conditions (3.33 mM) strongly inhibit QA synthetase activity. Since equivalent levels of NADP+ Regulation of De Novo Biosynthesis of and NADPH did not have as dramatic an effect, Nicotinamide Adenine Dinucleotide these compounds do not seem to play a substanNAD, albeit an essential cofactor, is required tial role in the modulation of QA synthetase by the cell in extremely small quantities. Lund- activity. In E. coli, QAPRTase does not appear quist and Olivera (71) have determined that the to be under repression-derepression control or intracellular concentrations of NAD and NADP feedback inhibition (35, 93). In contrast, during exponential growth of E. coli are 1.3 x QAPRTase synthesis is repressible in Bacillus 10-3 and 3.9 x 1O-4 M, respectively. Thus, strict subtilis (35, 93). Tritz and Chandler (110), in 1973, reported control must be exerted by the cell to maintain the presence in E. coli of a gene involved in the these levels. Regulation of de novo biosynthesis of QA was regulation of NAD biosynthesis. Extracts from first studied by Saxton et al. (93), in 1968. They certain putative nadB mutants of E. coli failed reported that extracts prepared from E. coli to complement extracts of nadA mutants for in grown in the presence of NA exhibited a de- vitro synthesis of QA. These particular nadB creased capacity for QA synthesis, indicating a mutants were redesignated nadR by virtue of repression of the QA synthetase system by NA the fact that the lesion appeared to affect the or a product formed from NA. In comparison, synthesis of both nadB and nadA gene products. NA added to the growth medium had no effect Merodiploid analysis of these regulatory muon QAPRTase activity. Chandler and Gholson tants by Tritz (109; see Genetics of Anaerobic (12) further demonstrated that the quantity of Nicotinamide Adenine Dinucleotide Biosyn-

actually a nadB mutant, not a nadR mutant. Therefore, results obtained by using a merodiploid with an actual genotype of nadR+/nadR require confirmation. Griffith et al. suggest that there may not be a separate nadR gene. The regulatory function may be part of the nadB protein. Abortive transduction experiments performed between numerous nadB mutants of S. typhimurium (some of which appear to be nadR) suggest the presence of a single gene at this locus (J. W. Foster, D. M. Kinney, and A. G. Moat; unpublished data).

92

FOSTER AND MOAT

thesis) appeared to show positive regulation of NAD biosynthesis by a diffusible product of the nadR gene. Griffith et al. (41), however, reported that the nadR mutant used in the study by Tritz (109) was actually a nadB mutant. Thus, the nature of the nadR locus and whether it is indeed distinct from nadB remain to be clarified. In their discussion, Griffith et al. (41) mention that the nadB protein binds to an NAD-agarose affinity column. This observation suggests that the nadB protein may possess regulatory features consistent with those of an aporepressor. An autoregulatory type of control system where the product of the nadB cistron serves to regulate the expression of the noncontiguous nadAB operon seems reasonable. Thus, nadR mutants may represent a specific class of nadB mutations. However, QA synthetase also exhibits feedback inhibition (41, 93), and binding of NAD to the nadB protein may simply reflect this fact. Experiments designed to determine whether a nadB protein isolated from an alleged nadR strain can bind to the aforementioned NADagarose column could prove useful in assessing the validity of the autoregulation theory. X. pruni utilizes the tryptophan pathway for NAD biosynthesis, as mentioned above. Brown and Wagner (10) studied the regulation of the tryptophan pathway in this species. They observed coordinate induction of tryptophan pyrrolase (tryptophan oxygenase), kynurenine formamidase, and kynureninase by L-tryptophan. L-Kynurenine was not effective as an inducer, which suggests that the effect is specifically due to L-tryptophan. Subsequent work by these authors revealed that tryptophan pyrrolase is feedback inhibited by NADH and NADPH as well as by NAMN and anthranilic acid (114). It is of interest that the oxidized forms of the coenzymes, NAD+ and NADP+, were ineffective as feedback inhibitors of this enzyme. The other enzymes of the tryptophan-NAD pathway were present, but no evaluation regarding their regulation was made (10). Remarkably, in this report evidence is presented which suggests the presence of a tryptophan-NAD pathway that utilizes anthranilic acid as an intermediate. NA-requiring mutants were isolated which could not grow on tryptophan but could substitute anthranilate for the NA requirement. The existence of an alternate pathway involving anthranilic acid may explain why this compound can inhibit tryptophan pyrrolase activity. It should be stressed that no enzymological evidence exists for this pathway as yet. N. crassa also uses the tryptophan-NAD pathway. Some studies have been conducted on the regulation of this system. Lester (69), in

MICROBIOL. REV.

1971, suggested that tryptophan oxygenase may be the rate-limiting step for NAD biosynthesis in N. crassa. He demonstrated repression and feedback inhibition (9, 69) by nicotinamide (NAm) or, more likely, NAD. One question which has persisted throughout the literature is whether or not kynureninase, whose substrates include hydroxykynurenine, can substitute for a hydroxykynureninase in NAD biosynthesis. Gaertner et al. (32) have demonstrated a hydroxykynureninase activity distinct from kynureninase. The hydroxykynureninase was shown to be constitutive in nature, whereas the kynureninase was inducible. Since a hydroxykynureninase would be involved with the synthesis of an essential cofactor, NAD, the presence of a constitutive enzyme would seem desirable. On the other hand, a constitutive kynureninase, which is involved only with the catabolism of tryptophan, would appear wasteful. S. cerevisiae is an interesting species in that it possesses both an anaerobic pathway and an aerobic pathway for NAD biosynthesis (1). Growth under aerobic conditions greatly favors the aerobic production of NAD from tryptophan over the anaerobic route. One explanation for this phenomenon has been offered by Schott et al. (94). They have demonstrated an apparent induction of L-kynurenine 3-hydroxylase by oxygen, which is not observed in other systems, such as that of N. crassa (95). Growth of yeast under anaerobic conditions yielded extracts which possessed very little of this activity (0.026 mU/mg of protein). A 10-fold increase in L-kynurenine 3-hydroxylase activity occurred when cells were grown aerobically. Work by Heilmann and Lingens (45) shows that 3-hydroxyanthranilate oxygenase of S. cerevisiae is not repressed or feedback inhibited by NAD or induced by tryptophan. Substrate concentrations higher than 3 x 10-4 M, however, inhibited the activity of this enzyme. Thus, these two activities may serve a major regulatory function in preventing an oversupply of NAD under aerobic conditions. Ahmad and Moat (1) have shown that more NAD is synthesized aerobically than anaerobically. Thus, induction of L-kynurenine 3-hydroxylase by oxygen would lead to an increase in hydroxyanthranilate and NAD levels, assuming kynurenine 3-hydroxylase is a rate-limiting step in this pathway. Too high a level of 3-hydroxyanthranilate would inhibit the activity of 3hydroxyanthranilate oxygenase, thus decreasing NAD biosynthesis. Regarding the question of kynureninase versus hydroxykynureninase activities and their roles in NAD biosynthesis, Shetty and Gaertner (98) have discovered that S. cerevisiae only has

VOL. 44, 1980

93


a constitutive hydroxykynureninase. This species does not appear to degrade tryptophan via the tryptophan-anthranilate cycle and therefore does not possess a kynureninase. This is in contrast to N. crassa, which has both an inducible kynureninase and a constitutive hydroxykynureninase.

PYRIDINE NUCLEOTIDE CYCLE METABOLISM Biochemistry Recycling pathways, collectively known as PNCs, function in a majority of the species studied. The basic outline for these cycles includes the degradation of NAD to NAm, the conversion of NAm to NA, and finally the synthesis of NAMN, an intermediate in the de novo pathway to NAD, from NA. Cells that possess an intact PNC, and many cells that do not, use the Preiss-Handler pathway to synthesize NAD from NA (52,85,86). Similarly, NAm deamidase, whose enzymatic function is to convert NAm to NA, is also present in all cells that can completely recycle NAD (52, 55, 92, 99, 106) through NA. The pathway of degradation from NAD to NAm, however, varies among different species. A majority of the eucaryotic species examined contain NAD glycohydrolase (2, 40, 57, 92, 108, 120), which catalyzes the following reaction: NAD+-- NAm + adenosine 5'-diphosphate-ribose Figure 8 illustrates the five-membered PNC (PNC V) which uses this enzymatic activity to cleave NAD for recycling. NAD glycohydrolase is relatively rare in procaryotic systems. An extensive search was made for NAD glycohydrolase in E. coli, but no such activity has been found (5). Species that lack NAD glycohydrolase appear to recycle NAD by first degrading the cofactor to NMN and subsequently converting the NMN to NAm. There has been some controversy concerning which enzyme or enzymes are used by E. coli to generate NMN from NAD. Procaryotic deoxyribonucleic acid ligase (68, 80), which repairs single-stranded deoxyribonucleic acid nicks, uses NAD as a substrate in the reaction shown in Fig. 9. Manlapaz-Fernandez and Olivera suggest a major role for this enzyme in NAD turnover by E. coli (75). They report, as unpublished data, a reduced turnover of NAD in a temperature-sensitive deoxyribonucleic acid ligase mutant. However, E. coli, yeast (22, 50), and apparently S. typhimurium (26) also contain NAD pyrophosphatase which can also degrade NAD to NMN and adenosine 5'-monophosphate. These two enzyme activities probably

QA -NAMN--+deNAD - NAD

I/

IPNCVY N A,(

NAm FIG. 8. PNC V.

1111111

1111111

l I 1111L 1 0O ° , J %H

NAm R

I

l

Ad R

1 11 11 1 1

T1 1 11 1

NAm I

Ad I

FIG. 9. Repair reaction catalyzed by deoxyribonucleic acid ligase. Abbreviations: R, ribose; (D), phosphate; Ad, adenine.

account for the majority of NAD hydrolysis in E. coli and presumably S. typhimurium. An enzyme has been discovered in E. coli (5), Azotobacter vinelandii (49), and S. typhimurium (J. W. Foster and A. G. Moat, unpublished data) which specifically converts NMN to NAm. This enzyme, NMN glycohydrolase, occurs both intracellularly and membrane bound in E. coli (5). The enzyme from A. vinelandii has been purified and studied quite extensively by Imai (49). His results show a pH optimum of between 8.5 and 9.0 and a temperature optimum of 39°C. Data presented also reveal activation of this NMN glycohydrolase by guanosine 5'-triphosphate, deoxyguanosine 5'-triphosphate, and guanosine 5'-tetraphosphate, with KA values of 0.025, 0.080, and 0.020 mM, respectively. The other procaryotic NMN glycohydrolases have not been examined as comprehensively, but they do not appear to exhibit this activation by guanosine nucleotides. NAD pyrophosphatase and NMN glycohy-

94

FOSTER AND MOAT

drolase, in essence, replace NAD glycohydrolase in the overall functioning of the PNC in most procaryotic species. Figure 10 illustrates the sixmembered PNC (PNC VI), which includes the aforementioned enzymatic activities. Mutants of E. coli deficient in the de novo biosynthesis of NAD are capable of growing on media supplemented with either NA or NAm. NA supports growth via the Preiss-Handler pathway, and NAm supports growth through conversion, by NAm deamidase, to NA (105). Sundaram et al. (106) proved this by isolating a Nad- mutant that would grow on NA but failed to grow on NAm. Analysis of cell-free extracts revealed that this mutant lacked NAm deamidase activity. The first direct evidence for a functional PNC in vivo came from Andreoli et al. (3) in 1969. A mutant of E. coli lacking NAm deamidase, after incubation with ['4C]NA, accumulated ['4C]NAm in the culture medium. They proposed that the added NA was metabolized to NAD, which was subsequently degraded to NAm. Since the NAm could not be recycled, it was excreted into the culture medium. Control experiments, using a strain with an intact PNC, did not accumulate any ['4C]NAm. ManlapazFernandez and Olivera (75) further examined the recycling of NAD by E. coli by means of pulse-chase labeling experiments with differentially labeled NA and adenine. Cells grown in the presence of ['4C]adenine and [3H]NA were transferred to cold medium, and the "4C-3H ratio of intracellular NAD was measured at various time intervals. Wild-type E. coli displayed a decrease in the '4C-3H ratio, indicating the preferential loss of adenine label due to NAD turnover. They also observed almost total conservation of the pyridine moiety in NAD during this recycling process. On the other hand, a mutant lacking an intact PNC VI as a result of NAm deamidase deficiency also exhibited a preferential loss of adenine label compared with niacin (NA) label from the pyridine nucleotide pool. The conclusion drawn form this experiment was that an alternate PNC, which does not utilize NAm as an intermediate, must exist in E. coli. Evidence for a similar alternative PNC in S. typhimurium was first provided when a mutant blocked in the de novo pathway at nadA and in PNC VI at NAm deamidase (pncA) grew well on exogenously supplied NMN. An alternate PNC had already been discovered in Clostridium sticklandii (31), A. vinelandii (48), and Propionibacterium shermanii (Propionibacterium freudenreichii subsp. shermanii) (29). The PNC in these species involves NMN deamidase, which specifically deamidates NMN to form NAMN and

MICROBIOL. REV.

therefore bypasses NAm and NA (PNC IV [Fig. 11]). A similar activity has recently been reported in S. typhimurium (62) and E. coli (75). The enzyme from S. typhimurium seems to differ quite dramatically from that of P. shermanii (30) and A. vinelandii (48). The enzymes from P. shermanii and A. vinelandii exhibit classic Michaelis-Menten kinetics and pH optima of 5.6 and 7.0, respectively. The Salmonella NMN deamidase seems to exhibit sigmoid kinetics suggestive of allosterism and has a pH optimum of 8.7. The purified enzyme from E. coli has a pH optimum of 9.0 and appears to exhibit linear kinetics (B. M. Olivera, D. Hillyard, P. Manlapaz-Ramos, J. Imperial, and L. J. Cruz, personal communication). Substantial evidence is accumulating which shows that both PNC IV and PNC VI appear to function in vivo, with PNC IV being the predominant PNC. A strain of E. coli was found to recycle NAD 72% via PNC IV and 28% via PNC VI (75). Similarly, several strains of S. typhimurium tested revealed relative contributions toward NAD turnover of 60 to 69% for PNC IV and 31 to 40% for PNC VI (J. W. Foster and A. M. Baskowsky-Foster, submitted for publication). Although the discovery of NMN deamidase in both S. typhimurium and E. coli proves there are two PNCs, there is evidence in the literature which appears to contradict this finding. Gholson et al. (37) have reported that a mutant blocked in the de novo pathway at nadB was capable of growing in media supplemented with NAD, whereas a mutant blocked in both the de novo pathway at nadB and in the PNC at NAm deamidase was incapable of growing in the same media. One would have expected the double mutant to grow on NAD if the alternate, PNC IV, pathway existed in vivo. Recent studies with S. typhimurium may have resolved this paradox. QA- NAMN

)

deNAD-

NAD

PNCVI

j

NA4- NAm( -NMN FIG. 10. PNC VI.

QA-NAMN -*deNAD -

NAD

PNC IV

NMN FIG. 11. PNC IV.


VOL. 44, 1980

Differentially labeled NAD transport experiments, using nadA and nadA pncA mutants, revealed that even though only nadA mutants took up the pyridine moiety of NAD, both nadA and nadA pncA cultures degraded NAD at the cell surface to NMN and adenosine 5'-monophosphate and accumulated these products extracellularly (26). The data suggest that the NMN resulting from NAD degradation is readily transported by NMN glycohydrolase and, in the process, converted to NAm. The fact that NMN glycohydrolase is associated with the membrane of E. coli (5) tends to support this theory. For NMN to traverse the membrane intact, it must apparently utilize a transport system other than NMN glycohydrolase. Effective utilization of this system seems to occur when exogenous NMN levels reach 10' M or more. Thus, one could predict that a large number of cells should be capable of degrading extracellular NAD to yield a concentration of NMN sufficient to permit entry by this alternate uptake system. Once intracellular, the NMN could be recycled via PNC IV and NMN deamidase. The result should be growth of the double mutant on NAD. This experiment was successfully performed with S. typhimurium (26). Subsequently, mutants unable to transport NMN via this alternate system have been isolated and designated pyridine nucleotide uptake mutants (pnuA). This gene has been mapped at 0 units via cotransduction with the thr gene (26, 62). A model for the utilization of exogenous NAD by S. typhimurium and E. coli is presented in Fig. 12. NADP is derived from NAD by an adenosine 5'-triphosphate (ATP)-dependent phosphorylation catalyzed by NAD kinase. The reaction proceeds as follows:

+ adenosine 5'-diphosphate

Studies with S. cerevisiae suggest there are three NAD kinases in this species. One is a cytoplasmic enzyme specific for NAD and inhibited by NADH; another is an NAD-specific mitochondrial enzyme; and the third, also mitochondrial, is specific for NADH (NADH kinase) (6). They appear to be separate and distinct enzymes, as evidenced by (i) three different rates of inactivation by heat and 3-(bromoacetyl)-pyridine, (ii) different Km values, (iii) different pH optima of activity, (iv) sigmoid Michaelis plots found with the mitochondrial kinases, and (v) inhibition of NAD kinase by high substrate concentrations. The control of NADP(H) synthesis is evidently very complex in this species. Part of the complex control could be in the form of allosteric modifiers, as suggested by the sigmoid kinetics of the mitochondrial kinases. Evidence indicating the presence of NAD kinase in E. coli has been presented by Imsande and Pardee (53) as well as Lundquist and Olivera (71). Although detailed enzymological studies were not performed, this enzyme seems to be inhibited by high ATP levels (53). Lundquist and Olivera (71) have presented data which suggest that the interconversion of NAD and NADP is important in the regulation of PNC metabolism. This aspect will be analyzed below (Regulation).

Pyridine Nucleotide Cycle Genetics Sundaram (105), in 1967, isolated a mutant of E. coli which lacked NAm deamidase. This enzyme converts NAm to NA and functions as an

NAm

N MN

NAD

NAD+ + ATP -- NADP+

PNC

dNAD

INSIDE NA

NAMN

FIG. 12. Model for utilization of exogenous NAD by S. typhimurium and E. coli. Abbreviations: AMP, adenosine 5'-monophosphate; dNAD, deNAD.

96

FOSTER AND MOAT

integral part of PNC VI. Penicillin selection procedures were used to isolate nadC mutants capable of growing on NA but incapable of growing on NAm. The mutation is located at 39 min on the E. coli linkage map (7, 24) and is designated pncA. Extracts from pncA mutants, as predicted, lacked NAm deamidase. SimilarpncA mutants were isolated in S. typhimurium on the basis of their resistance to the NAm analog 6amino-NAm. In contrast to E. coli, the pncA locus mapped at 27 units in this species (Fig. 4), between the gal and trp operons (26). This locus in E. coli resides between the trp and his operons. Thus, pncA appears to reside within the large inverted region described in the reviews by Sanderson and Hartman (90) and Riley and Anilionis (87). However, placement of pncA in S. typhimurium at 27 units suggests that the inversion extends beyond the 10 units proposed to nearly 14 units (Fig. 6). Further experimentation with mutants in this region is needed to confirm this theory. Pardee et al. (83), interested in how microconstitutive enzymes are maintained at low levels, isolated a mutant of E. coli capable of growing on NAm as the sole source of nitrogen. The mutant was a hyperproducer of NAm deamidase, and the relevant gene was designated pncH. This gene mapped at 39 min in E. coli and was cotransducible with pncA. The purpose behind the search for this type of mutant was to study the regulation of NAm deamidase production. Even though this enzyme is constitutive in nature, Pardee has postulated that enzymes produced in low quantities may have developed a system of regulation distinct from the classic induction-repression control mechanisms which utilize a diffusible regulatory element. Induction-repression regulation appears too costly for the cell when one considers the amount of enzyme produced. For example, wild-type E. coli produces NAm deamidase with a specific activity of 3 nmol of NAm hydrolyzed per min per mg of protein (24). This is a value several orders of magnitude below those of many enzymes involved in other major bacterial pathways. The pncH mutants produced enzyme with a specific activity up to 50 times normal levels. This was shown to be a true increase in the amount of enzyme produced and not simply a more active enzyme. Although the actual cause for this increased synthesis is not known, one can speculate that perhaps a promoter "up" mutation occurred which would allow for more frequent messenger ribonucleic acid initiation by ribonucleic acid polymerase. R. Lemmon, J. Rowe, and G. Tritz reported (Abstr. Annu. Meet. Am. Soc. Microbiol. 1977, K229, p. 224) that mutants resistant to the NA

MICROBIOL. REV.

analog 6-amino-NA lack NAPRTase activity. The relevant gene in these mutants was designated pncB. Attempts to map these mutants by conjugation have been frustrated by the high frequency of spontaneous resistance associated with this analog. First reports, however, placed pncB between 20 and 30 min on the E. coli linkage map. Foster et al. have isolated pncB mutants of S. typhimurium which also lack, or possess reduced levels of, NAPRTase (26). However, through cotransduction and mutagenesis they successfully introduced nadA and nadB mutations into these pncB mutants. This served to eliminate the problem of the development of spontaneous 6-amino-NA-resistant mutants during mapping procedures. The pncB locus has been mapped in this species via conjugation and was found to reside near 25 units (Fig. 6). Regulation NAD, as a cofactor, is used for catabolic energy-yielding oxidations, whereas NADP serves as the source of biosynthetic reducing power. Since NAD also serves as a substrate in cellular metabolism, it is reasonable to expect that the synthesis and breakdown of NAD and NADP are carefully regulated. Imsande, in the early 1960s, first demonstrated the regulation of one of the PNC enzymes in E. coli (50, 51, 53). The activity of NAPRTase in extracts prepared from cells grown in medium containing NA was 100fold lower than the activity found in extracts prepared from cells grown in NA-free medium. In contrast, this enzyme was not regulated by either repression or feedback inhibition in Serratia marcescens, B. subtilis, Torula cremoris, or Tetrahymenapyriformis (51). Presumably, in E. coli NA is converted via the Preiss-Handler pathway to NAD. NAD would then function in some capacity to repress transcription of the NAPRTase structural gene, pncB. NAD cannot be used in this type of study because exogenous NAD is not utilized directly by the cell, but must be degraded and recycled through the PNC (26, 37). Imsande (51) in 1964 and Foster et al. (26) in 1979 have measured NAPRTase levels in S. typhimurium. Repression of this enzyme does occur, but at a lower level than with E. coli. A detailed examination of NA concentration versus NAPRTase activity revealed that repression did not occur until the NA concentration was at least 7 x 10-7 M (27). Maximal repression occurred around 5 x 10' M NA. This study also revealed that in order to obtain repression with other intermediates of the PNC, they must be used at a concentration which will result in a nadA mutant generation time of 65 min or less.


VOL. 44, 198

Although the exact mechanism of repression has not been proven, there is evidence which implicates NAD as the true corepressor molecule. Experiments with labeled precursors in vivo result in most, if not all, of the label residing in NAD and NADP (26, 72; Foster and Baskowsky-Foster, submitted for publication). This finding suggests that intracellular levels of the other intermediates are too low to cause repression. Furthermore, studies with S. typhimurium have shown that although NAm and NMN will cause repression of NAPRTase in a wild-type strain, the introduction of a pncA mutation prevents this repression (26). These NAm deamidase mutants do not recycle exogenous NAm or NMN via PNC VI to NAD, suggesting that repression of NAPRTase by various PNC intermediates is the result of their conversion to NAD. The concentration of NMN used was not sufficient to produce a nadA pnecA mutant generation time of 65 min, which explains why there was no repression with this nucleotide even though it could be recycled by PNC IV. The possibility of feedback inhibition by NAD was explored by Imsande and Pardee in E. coli (53) and Foster et al. in S. typhimurium (26), but with negative results. In light of the results obtained with tryptophan oxygenase, where NADH but not NAD+ successfully inhibited activity (114), perhaps NADH should be examined for any effect it may have on NAPRTase activity. Other PNC enzymes of E. coli, when studied, failed to show any regulation. Pardee et al. (83) and Baecker et al. (8) found no change in NAm deamidase activity when E. coli was grown in media supplemented with NA, NAm, NAD, or NADP. deNAD pyrophosphorylase, NAD synthetase, and NAD kinase activities failed to show any decrease in activity when these enzymes were extracted from cells grown in 1o-4 M NA-supplemented minimal medium (53). These authors, therefore, suggested that NAPRTase is one of the key control elements of the PNC. Further evidence indicating the precise control exerted over NAD metabolism comes from Lundquist and Olivera (71). Exponentially grown E. coli maintained a steady-state balance between the levels of NAD and NADP. In addition, the breakdown of NADP to NAD was found to be one of the major processes which determines the relative levels of NAD and NADP. Lundquist and Olivera (71) present a model, based on labeling data, which represents this steady state as follows:

NAcell wall

NAD

R2

NADP

97

R. is the rate of fonnation of NAD from external

NA in each cell, R1 is the average rate of convesion of NAD to NADP per cell, and R2 is the average rate of breakdown of NADP to NAD per cell. Interestingly, other intermediates of the PNC are present in such infinitesimal amounts that the only pyridine compounds labeled after a ['4C]NA pulse-chase experiment are NAD and NADP. Regulation, therefore, appears to involve the interconversion of NAD and NADP (the normal ratio of NADP to NAD being calculated at 0.3) and the regulation of NAD synthesis through the repression of QA synthetase and NAPRTase. An imbalance in the ratio of NAD to NADP due to excess NAD would result in the repression of NAPRTase. Normal recycling of the NAD would occur due to the apparent constitutivity of the other PNC enzymes. The lowered NAPRTase activity would cause the loss of some NA through excretion and therefore reduce the level of intracellular NAD. These events should restore the NADP-NAD ratio back to 0.3. Support for this theory is provided by Wimpenny and Firth (118). They have measured NAD(H) levels in E. coli and Klebsiella aerogenes (K. pneumoniae) during transition from aerobic to anaerobic growth. Rapid drops in NAD levels were observed which could not be explained by simple reduction to NADH. They presume that this loss is due to NAD turnover. This hypothesis is feasible, since both cycles have been shown to function in vivo in E. coli (75) and S. typhimurium (Foster and Baskowsky-Foster, submitted for publication). However, Olivera has characterized a PolIEndoI- strain of E. coli that does not appear to recycle intracellular NAD via PNC VI. (Olivera et al., personal communication). Thus, a definitive role for PNC VI in regulating intracellular NAD levels remains to be established. The importance the PNC commands in the regulation of NAD biosynthesis in E. coli is reported by McLaren et al. (76). They conclude that the de novo biosynthetic pathway is the least preferred route for NAD biosynthesis, whereas the Preiss-Handler pathway takes precedence. At an NA concentration of 8 x 10-7 M, the average E. coli cell took up 90 molecules of NA per s for NAD biosynthesis and synthesized 340 NAD molecules per s de novo. Increasing the NA concentration to 2 x 106 M resulted in 215 molecules of NAD being derived from the medium per s and 215 molecules per s synthesized de novo. Finally, at 4 x 10-6 M NA, endogenous synthesis of NAD is completely suppressed. All of the intracellular NAD is then derived from the medium. Their conclusion that the PNC takes precedence over de novo synthe-

98

FOSTER AND MOAT

MICROBIOL. REV.

sis is supported by Chandler and Gholson (12, 13), who demonstrated a repression of QA synthetase activity by NA (presumably through conversion to NAD; see Regulation of De Novo Biosynthesis of Nicotinamide Adenine Dinucleotide). Foster et al. (27) arrived at a concentration similar to that mentioned above (5 x 106 M NA) for the maximal repression of NAPRTase in S. typhimurium. Some enzymes of the PNC have been shown to be associated with the cell surface. NMN glycohydrolase (5) was shown to be an integral component of the cytoplasmic membrane, and NAPRTase appears to be located within the periplasmic space of E. coli (8). The study with NAPRTase utilized an osmotic shock treatment termed the ethylenediaminetetraacetate-lysozyme-freeze-thaw method to release NAPRTase from the periplasmic space. This procedure released more enzyme as compared with sonic oscillation or alumina grinding. In addition, growth of E. coli supplemented with NA produced only a fourfold reduction in total enzyme activity, compared with the 100-fold decrease reported by others using alumina grinding techniques (51, 53, 93). Thus, the repression of NAPRTase in E. coli is not as strong as originally believed. NAD pyrophosphatase or an enzyme similar in function may also be associated with the cell membrane, at least in S. typhimurium (26). S. typhimurium fed ['4C]NAD rapidly degraded this NAD to [14C]NMN and adenosine 5'-monophosphate, both of which accumulate extracellularly. This degradation and extracellular accumulation occurred even in pncA mutants which did not readily incorporate '4C intracellularly from exogenously supplied ["4C]NAD. These findigs suggest that NAD degradation occurs at the membrane level. Thus, in summary, there is evidence suggesting a membrane association for two or possibly three PNC enzymes in E. coli and S. typhimurium.

Organisms of Special Interest Haemophilus. It is well known that Haemophilus influenzae and H. parainfluenzae require NAD for growth (25, 73). Neither NA nor NAm can substitute for NAD. H. haemoglobinophilus, however, does not require NAD as a growth factor (89). Kasarov and Moat (58) have demonstrated that this species synthesizes NAD from NAm by the following series of reactions: (i) NAm + 5-phosphoribosyl pyrophosphate + ATP

NAmPRTase

NMN

(ii) NMN + ATP

deNAD (NAD) pyrophosphorylase or NMN adenylyltranferase

NAD

This organism could not convert either QA or NA to NAD, indicating the lack of either a de novo biosynthetic pathway or a Preiss-Handler pathway. Mycobacteriunm A test commonly used to distinguish human strains of M. tuberculosis from bovine strains and the atypical mycobacteria is the NA test (63, 64). Human varieties of M. tuberculosis accumulate large quantities of NA in the culture medium, whereas other mycobacteria do not. A study by Kasarov and Moat (57) revealed that extracts prepared from a human strain of M. tuberculosis had very high levels of NAD glycohydrolase and NAm deamidase when compared with the bovine strain. Furthermore, whereas enzyme preparations exhibited activity with regard to the enzymes of the biosynthetic pathway from QA through NAD, NAPRTase levels were extremely low or absent. The human strain of M. tuberculosis rapidly degrades NAD to NA but cannot recycle the NA to NAD. The NA then accumulates extracellularly. The bovine strain, also with low NAPRTase activity, does not degrade NAD as rapidly, thereby accumulating much less NA. Procedures used to assay NAD glycohydrolase from M. tuberculosis include a heat activation step (108). Gopinathan et al. (39, 40) have shown that heating the extract destroys an inhibitor of mycobacterial NAD glycohydrolase. In fact, one of the more plausible theories of isoniazid inhibition involves the binding of the NAD glycohydrolase inhibitor by isoniazid, thus releasing the enzyme activity. The increased activity would consequently deplete the intracellular NAD pool (101). The presence of NAD glycohydrolase and the absence of a PNC appear to provide a logical explanation for the fact that isoniazid is effective against M. tuberculosis and not against a variety of other bacteria. Clostridium butylicum. C. butylicum (C. acetobutylicum), a strict anaerobe, synthesizes NAD from aspartate and formic acid as mentioned previously. C. butylicum also accumulates NA in culture media (54). In contrast to M. tuberculosis, however, this NA is not derived from the degradation of NAD. Rather, it is the result of NAMN and deNAD degradation (59). The pathway of degradation was shown to be as follows: deNAD -- NAMN -- NA-riboside -- NA Unexpectedly, 5 to 10 mM ATP stimulated degradation, whereas 20 mM ATP was required

VOL. 44, 1980

99


to reduce this degradation and allow conversion of NAMN and deNAD to NAD. Attempts to demonstrate recycling of NA were unsuccessful, indicating the lack of NAPRTase in this species (13,59). NAD pyrophosphatase and NAm deamidase activities were found (59), but no means of converting NMN to NAm was observed. Although it appears unusual for a species to irre-

versibly degrade a cofactor as important as NAD, this degradation may represent a primitive mechanism for metabolic control over the intracellular concentration of this cofactor (see Pyridine Nucleotide Cycle Metabolism, Regulation). Other studies have shown that C. sticklandii extracts have an NMN deamidase activity which converts NMN to NAMN and thus recycles NAD (31). C. butylicum was not examined for this enzyme. Other clostridia and anaerobic genera should be examined for this type of recycling capability. Azotobacter vinelandii Reports from Imai (48, 49) reveal that A. vinelandii, an aerobic, nitrogen-fixing species, possesses a PNC IV and a PNC VI similar to those of E. coli and S. typhimurium. The following activities have been measured: QAPRTase, NAPRTase, NAMN adenylyltransferase, NAD synthetase, NAD kinase, NAD glycohydrolase, NAD pyrophosphorylase, NMN deamidase (48), and NMN glycohydrolase (49). One can see from this list of enzymes that in addition to PNC IV and PNC VI, there is a PNC V which utilizes NAD glycohydrolase. This is the only species in which enzymes for all three types of PNC have been demonstrated. Lactobacillus and Leuconostoc. Members of Lactobacillus and Leuconostoc appear incapable of synthesizing NAD de novo. However, different species have developed different pathways to satisfy their pyridine nucleotide requirement (79). For example, Leuconostoc mesenteroides specifically requires NA, but Lactobacillus fructosus (Lactobacillus fructivorans) will only use NAm to synthesize NAD. Lactobacillus plantarum and Lactobacillus casei, how-

ever, can use either NA or NAm. A study by Ohtsu et al. (79) in 1967 explains this phenomenon on the basis of each species' complement of PNC enzymes. Table 3 summarizes their data. One can see that those species which specifically rquire NAm convert NAm to NAD via NMN in a manner similar to that shown for H. haemoglobinophilus (58). Lactobacilli which only use NA lack NMN pyrophosphorylase and NAm deamidase. Species which use either NA or NAm appear to have an intact PNC. The pathway from NAm to NAD via NMN may be expected to operate in all species which are strictly dependent upon NAm for growth. However, other species of this type, e.g., Pasteurella multocida (65) and Haemophilus species which do not require NAD for growth (121), have not been studied with regard to the mechanism of conversion of NAm to NAD. EVOLUTIONARY ASPECTS Evolution of Biosynthetic Pathways Since NAD is an essential cofactor in both aerobic and anaerobic processes, a pathway for its synthesis must have existed in primitive cellular organisms. Furthermore, since the earth's environment was initially anaerobic, it follows that the first biosynthetic pathway to NAD could not have involved molecular oxygen. Thus, a de novo biosynthetic pathway similar to the Escherichia-Salmonella aspartate pathway or the clostridial aspartate pathway would fulfill the requirements of a primordial system. Gaertner and Shetty (33) have discussed the possible evolution of the aerobic tryptophan pathway to NAD. They suggest five steps in the evolutionary process. Stage I is the development of the anaerobic biosynthetic pathway described above. Stage II is the appearance of the inducible kynureninases necessary for the catabolism of tryptophan as a carbon source. The subsequent formation of the aerobic NAD biosynthetic pathway from the inducible tryptophan catabolic enzymes constitutes stage III. Stage IV includes the further refinement of the aerobic

TABLE 3. PNC enzymes present in various microorganisms displaying different specificities for NA or NAm Enzyme present

Species

NAD

Requirement NAPRTase

NAm deamidase

NMN pyrophosphorylase

(deNAD) pyrophosphorylase

L. mesenteroides NA + + L. fructosus NAm + + L. plantarum NA or NAm + + + L. casei NA or NAm + + +/-a + a +/-, L. casei possessed barely measurable levels of NMN pyrophosphorylase activity. (Adapted from Ohtsu et al. [791.)

100

FOSTER AND MOAT

MICROBIOL. REV.

pathway, elimination of the inducible kynureni- Ramos, J. S. Imperial, and J. Cruz (manuscript nases, and loss of the anaerobic NAD biosyn- in preparation), and Foster and Baskowsky-Fosthetic pathway; interestingly, S. cerevisiae and ter (submitted for publication) support the conpossibly other fungi appear to be transitional cept of an intracellular preference for recycling species, possessing both anaerobic and aerobic NAD via PNC IV. PNC VI, although predomipathways to NAD (1, 98) but lacking an induc- nantly used for the transport of preformed pyrible kynureninase (98). Gaertner and Shetty sug- idine compounds, may also be used to fine-tune gest that the final stage, stage V, is the complete intracellular NAD levels through the repression elimination of all de novo NAD synthesizing of NAPRTase. One question presents itself. Why capacity, e.g., H. influenzae. does a genus, such as Azotobacter, possess three PNCs? Much more research will be recan be Evolution of the Pyridine Nucleotide Cycle quired before a definitive conclusion made. However, one may suggest that AzotoEvery species that has been studied regarding bacter represents an evolutionary branch point NAD turnover, be it procaryote or eucaryote, in PNC metabolism. has the ability to degrade NAD. Not all appear Nevertheless, all of the enzymes involved in to be fully capable of recycling the degradation the recycling of NAD as well as the regulation products, but degradation products are formed. associated with them represent a phenomenal One could conclude that the turnover of NAD is effort by the cell to preserve a compound, NA, very important to the cell. This importance is whose primary purpose involves the biosynmost probably manifested in the maintenance of thesis of NAD (Friedmann and his co-workers a certain NADP-NAD ratio, as suggested by [29a, 30] have shown that in some microorgaLundquist and Olivera (71), as well as the main- nisms NAMN is also involved in the synthesis tenance of a specific quantitative level of NAD, of cobalamin). Perhaps it is just coincidence that as suggested by Foster et al. (27). these enzymes and controls are present. Then Several enzymes involved with PNC metabo- again, perhaps it is not. There may be essential lism appear to be the result of gene duplication functions of NAD or its recycling yet to be and divergent evolution (87). For example, NA- discovered in both procaryotic and eucaryotic PRTase and QAPRTase are very similar with microbial metabolisms. regard to their substrate specificities, the only SUMMARY difference between NA and QA being an extra From the various studies that have been recarboxyl group on QA. NAm deamidase and NMN deamidase perform similar functions, as viewed with regard to the biosynthesis of NAD, do NAD pyrophosphatase and deNAD pyro- it can be generally stated that mammalian cells phosphorylase. In addition, NAD glycohydro- of most types, yeasts and molds metabolizing lase could have arisen from a modification of under aerobic conditions, and one or two unuNMN glycohydrolase or vice versa. Speculation sual genera of bacteria are all capable of conas to the order in which these evolutionary verting the ring carbon and nitrogen of tryptochanges took place would be unwarranted at this phan to the pyridine ring of NAD (Fig. 13). Most point. Suffice it to say the changes did occur and of the procaryotic species that have been examhave been maintained. Maintenance of a genetic ined to date utilize aspartate and DHAP to fonn duplication or alteration, however, implies a cer- the pyridine ring. The anaerobe C. butylicum tain amount of selective advantage associated appears to be the only species that utilizes a with it. Why, though, should such organisms as stepwise pathway involving the condensation of Propionibacterium (30), E. coli (75), and S. ty- formate with aspartate to form N-formylasparphimurium (62) possess two recycling pathways tate and further condensation with acetyl coen(PNCs IV and VI) and A. vinelandii possess zyme A to yield the pyridine ring of QA. QA is three (48,49)? The major difference between the the common intermediate in all of these paththree cycles is that PNC IV preserves the ribose- ways, and the steps from QA to NAD appear to 5-phosphate moiety of NMN, whereas PNC V be identical in all species capable of converting and PNC VI promote the turnover of this com- this compound to NAD. Among species that require either NA or pound with subsequent ATP-dependent replacement. The suggestion has been offered that PNC NAm, those that are not specific in their requireIV could be advantageous due to its minimal ment (i.e., that can use either NA or NAm to expenditure of energy (62). PNC V and PNC VI, satisfy their NAD requirement) convert NA to on the other hand, appear to be most suited for NAMN via NAPRTase. NAm is deamidated to the transportation and conversion of preformed NA. Species that exhibit a specific requirement pyridine compounds found in the environment. for NAm convert this compound to NMN and Results from Manlapaz-Fernandez and Olivera then to NAD. Although the NMN pathway of (75), B. M. Oliver, D. Hillyard, P. Manlapaz- NAD synthesis appears to be relatively rare in


VOL. 44, 1980 BACTERIA Mycobacterium tuberculosiJ (scherichia co/i Su/mone//l typhimurium Bocillu# sublilis FUNGI (ANAEROBICALLY) SO*E PLANTS

COOH H 'H I + SH H2C-O-® H2N COOH DHAP ASPARTATE H2 COH O-C

HCOOH +

COOH HCH I H2N

OOH

FORMATE ASPARTATE

cII

101

H2CHCOOH

TRYPTOPHAN MAMMALS 6

"COMMON PATHWAY

* YEASTS,MOLDS

STEPS# SOME PLANTS * Xatehomoea prua/ * COOH

(PREISS-HANDLER) NAMN -'bDEAMIDO-NAD----*NAD

NiCOOH QUINOLINATE

CH3 COSCoA fOOH HCH HCO CH \N 'OOH

-~~~~~~~

4 NA

NMN

NAm

NAm

~~

H N-FORMYLASPARTATE

ANAEROBIC BACTERIA (Clostridium buty/icum)

NA or NAmrequiring organisms

NAm-specific organisms Heemophi/us

haemog/obhinophilus L octoboci//us fruclosus

FIG. 13. Summary of NAD biosynthetic pathways. [XI represents unknown intermediate(s). Abbreviations:

,) phosphate; CoA, coenzyme A. microbial cells, it is present in a wide variety of mammalian cells and is intranuclear in location. Species capable of degrading NAD and recycling the various degradation products to reform NAD utilize one or more alternative pathways referred to collectively as PNCs. Most procaryotes examined are capable of degrading NAD to NMN and then to NAm and NA. Final recycling of NA to NAD occurs via the Preiss-Handler pathway. In E. coli, S. typhimurium, Azotobacter, and Propionibacterium, an alternate cycle (PNC IV) is present as a result of the activity of NMN deamidase. M. tuberculosis appears to be one of the few procaryotic species that degrades NAD via NAD glycohydrolase. However, this species lacks NAPRTase activity and, instead of recycling NA, excretes NA into the medium. Most eucaryotes, including yeasts and molds, degrade NAD via NAD glycohydrolase. The NAm produced is deamidated to NA and recycled to NAD via the Preiss-Handler pathway. Although the various pathways leading to the formation of NAD appear to be rather complex, this is not altogether surprising in view of the essential nature of NAD in cellular metabolism. What is surprising is the relatively late development of interest in detailed studies of the systems and regulation involved in the biosynthesis and metabolism of NAD. A much broader survey of the numerous microbial genera would

be of interest to determine the degree of conservation or divergence which may have occurred in the evolutionary development of pathways of synthesis and metabolism of NAD. Several aspects of NAD metabolism have yet to be resolved. The problem which has attracted the most attention, and yet has proven to be the most frustrating, concerns the isolation and identification of the intermediate(s) involved in the aspartate-DHAP pathway in procaryotes. Although several possible structures have been proposed, none has been well characterized chemically. Progress has been achieved in this area, however, in that several potential compounds have been isolated, and some defmiitive answers should be forthcoming. Another question which requires a satisfactory explanation is the necessity of multiple pathways for the recycling of NAD, particularly within a single species. Is NAD so precious a cofactor that evolutionary selective processes resulted in the emergence of several recycling capabilities? The most obvious rationalization would be that the ability to preserve the pyridine ring is essential to optimal metabolic capability, particularly if there are surges in the requirement for NAD as a substate. However, a role for these cycles in the regulation of intracellular levels of NAD cannot be overlooked. Further investigation with mutants blocked in each of the PNC pathways should provide a fruitful

102

FOSTER AND MOAT

avenue of approach to this problem. Further investigation will be necessary to characterize the genetic loci which code for the enzymes involved in NAD metabolism. Significant progress has been made in this regard with S. typhimurium and E. coli, but a great deal of additional work will be required to fully elucidate the genetic control of pyridine nucleotide metabolism. Additional studies with a complete set of mutants blocked in each of the steps in the metabolic pathways would provide a more complete and accurate picture of the genetics and regulation of NAD biosynthesis and metabolism. ACKNOWLEDGMENTS We are indebted to H. Friedmann for his critical reading of the manuscript and to R. K. Gholson and B. M. Olivera for their helpful discussion. LITERATURE CITED 1. Ahmad, F., and A. G. Moat. 1966. Nicotinic acid biosynthesis in prototrophs and tryptophan auxotrophs of Saccharomyces cerevisiae. J. Biol. Chem. 241:775-778. 2. Albertson, J. N., Jr., and A. G. Moat. 1965. Biosynthesis of nicotinic acid by Mycobacterium tuberculosis. J. Bacteriol. 89:540-541. 3. Andreoli, A. J., T. Grover, R. K. Gholson, and T. S. Matney. 1969. Evidence for a functional pyridine nucleotide cycle in Escherichia coli. Biochim. Biophys. Acta 192:539-541. 4. Andreoli, A. J., T. W. Ikeda, T. Nishizuka, and 0. Hayaishi. 1963. Quinolinic acid: a precursor to nicotinamide adenine dinucleotide in Escherichia coli. Biochem. Biophys. Res. Commun. 12:92-97. 5. Andreoli, A. J., T. W. Okita, R. Bloom, and T. A. Grover. 1972. The pyridine nucleotide cycle: presence of a nicotinamide mononucleotide specific glycohydrolase in Escherichia coli. Biochem. Biophys. Res. Commun. 49: 264-269. 6. Apps, D. K. 1970. The NAD kinases of Saccharomyces cerevisiae. Eur. J. Biochem. 13:223230. 7. Bachmann, B. J., K. B. Low, and A. L. Taylor. 1976. Recalibrated linkage map of Escherichia coli K-12. Bacteriol. Rev. 40:116-167. 8. Baecker, P. A., S. G. Yung, M. Rodriguez, E. Austin, and A. J. Andreoli. 1978. Periplasmic location of nicotinate phosphoribosyltransferase in Escherichia coli. J. Bacteriol. 133:1108-1112. 9. Brody, S. 1972. Regulation of pyridine nucleotide levels and ratios in Neurospora crassa. J. Biol. Chem. 247:6013-6017. 10. Brown, A. T., and C. Wagner. 1970. Regulation of enzymes involved in the conversion of tryptophan to nicotinamide adenine dinucleotide in a colorless strain of Xanthomonas pruni. J. Bacteriol. 101:456-463. 11. Casciano, D. A., and F. H. Gaertner. 1973. A specific and sensitive fluorometric assay for tryptophan oxygenase. Arch. Biochem. Bio-

MICROBIOL. REV. phys. 156:658-664. 12. Chandler, J. LR., and R. K. Gholson. 1972. De novo biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli: excretion of quinolinic acid by mutants lacking quinolinate phosphoribosyl transferase. J. Bacteriol. 111: 98-102. 13. Chandler, J. L. R., and R. K. Gholson. 1972. Studies on the biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli. III. Precursors of quinolinic acid in vitro. Biochim. Biophys. Acta 264:311-318. 14. Chandler, J. L. R., and R. K. Gholson. 1972. Studies on the de novo biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli. II. Quantitative method for isolating quinolinic acid from biological materials. Anal. Biochem. 48:529-535. 15. Chandler, J. L. R., R. K. Gholson, and T. A. Scott. 1970. Studies on the de novo biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli. I. Labelling patterns from precursors. Biochim. Biophys. Acta 222:523526. 16. Chaykin, S. 1967. Nicotinamide coenzymes. Annu. Rev. Biochem. 36:149-170. 17. Chen, J.-L., 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. 18. Chen, J., and G. J. Tritz. 1976. Detection of precursors of quinolinic acid in Escherichia coli. Microbios 16:207-218. 19. Cobb, J. R., S. C. Pearcy, and R. K. Gholson. 1977. Metabolism of 6-aminonicotinic acid in Escherichia coli. J. Bacteriol. 131:789-794. 20. Colrier, R. J. 1967. Effect of diphtheria toxin on protein synthesis: inactivation of one of the transfer factors. J. Mol. Biol. 25:83-98. 21. Collier, R. J. 1975. Diphtheria toxin: mode of action and structure. Bacteriol. Rev. 39:5445. 22. Dahmen, W., B. Webb, and J. Preiss. 1967. The deamido-diphosphopyridine nucleotide and diphosphopyridine nucleotide pyrophosphorylases of Escherichia coli and yeast. Arch. Biochem. Biophys. 120:440-450. 23. Dalgliesh, C. E. 1955. Metabolism of the aromatic amino acids. Adv. Protein Chem. 10:31150. 24. Dickinson, E. S., and T. K. Sundaram. 1970. Chromosomal location of a gene defining nicotinamide deamidase in Escherichia coli. J. Bacteriol. 101:1090-1091. 25. Evans, N. M., D. D. Smith, and A. J. Wicken. 1974. Haemin and nicotinamide adenine dinucleotide requirements of Haemophilus influenzae and Haemophilus parainfluenzae. J. Med. Microbiol. 7:359-365. 26. Foster, J. W., D. M. Kinney, and A. G. Moat. 1979. Pyridine nucleotide cycle of Salmonella typhimurium: isolation and characterization of pncA, pncB, and pncC mutants and utilization of exogenous nicotinamide adenine dinucleotide. J. Bacteriol. 137:1165-1175. 27. Foster, J. W., D. M. Kinney, and A. G. Moat.

VOL. 44, 1980


1979. Pyridine nucleotide cycle of Salmonella typhimurium: regulation of nicotinic acid phosphoribosyltransferase and nicotinamide deamidase. J. Bacteriol. 138:957-961. 28. Foster, J. W., and A. G. Moat. 1978. Mapping

and characterization of the nad genes in Salmonella typhimurium LT-2. J. Bacteriol. 133: 775-779. 29. Friedmann, H. C. 1971. Preparation of nicotinic acid mononucleotide from nicotinamide mononucleotide by enzymatic deamidation. Methods Enzymol. 18B:192-197. 29a.Friedmann, H. C., and L. M. Cagen. 1970. Microbial biosynthesis of B12-like compounds. Annu. Rev. Microbiol. 24:159-208. 30. Friedmann, H. C., and C. Gartski. 1973. The pyridine nucleotide cycle: presence of a nicotinamide mononucleotide-specific amidohydrolase in Propionibacterium shermanii. Biochem. Biophys. Res. Commun. 50:54-58. 31. Fyfe, S. A., and H. C. Friedmann. 1969. Vitamin B-12 biosynthesis enzyme studies on the formation of the a-glycosidic nucleotide precursor. J. Biol. Chem. 244:1659-1666. 32. Gaertner, F. H., K. W. Cole, and G. R. Welch. 1971. Evidence for distinct kynureninase and hydroxykynureninase activities in Neurospora crassa. J. Bacteriol. 108:902-909. 33. Gaertner, F. H., and A. S. Shetty. 1977. Kynureninase-type enzymes and the evolution of the aerobic tryptophan to nicotinamide adenine dinucleotide pathway. Biochim. Biophys. Acta 482:453-460. 34. Gholson, R. K. 1966. The pyridine nucleotide cycle. Nature (London) 212:933-935. 35. Gholson, R. K. 1968. Regulation of NAD biosynthesis via the pyridine nucleotide cycle. J. Vitaminol. 14(Suppl.):114-122. 36. Gholson, R. K., and J. Kori. 1964. Enzyme repression in the regulation of nicotinamide adenine dinucleotide biosynthesis in Bacillus subtilis. J. Biol. Chem. 239:PC2399-PC2400. 37. Gholson, R. K., G. J. Tritz, T. S. Matney, and A. J. Andreoli. 1969. Mode of nicotinamide adenine dinucleotide utilization by Escherichia coli. J. Bacteriol. 99:895-896. 38. 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. 39. Gopinathan, K. P., T. Ramakrishnan, and C. S. Vaidyanathan. 1966. Purification and properties of an inhibitor for nicotinamide adenine dinucleotide glycohydrolase from Mycobacterium tuberculosis H37Rv. Arch. Biochem. Biophys. 113:376-382. 40. Gopinathan, K. P., M. Sirsi, and C. S. Vaidyanathan. 1964. Nicotinamide adenine dinucleotide glycohydrolase of Mycobacterium tuberculosis H17Rv. Biochem. J. 91:277-282. 41. Griffith, G. R., J. L. R. Chandler, and R. K. Gholson. 1975. Studies on the de novo biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli: the separation of the nadB gene product from the nadA gene product.

103

Eur. J. Biochem. 54:239-245. 42. Gross, D., A. Fuge, R. Stecher, A. Zureck, and H. R. Schutte. 1965. Untersuchungen zur Biosynthesis der Nicotinsaure bei Mycobacterium tuberculosis. Z. Naturforsch. Teil B 20: 116-119. 43. Hadwiger, L A., S. E. Badiei, G. R. Wailer, 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. 44. Hayaishi, O., and K. Ueda. 1977. Poly (ADPribose) and ADP-ribosylation of proteins. Annu. Rev. Biochem. 46:95-116. 45. Heilmann, V. H., and F. Lingens. 1968. Reinigung und Eigenschaften der 3-Hydroxyanthranilat-oxygenase aus Saccharomyces cerevisiae. Hoppe-Seyler's Z. Physiol. Chem. 349: 223-230. 46. Hitchock, M. J. M., and E. Katz. 1978. Actinomycin biosynthesis by protoplasts derived from Streptomyces parvulus. Antimicrob. Agents Chemother. 13:104-114. 47. Honjo, T., and 0. Hayaishi. 1973. Enzymatic ADP-ribosylation of proteins and regulation of cellular activity. Curr. Top. Cell. Regul. 7:87127. 48. Imai, T. 1973. Purification and properties of nicotinamide mononucleotide amidohydrolase from Azotobacter vinelandii. J. Biochem. (Tokyo) 73:139-153. 49. Imai, T. 1979. Isolation and properties of a glycohydrolase specific for nicotinamide mononucleotide from Azotobacter vinelandii. J. Biochem. (Tokyo) 85:887-899. 50. Imsande, J. 1961. Pathway of diphosphopyridine nucleotide biosynthesis in Escherichia coli. J. Biol. Chem. 236:1494-1497. 51. Imsande, J. 1964. A comparative study of the regulation of pyridine nucleotide formation. Biochim. Biophys. Acta 82:445-453. 52. Imsande, J., and P. Handler. 1961. Biosynthesis of diphosphopyridine nucleotide. III. Nicotinic acid mononucleotide pyrophosphorylase. J. Biol. Chem. 236:525-530. 53. Imsande, J., and A. B. Pardee. 1962. Regulation of pyridine nucleotide biosynthesis in Escherichia coli. J. Biol. Chem. 237:13051308. 53a. International Union of Biochemistry. 1973. Enzyme nomenclature. American Elsevier Publishing Co., Inc., New York. 54. Isquith, A. J., and A. G. Moat. 1966. Biosynthesis of nicotinamide adenine dinucleotide and nicotinic acid by Clostridium butylicum. Biochem. Biophys. Res. Commun. 22:565-571. 55. Joshi, J. G., and P. Handler. 1962. Purification and properties of nicotinamidase from Torula cremoris. J. Biol. Chem. 237:929-935. 56. Kasarov, L. B., and A. G. Moat. 1972. Convenient method for the enzymatic synthesis of 14Cnicotinamide riboside. Anal. Biochem. 46:181186. 57. Kasarov, L B., and A. G. Moat. 1972. Metabolism of nicotinamide adenine dinucleotide in human and bovine strains of Mycobacterium

104

FOSTER AND MOAT

tuberculosis. J. Bacteriol. 110:600-603. 58. Kasarov, L. B., and A. G. Moat. 1973. Biosynthesis of nicotinamide adenine dinucleotide in Haemophilus haemoglobinophilus. Biochim. Biophys. Acta 320:372-378. 59. Kasarov, L. B., and A. G. Moat. 1973. Metabolism of the pyridine nucleotides involved in nicotinamide adenine dinucleotide biosynthesis in Clostridium butylicum. J. Bacteriol. 115:35-42. 60. Katz, E., and H. Weissbach. 1967. Studies on the cell-free synthesis of actinomycins. Dev. Ind. Microbiol. 8:67-87. 61. 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. 62. Kinney, D. M., J. W. Foster, and A. G. Moat. 1979. Pyridine nucleotide cycle of Salmonella typhimurium: in vitro demonstration of nicotinamide mononucleotide deamidase and characterization of pnuA mutants defective in nicotinamide mononucleotide transport. J. Bacteriol. 140:607-611. 63. Konno, K., R. Kurzmann, and K. T. Bird. 1957. The metabolism of nicotinic acid in mycobacteria. A method for differentiating tubercle bacilli of human origin from other mycobacteria. Am. Rev. Tuberc. 75:529-537. 64. Konno, K., R. Kurzmann, K. T. Bird, and A. Sbarra. 1958. Differentiation of human tubercle bacilli from atypical acid-fast bacilli. Am. Rev. Tuberc. 77:669-674. 65. Koser, S. A. 1968. Vitamin requirements of bacteria and yeasts. Charles C Thomas, Publisher, Springfield, Ill. 66. Krehl, W. A., L. J. Trepley, P. S. Sarma, and C. A. Elvejhem. 1945. Growth retarding effect of corn in niacin-low rations and its counteraction by tryptophan. Science 101:489-490. 67. Langley, D., and J. R. Guest. 1974. Biochemical and genetic characteristics of deletion and other mutant strains of Salmonella typhimurium LT-2 lacking a-keto acid dehydrogenase complex activities. J. Gen. Microbiol. 82:319335. 68. Lehman, L. R. 1974. DNA ligase: structure, mechanism, function. Science 186:790-797. 69. Lester, G. 1971. End-product regulation of the tryptophan-nicotinic acid pathway in Neurospora crassa. J. Bacteriol. 107:448-455. 70. Lingens, F., and P. Vollprecht. 1964. Zur Biosynthese der Nicotinsaure in Streptomyceten, Algen, Phycomyceten und Hefe. Z. Physiol. Chem. 339:64-74. 71. Lundquist, R., and B. M. Olivera. 1971. Pyridine nucleotide metabolism in Escherichia coli. I. Exponential growth. J. Biol. Chem. 246: 1107-1116. 72. Lundquist, R., and B. M. Olivera. 1973. Pyridine nucleotide metabolism in Escherichia coli. II. Niacin starvation. J. Biol. Chem. 248: 5137-5143. 73. Lwoff, A., and M. Lwoff. 1937. Studies on codehydrogenases. I. Nature of growth factor "V." Proc. R. Soc. London Ser. B 122:352-359.

MICROBIOL. REV. 74. Mahler, H. R., and E. H. Cordes. 1971. Biological chemistry, 2nd ed., p. 801-803. Harper & Row, Publishers, New York. 75. Manlapaz-Fernandez, P., and B. M. Olivera. 1973. Pyridine nucleotide metabolism in Escherichia coli. IV. Turnover. J. Biol. Chem. 248: 5067-5073. 76. McLaren, J., D. T. C. Ngo, and B. M. Olivera. 1973. Pyridine nucleotide metabolism in Escherichia coli. III. Biosynthesis from alternate precursors in vivo. J. Biol. Chem. 248:51445159. 77. Moat, A. G. 1979. Microbial physiology. John Wiley & Sons, Inc., New York. 78. 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. 79. Ohtsu, E., A. Ichiyama, Y. Nishizuka, and 0. Hayaishi. 1967. Pathways of nicotinamide adenine dinucleotide biosynthesis in nicotinic acid or nicotinamide requiring microorganisms. Biochem. Biophys. Res. Commun. 29:635-641. 80. Olivera, B., and I. R. Lehman. 1967. Diphosphopyridine nucleotide: a cofactor for the polynucleotide-joining enzyme from Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 57:17001704. 81. Ortega, M. V., and G. M. Brown. 1960. Precursors of nicotinic acid in Escherichia coli. J. Biol. Chem. 235:2939-2945. 82. Palleroni, N. J., and R. Y. Stanier. 1964. Regulation mechanism governing synthesis of the enzymes for tryptophan oxidation by Pseu-

domonas fluorescens. J. Gen. Microbiol. 35: 319-334. 83. Pardee, A. B., E. J. Benz, D. A. St. Peter, J. N. Krieger, M. Neuth, and H. W. Trieshman. 1971. Hyperproduction and purification of nicotinamide deamidase, a microconstitutive enzyme of Escherichia coli. J. Biol. Chem. 246:6792-6796. 84. Partridge, C. W. H., D. M. Bonner, and C. Yanofsky. 1952. A quantitative study on the relationship between tryptophan and niacin in Neurospora. J. Biol. Chem. 194:269-278. 85. Preiss, J., and P. Handler. 1958. Biosynthesis of diphosphopyridine nucleotides. I. Identification of the intermediates. J. Biol. Chem. 233: 483492. 86. Preiss, J., and P. Handler. 1958. Biosynthesis of diphosphopyridine nucleotides. II. Enzymatic aspects. J. Biol. Chem. 233:493-500. 87. Riley, M., and A. Anilionis. 1978. Evolution of the bacterial genome. Annu. Rev. Microbiol. 32:519-560. 88. Rosenfeld, H., and P. Feigelson. 1969. Synergistic and product induction of the enzymes of tryptophan metabolism in Pseudomonas acidovorans. J. Bacteriol. 97:697-704. 89. Ryan, W. J. 1968. An X-factor requiring Haemophilus species. J. Gen. Microbiol. 52:275286. 90. Sanderson, K. E., and P. E. Hartman. 1978. Linkage map of Salmonella typhimurium, edition V. Microbiol. Rev. 42:471-519.

VOL. 44, 1980


91. Sanwal, B. D. 1970. Allosteric controls of amphibolic pathways in bacteria. Bacteriol. Rev. 34:20-39. 92. Sarma, D. S. R., S. Rajalokshmi, and P. S. Sarma. 1964. Studies on the enzymes involved in nicotinamide adenine dinucleotide metabolism in Aspergillus niger. Biochim. Biophys. Acta 81:311-322. 93. Saxton, R. E., V. Rocha, R. J. Rosser, A. J. Andreoli, M. Shimoyoma, A. Sosska, J. L. R. Chandler, and R. K. Gholson. 1968. A comparative study of the regulation of nicotinamide adenine dinucleotide biosynthesis. Biochim. Biophys. Acta 156:77-84. 94. Schott, H., H. Staudinger, and V. Uflrich. 1971. The regulatory function of L-kynurenine 3-hydroxylase (EC 1.14.1.2) for the biosynthesis of pyridine nucleotides in anaerobically and aerobically grown Saccharomyces cerevisiae. Hoppe-Seyler's Z. Physiol. Chem. 352: 1654-1658. 95. Schott, H., V. Ullrich, and H. Staudinger. 1970. Enzymatic properties of L-kynurenine 3hydroxylase (EC 1.14.1.2) in Neurospora crassa. Hoppe-Seyler's Z. Physiol. Chem. 351: 99-101. 96. Scott, T. A., E. Bellion, and M. Mattey. 1969. The conversion of N-formyl aspartate into nicotinic acid by extracts of Clostridium butylicum. Eur. J. Biochem. 10:318-323. 97. Scott, T. A., and M. Mattey. 1968. The incorporation of formate into nicotinic acid by Clostridium butylicum. Biochem. J. 107:606-607. 98. Shetty, A. S., and F. H. Gaertner. 1973. Distinct kynureninase and hydroxykynureninase activities in microorganisms: occurrence and properties of a single physiologically discrete enzyme in yeast. J. Bacteriol. 113:1127-1133. 99. Shimoyama, M., Y. Tanigawa, T. Ito, R. Murashima, L. Ueda, and T. Tomoda. 1971. Nicotinamide deamidation by microorganisms in rat stomach. J. Bacteriol. 108:191-195. 100. Spencer, R. L., and J. Preiss. 1967. Biosynthesis of diphosphopyridine nucleotide. The purification and properties of diphosphopyridine nucleotide synthetase from Escherichia coli B. J. Biol. Chem. 242:385-392. 101. Sriprakash, K. S., and T. Ramarkrishnan. 1969. Isoniazid and nicotinamide adenine dinucleotide synthesis in Mycobacterium tuberculosis. Indian J. Biochem. 6:49-52. 102. Stouthamer, A. H. 1969. A genetical and biochemical study of chlorate-resistant mutants of Salmonella typhimurium. Antonie van Leeuwenhoek J. Microbiol. Serol. 35:505-521. 103. Stouthamer, A. H., and C. W. Bettenhausen. 1970. Mapping of a gene causing resistance to chlorate in Salmonella typhumurium. Antonie van Leeuwenhoek J. Microbiol. Serol. 36:555565. 104. Sugimura, T. 1973. Poly(adenosine diphosphate ribose). Prog. Nucleic Acid Res. Mol. Biol. 13: 127-151. 105. Sundaram, T. K. 1967. Biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli. Biochim. Biophys. Acta 136:586-588.

105

106. Sundaram, T. K., K. V. Rajagopalan, C. V. Pichappa, and P. S. Sarma. 1960. Studies on pyridine nucleotide biosynthesis. Biochem. J. 77:145-149. 107. Suzuki, N., J. Carlson, G. Griffith, and R. K. Gholson. 1973. Studies on the de novo biosynthesis of nicotinamide adenine dinucleotide in Escherichia coli. V. Properties of the quinolinic acid synthetase system. Biochim. Biophys. Acta 304:309-315. 108. Swartz, M. N., N. 0. Kaplan, and M. E. French. 1956. Significance of "heat activated" enzymes. Science 123:50-53. 109. Tritz, G. J. 1974. Characterization of the nadR locus in Escherichia coli. Can. J. Microbiol. 20:205-209. 110. Tritz, G. J., and J. LR. Chandler. 1973. Recognition of a gene involved in the regulation of nicotinamide adenine dinucleotide biosynthesis. J. Bacteriol.114:128-136. 111. 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. 104:45-49. 112. 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. 113. Viswamitra, M. A. 1975. Molecular structure of NAD. Nature (London) 258:540-542. 114. Wagner, C., and A. T. Brown. 1970. Regulation of tryptophan pyrrolase activity in Xanthomonaspruni. J. Bacteriol. 104:90-97. 115. Wicks, F. D., S. Sakakibara, and R. K. Gholson. 1978. Evidence for an intermediate in quinolinate biosynthesis in Escherichia coli. J. Bacteriol. 136:136-141. 116. Wicks, F. D., S. Sakakibara, R. K. Gholson, and T. A. Scott. 1977. The mode of condensation of aspartic acid and dihydroxyacetone phosphate in quinolinate synthesis in Escherichia coli. Biochim. Biophys. Acta 500:213216. 117. Wilson, R. G., and L. M. Henderson. 1963. Tryptophan-niacin relationship in Xanthomonaspruni. J. Bacteriol. 85:221-229. 118. Wimpenny, J. W. T., and A. Firth. 1972. Levels of nicotinamide adenine dinucleotide and reduced nicotinamide adenine dinucleotide in facultative bacteria and the effect of oxygen. J. Bacteriol. 111:24-32. 119. Yanofsky, C. 1954. The absence of a tryptophan-niacin relationship in Escherichia coli and Bacillus subtilis. J. Bacteriol. 68:577-584. 120. Yuan, J. H., and B. M. Anderson. 1973. Bull semen nicotinamide adenine dinucleotide nucleosidase. V. Kinetic studies. J. Biol. Chem. 248:417-421. 121. Zinneman, K., and E. L. Biberstein. 1974. Genus Haemophilus Winslow, Broadhurst, Buchanan, Krumwiede, Rogers and Smith 1917, 561, p. 364-368. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore.