OUP user menu

Why do some yeast species require niacin for growth? Different modes of NAD synthesis

You-Fang Li, Wei-Guo Bao
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00231.x 657-664 First published online: 1 August 2007


NAD holds a key position in metabolism and cellular regulatory events as a major redox carrier and a signalling molecule. NAD biosynthesis pathways have been reconstructed and compared in seven yeast species with completely sequenced genomes, including Saccharomyces cerevisiae, Kluyveromyces lactis, Candida glabrata, Debaryomyces hansenii, Candida albicans, Yarrowia lipolytica and Schizosaccharomyces pombe. Both amino acid and nucleotide sequence similarity analysis in silico indicated that de novo NAD biosynthesis might not exist in K. lactis, C. glabrata and Schiz. pombe, while other species have the kynurenine pathway. It also showed that the NAD salvage pathway via nicotinic acid and nicotinic acid mononucleotide is conserved in all of these yeasts. Deletion of KlNPT1 (the gene for nicotinate phosphoribosyl-transferase) is lethal, which demonstrates that this salvage pathway, utilizing exogenous nicotinic acid, is the unique route to synthesize NAD in K. lactis. The results suggested that the basis of the variation of niacin requirements in yeasts lies in their different combinations of NAD biosynthesis pathways. The de novo pathway is absent but the salvage pathway is conserved in niacin-negative yeasts, while both pathways coexist in niacin-positive yeasts.

  • NAD biosynthesis
  • nicotinic acid
  • comparative yeast genome analysis


NAD is well known as an important coenzyme in a variety of metabolic reactions. Recent studies have shown that this versatile compound plays roles in many different cellular processes, such as protein deacetylation (Lin et al., 2000; Hekimi & Guarente, 2003; Howitz et al., 2003), protein mono- and poly-ADP-ribosylation (Berger et al., 2004), calcium signalling (Guse, 2004) and modulation of transcription activity (Rutter et al., 2001; Zhang et al., 2002).

NAD is an essential cofactor in living systems. Two alternative de novo NAD biosynthesis pathways (starting from either l-aspartic acid or l-tryptophan) can be utilized in different organisms. Many prokaryotes, such as Escherichia coli (Begley et al., 2001), and some plants, such as Arabidopsis thaliana (Katoh & Hashimoto, 2004; Katoh et al., 2006), synthesize NAD via the aspartate pathway. In contrast, most eukaryotes use the tryptophan degradation (kynurenine) pathway (Katoh & Hashimoto, 2004). Recently, it has been demonstrated that the kynurenine pathway is also found in some bacteria (Kurnasov et al., 2003). The plant Oryza sativa probably has both pathways, according to a blast similarity search (Katoh & Hashimoto, 2004). The aspartate and kynurenine pathways converge at quinolinic acid, and then use the same reaction steps to finally form NAD. On the other hand, in different NAD-consuming reactions, NAD is converted into nicotinamide (Berger et al., 2004), which can be recycled back to NAD via nicotinic acid (Na) and nicotinic acid mononucleotide (NaMN), or via nicotinamide mononucleotide (NMN) by salvage pathways (Rongvaux et al., 2003). Such NAD recycling circuits link niacin (a generic name referring to both nicotinic acid and nicotinamide) to NAD biosynthesis. While nicotinamide can use two different salvage pathways, nicotinic acid is universally converted into NAD by the conserved ‘Preiss–Handler pathway’ (three steps from nicotinic acid to NAD in the salvage pathway via Na and NaMN) (Rongvaux et al., 2003). Thus, through these NAD biosynthesis routes (summarized in Fig. 1), both amino acids (aspartic acid and tryptophan) and niacin can act as precursors of NAD.


Schematic representation of various NAD synthesis pathways. NAD biosynthesis includes two de novo pathways – the kynurenine pathway and the aspartate pathway (starting from tryptophan or aspartic acid), and also two salvage pathways – via nicotinic acid (Na) and nicotinic acid mononucleotide (NaMN) and via nicotinamide mononucleotide (NMN). Reaction steps requiring O2 incorporation are indicated.

Most yeast species do not require niacin for growth. Such well-known species as Saccharomyces cerevisiae, Debaryomyces hansenii, Candida albicans and Yarrowia lipolytica are all niacin-positive. However, there are a minority of yeasts that require niacin (35 among 590 species described are niacin-negative) (Barnett et al., 1990). These include Kluyveromyces lactis, Candida glabrata and Schizosaccharomyces pombe (Fig. 2).


Nicotinic acid auxotrophy tests of different yeast species. Yeast strains representing S. cerevisiae (S288C), C. glabrata (CBS138), K. lactis (CBS2359), D. hansenii (CBS767), C. albicans (CBS562), Y. lipolytica (CLIB89) and Schiz. pombe (CBS356) were grown to stationary phase at 28°C in complete glucose medium. The cultures were centrifuged, and the pellets were washed with saline (0.15 M NaCl). Ten microliters of 100-fold diluted cell suspensions were dropped on to minimal glucose media with (upper panel) and without nicotinic acid (lower panel), and then grown at 28°C for 3 days (except for D. hansenii, which was grown for 4 days).

Niacin is involved in parts of the NAD synthesis pathways. Therefore it is possible that the different niacin requirements of various yeasts come from a divergence in their mode of NAD biosynthesis. Since most of the genes directly involved in the various pathways of NAD formation are already known, it is possible to reconstruct, from the genome sequence data, the NAD synthesis pathways operating in different yeast species. For the above-cited seven species, the complete genome sequences are available (Goffeau et al., 1996; Wood et al., 2002; Dujon et al., 2004; Jones et al., 2004). By comparing these pathways, we sought a possible basis for the variation in niacin requirement among these yeasts.

One other interest that prompted us to examine NAD biosynthesis pathways is the connection between aerobic/anaerobic growth and NAD biosynthesis pathways in S. cerevisiae, where NAD synthesis is conducted by both the kynurenine and salvage pathways (Panozzo et al., 2002). The kynurenine pathway is an O2-dependent process in which three reaction steps require the incorporation of molecular oxygen (Fig. 1). For this reason, the salvage pathway is required for this yeast to grow in an O2-depleted environment. While S. cerevisiae can grow anaerobically, many of these other yeasts cannot grow without oxygen. We therefore investigated the possible connection between the ability to grow anaerobically and the presence or absence of different NAD biosynthesis pathways in these yeasts.

Materials and methods

Strains and media

Yeast strains are listed in Table 1. Yeast cells were routinely grown at 28°C in a complete medium containing 1% Bacto-yeast extract (Difco), 1% Bacto-peptone (Difco) and 2% glucose. 200 μg mL−1 of antibiotic G418 was added, when required. To sporulate, a K. lactis diploid strain was grown on ME plate (5% Bacto-malt extract and 2% Bacto-agar, Difco). A synthetic minimal medium was prepared as described (Sherman, 1991), and nicotinic acid was dropped out as specified.

View this table:

Yeast strains

Yeast strainsGenotypeSource
S. cerevisiae
S288CPrototrophH. Fukuhara, Institut Curie
K. lactis
CK11 MATa/MATαuraA1/uraA1, adeT-600/+, +/lysA1X.J. Chen, UT Southwestern Medical Center
CK11/Δklnpt1CK11 klnpt1::KanMX/KlNPT1This work
CBS2359PrototrophH. Fukuhara, Institut Curie
C. glabrata
CBS138PrototrophH. Fukuhara, Institut Curie
D. hansenii
CBS767PrototrophH. Fukuhara, Institut Curie
C. albicans
CBS562PrototrophH. Fukuhara, Institut Curie
Y. lipolytica
CLIB89PrototrophH. Fukuhara, Institut Curie
Schiz. pombe
CBS356PrototrophH. Fukuhara, Institut Curie

Disruption of KlNPT1

To construct a K. lactis npt1 mutant, the ‘split-marker recombination’ procedure was used as described (Fairhead et al., 1996). The DNA fragments upstream (−586 to +4) and downstream (+1285 to +1965) the KlNPT1 ORF (the initial nucleotide of ORF was designated as +1) were amplified by PCR (two pairs of primers: 5′-CGGGATCCAAAGCAAATTAAGGCAG-3′, 5′-GGGGTACCATTGTGTCTTGGTTTGCT-3′ and 5′-CGGGATCCTTGAACCAAACCGGATTTTA-3′, 5′-GGGGTACCCAGCATACTGTGGATTC-3′), and cloned into pKA and pAN vectors, respectively. The resulting plasmids were cotransformed into K. lactis diploid strain CK11. The expected structure of chromosomal integration was confirmed by Southern hybridization: one of two KlNPT1 ORFs, from the second codon to the stop codon TAA, was found to be deleted and replaced by a KanMX selection marker to give a Δklnpt1::KanMX/KlNPT1 heterozygous diploid mutant.

Comparative genome analysis and reconstruction of NAD biosynthesis pathways in yeasts

The database websites of the yeast species used in this study are listed at the end of Table 2. blast search was performed using tools implemented therein.

View this table:

Comparison of NAD biosynthesis pathways in yeasts

S. cerevisiae 1C. glabrata 2K. lactis 2D. hansenii 2C. albicans 3Y. lipolytica 2Schiz. Pombe 4Other reference sequences5
de novo NAD biosynthesis
—Kynurenine pathway
BNA2/YJR078W: Tryptophan 2,3-dioxygenase--DEHA0G12947g (364/e-101)orf19.583 (919/9.9e-94)YALI0F26455g (304/6e-83)-
BNA3/YJL060W: ArylformamidaseCAGL0J05126g (645/0.0)KLLA0F01617g (633/0.0)DEHA0A04840g (471/e-133)orf19.5809 (1205/4.9e-124)YALI0E28787g (474/e-134)SPAC6B12.04c (1074/1.6e-110)
BNA4/YBL098W: Kynurenine 3-mono oxygenase--DEHA0C06743g (365/e-101)orf19.5443 (1049/1.7e-107)YALI0D09867g (358/5e-99)-
BNA5/YLR231C: Kynureninase--DEHA0G14949g (509/e-144)orf19.394 (1248/1.4e-128)YALI0B22902g (416/e-116)-
BNA1/YJR025C: 3-hydroxyanthranilic acid dioxygenase--DEHA0G24376g (236/4e-63)orf19.3515 (616/1.3e-61)YALI0B02852g (243/4e-65)-
BNA6/YFR047C: Quinolinate phosphoribosyl transferase--DEHA0F27995g (149/4e-37) frameshift DEHA0F28006g (251/1e-67)orf19.5054 (1008/3.7e-103)YALI0E07073g (419/e-118)-
Aspartate pathway
-*-*-*-*-*-*-*NP_417069 (Escherichia colil-aspartate oxidase)
-------NP_415271 (Escherichia coli quinolinate synthase)
Salvage NAD biosynthesis
—Salvage pathway via nicotinic acid and nicotinic acid mononucleotide (NaMN)
PNC1/YGL037C: nicotinamidaseCAGL0A01716g (279/7e-76)KLLA0F22242g (219/1e-57)DEHA0C14564g (148/3e-36)PNC1/orf19.6684 (371/1.2e-35)YALI0A21153g (153/6e-38)SPBC365.20c (223/2.3e-20)
NPT1/YOR209C: nicotinate phosphoribosyl-transferaseCAGL0L02805g (634/0.0)KLLA0D06655g (624/e-179)DEHA0G25201g (493/e-140)NPT1/orf19.7176 (1281/4.3e-132)YALI0B00220g (434/e-122)SPAC1486.06 (912/2.3e-93)
NMA2/YGR010W: nicotinic acid mononucleotide adenylyltransferaseCAGL0A01023g (546 e-155,KLLA0C18051g (498 e-141,DEHA0C12639g (484/e-137,orf19.7499 (1202/6.9e-128,YALI0E25652g (455/e-128,SPAC806.06c (1043/3.0e-107,
NMA1/YLR328W: nicotinic acid mononucleotide adenylyltransferase547/e-156)503/e-143)479/e-136)1221/2.3e-129)472/e-133)1045/1.8e-107)
QNS1/YHR074W: glutamine-dependent NAD synthetaseCAGL0J10758g (1255/0.0)KLLA0D13024g (1214/0.0)DEHA0A01969g (1004/0.0)orf19.1460 (2625/1.6e-274)YALI0A20108g (1042/0.0)SPCC553.02 (2276/6.6e-238)
Salvage pathway via nicotinamide mononucleotide (NMN)
-------P43490 (Homo sapiens nicotinamide phosphoribosyl transferase)
Nicotinic acid transport
TNA1/YGR260W: high affinity nicotinic acid plasma membrane permeaseCAGL0F08371g (773/0.0) CAGL0F00209g (186/7e-48)KLLA0A09449g (693/0.0) KLLA0C19019g (288/2e-78)DEHA0E10494g (257/4e-69)TNA1/orf19.4335 (642/2.2e-64)YALI0E20471g (271/3e-73) YALI0F28193g (259/6e-70) YALI0C02783g (255/1e-68) YALI0E20273g (244/3e-65) YALI0F19536g (241/3e-64) YALI0E27247g (240/4e-64) YALI0D12100g (237/3e-63) YALI0F05984g (234/2e-62) YALI0F28369g (229/7e-61) YALI0D10043g (229/9e-61) YALI0C08569g (223/5e-59) YALI0C22363g (219/1e-57) YALI0F21109g (218/2e-57) YALI0A09383g (169/8e-43) YALI0A17666g (166/7e-42)SPAC11D3.18c (786/5.1e-80) SPBC1683.12 (715/1.7e-72) SPAC1002.16c (684/3.3e-69) SPAC1039.04 (583/1.7e-58)


Genes coding for enzymes in de novo NAD biosynthesis pathways are absent from the genomes of K. lactis, C. glabrata and Schiz. Pombe. In the yeast S. cerevisiae, the kynurenine pathway is the recognized route for de novo NAD biosynthesis, and genes BNA1 to BNA6 (which encode all enzymes catalyzing the conversion from tryptophan to nicotinic acid mononucleotide) have been identified, with the exception of BNA3 (Schott et al., 1971; Iwamoto et al., 1995; Kucharczyk et al., 1998; Panozzo et al., 2002). Using the amino-acid sequences from S. cerevisiae, we searched for putative coding sequences for kynurenine pathway enzymes in the genome sequences of six other yeast species with the blast algorithm. The results are presented in Table 2. In D. hansenii, C. albicans and Y. lipolytica, orthologs to the BNA1 to BNA6 genes were found, suggesting that these three yeasts have a kynurenine pathway similar to that in S. cerevisiae. However, no genes orthologous to BNA1, BNA2, BNA4, BNA5 and BNA6 could be found in the genomes of C. glabrata, K. lactis or Schiz. pombe, indicating that the whole kynurenine pathway is absent in these yeasts. Is it possible that these species could use the aspartate pathway for NAD biosynthesis, even though it has been found only in prokaryotes and plants up to now? To examine this possibility, the genomes were searched for coding sequences corresponding to the amino-acid sequences of E. colil-aspartate oxidase and quinolinate synthase. But no ortholog could be detected in the genomes of any of these seven yeasts (Table 2). The search was repeated, using the amino-acid sequences of the same enzymes from the plant A. thaliana, with the same result (data not shown). The results revealed that C. glabrata, K. lactis and Schiz. pombe also lack the aspartate pathway found in bacteria and some plants. This is congruent with the absence of a BNA6 ortholog (in S. cerevisiae, this gene encodes quinolinate phosphoribosyl transferase which converts quinolinic acid into nicotinic acid mononucleotide). The aspartate and kynurenine pathways converge at quinolinic acid. In order to further clarify whether there is any trace of genes for de novo NAD synthesis in these three yeasts, we performed the blast search again with the nucleotide sequences. No significant hits were found except for BNA3 (data not shown). To conclude, bioinformatic analyses showed that some yeasts, such as C. glabrata, K. lactis and Schiz. pombe, probably lack any means of de novo NAD biosynthesis, as both the kynurenine pathway and the aspartate pathway are absent.

Two points concerning individual genes in the kynurenine pathway should be noted. While BNA1, BNA2, BNA4, BNA5 and BNA6 have orthologs in the genomes of D. hansenii, C. albicans and Y. lipolytica, all of them are absent from those of K. lactis, C. glabrata and Schiz. pombe. BNA3, however, is an exception, since it is present in every genome, including those which apparently lack the kynurenine pathway genes (Table 2). Although BNA3 has been presumed to encode arylformamidase catalyzing the second step (N-formyl-l-kynurenine to l-kynurenine) of the kynurenine pathway, its precise role is not well established, unlike that of the other five genes (Panozzo et al., 2002). Our comparative genomic data suggest that this gene could be involved in another function. Secondly, two partial overlapping sequences showing significant similarity to BNA6, at both protein and DNA levels, were found in D. hansenii (Table 2). DEHA0F27995g could predict a polypeptide similar to the N-terminal part of Bna6p and DEHA0F28006g to the C-terminal part. These two sequences have been annotated as pseudogenes in the genome database and their deduced amino acid sequences have been excluded from the inventory of D. hansenii proteins in the database. The presence of these two ORFs, and the fact that D. hansenii can grow in the absence of nicotinic acid (Fig. 2), strongly imply that these two sequences might act as a BNA6 homolog, once the probable frameshift error has been corrected.

NAD salvage pathway is conserved in all seven yeasts

In addition to the de novo biosynthesis route, S. cerevisiae can utilize the salvage pathway (via Na and NaMN) to recycle nicotinamide produced from various NAD-consuming reactions. Using this pathway, nicotinic acid is incorporated into NAD synthesis by the action of nicotinate phosphoribosyltransferase, which is highly conserved from prokaryotes to eukaryotes. Data shown in Table 2 indicate that this salvage pathway is also well-conserved among yeast species. In particular, the genes for nicotinate phosphoribosyltransferase show a high similarity with NPT1 in S. cerevisiae (Panozzo et al., 2002) at the amino-acid sequence level. Mammals have been reported to use another salvage pathway (via NMN) to recycle nicotinamide through a reaction catalyzed by nicotinamide phosphoribosyltransferase (NAmPRTase) (Rongvaux et al., 2002, 2003). However, our bioinformatics investigations showed that there is no gene encoding a putative NAmPRTase in the yeasts examined in this work (Table 2).

De novo pathways and the salvage pathway using nicotinic acid are convergent at NaMN, so they share two reaction steps from NaMN to NAD catalyzed by nicotinic acid mononucleotide adenylyltransferase and glutamine-dependent NAD synthetase, respectively. As shown in Table 2, genes for these two enzymes are highly conserved in all seven yeasts.

KlNPT1 is an essential gene in K. lactis

The results above showed that K. lactis, C. glabrata and Schiz. pombe might not undertake de novo NAD biosynthesis, while they do possess the NAD salvage pathway (via Na and NaMN). If that is the case, then the salvage pathway would be the sole route by which these yeasts generate NAD. To verify this point, we deleted the KlNPT1 gene, which is predicted to encode nicotinate phosphoribosyltransferase, the key enzyme converting nicotinic acid into NaMN in the salvage pathway. Deletion of KlNPT1 was carried out in a K. lactis diploid strain CK11. The resulting heterozygous diploid Δklnpt1/KlNPT1 was sporulated, and 20 asci were dissected on to complete glucose medium. The tetrad dissection showed the segregation pattern of a nonessential gene: 18 asci gave four viable spores and two asci gave three viable spores [(a) in Fig. 3]. This result was unexpected. However, we noticed that some colonies were much smaller, and that the spore segregation in 18 asci presented a typical Mendelian genetics situation: two normal colonies vs. two smaller colonies. All 40 smaller colonies were antibiotic G418 resistant while 38 normal colonies were sensitive [(b) in Fig. 3]. The G418-resistant cells grew poorly and growth seemed to stop after 2 days. This observation reminded us that the Δklnpt1 mutants could survive on some metabolites, such as NAD diffused by their KlNPT1 neighbours even though the disruption would be lethal. It is known that cells can live on a very low concentration of NAD. To check this hypothesis, the G418-resistant colonies were replicated a second time on to a G418-containing plate. No cells could grow any more [(c) in Fig. 3]. At the same time, a feeding experiment was carried out with all these G418-resistant colonies, and the results showed that their survival was dependent on the presence of KlNPT1 cells [(d) in Fig. 3]. Therefore, we conclude that disruption of KlNPT1 is lethal, showing that the salvage pathway is the sole means of NAD biosynthesis in K. lactis.


Tetrad dissection analysis of the K. lactis heterozygous diploid Δklnpt1/KlNPT1. Asci were dissected and grown on complete medium (a), and then replicated onto a plate containing 200 μg mL−1 of antibiotic G418 (b). The second replication was performed with G418 plate (c). At the same time, cells from each G418-resistant colony were suspended in saline (0.15 M NaCl) and spotted on complete medium for a feeding experiment in which the diploid Δklnpt1/KlNPT1 was used as a KlNPT1 holding strain (d). All plates were incubated at 28°C for 3 or 4 days and then photographed.


By analysis of the genomic sequence data, we found that three niacin-requiring species (K. lactis, C. glabrata and Schiz. pombe) lack both the kynurenine pathway and the aspartate pathway. Only the genes of the salvage pathway are present. In fact, all yeast species examined appear to possess the genes for this pathway (nicotinamidase, nicotinate phosphoribosyltransferase). For the niacin-negative species, this pathway is therefore absolutely necessary for NAD biosynthesis. This was experimentally confirmed by the fact that the K. lactis mutant deleted from NPT1 was incapable of growth even in the presence of nicotinic acid. To our knowledge, all other organisms previously investigated, from lower prokaryotes to mammals or green plants, have either a kynurenine pathway or an aspartate pathway, or both, for de novo NAD synthesis. Therefore, although niacin is routinely regarded as a vitamin (B3), it is not a vitamin in the strict sense of the word because the nicotinate moiety of NAD could be also synthesized from tryptophan or aspartic acid in those organisms. However, our work revealed that niacin is really a vitamin to some yeasts, such as K. lactis, C. glabrata and Schiz. pombe.

According to available taxonomic data (Barnett et al., 1990), there are only about 35 niacin-requiring yeast species. Such species are over-represented in particular genera. For example, all species closely related to K. lactis are niacin-negative (K. marxianus, K. aestuarii, K. africanus, K. wickerhamii, etc). Hanseniaspora species also often require niacin for growth. We may suppose that all these species have lost the amino-acid-derived NAD synthesis pathways, and have made the salvage pathway sufficiently efficient to provide an alternative. It is possible that this mode of NAD acquisition might have preceded the development of the kynurenine pathway of eukaryotes.

It has been shown that the NAD salvage pathway using nicotinic acid is necessary for anaerobic growth in the yeast S. cerevisiae (Panozzo et al., 2002). A relationship between anaerobiosis and the presence of the salvage pathway was therefore worth investigating. However, because the salvage pathway is conserved among different yeast species irrespective of their ability to grow in the absence of oxygen (Table 2), we investigated their ability to transport nicotinic acid. Although the major route of absorption of nicotinic acid was believed to be a process of passive diffusion (Henderson, 1983), TNA1, a gene for high affinity nicotinic acid transporter has been identified (Klebl et al., 2000; Llorente & Dujon, 2000). The search for TNA1 orthologs revealed that genes for such a putative transporter exist in the different yeast genomes and that this function is often quite redundant (Table 2). Thus, our data did not suggest any direct link between the differences in NAD biosynthesis pathways and the yeasts' ability to grow anaerobically.


We are very grateful to Dr Hiroshi Fukuhara (Institut Curie), Dr Monique Bolotin-Fukuhara (IGM, Orsay) and Prof. Stephen Oliver (Manchester) for critical reading of the manuscript. We thank Dr Hiroshi Fukuhara and Dr Xin Jie Chen (Southwestern Medical Center) for kindly providing us with strains.

This work received support from the ‘Institut Fédératif de Recherche Génomes-IFR115-’.


View Abstract