Tài liệu Báo cáo khoa học: Identification of ATP-NADH kinase isozymes and their contribution to supply of NADP(H) in Saccharomyces cerevisiae docx

The phenotypic analysis of the single, double and triple mutants, which was unexpectedly found to be viable, for UTR1, YEF1 and POS5 demonstrated the critical contribution of Pos5p to mi

contribution to supply of NADP(H) in Saccharomyces

cerevisiae

Feng Shi1,2, Shigeyuki Kawai1, Shigetarou Mori1, Emi Kono1and Kousaku Murata1

1 Department of Basic and Applied Molecular Biotechnology, Division of Food and Biological Science, Graduate School of Agriculture, Kyoto University, Uji, Kyoto, Japan

2 School of Biotechnology, Southern Yangtze University, Wuxi, Jiangsu, China

The genomic DNA sequence of the widely studied

yeast Saccharomyces cerevisiae, which is a model

organism for eukaryotic cells, contains three NAD

kin-ase homologues, namely, Utr1p, Pos5p and Yel041wp

[1–3] NAD kinase (EC 2.7.1.23) catalyses NAD phos-phorylation by using phosphoryl donors (ATP or inor-ganic polyphosphate [poly(P)]), constituting the last step of the NADP biosynthetic pathway [4,5] For the

Keywords

ATP-NADH kinase; Pos5p; Saccharomyces

cerevisiae; Utr1p; Yef1p

Correspondence

K Murata, Department of Basic and Applied

Molecular Biotechnology, Division of Food

and Biological Science, Graduate School of

Agriculture, Kyoto University, Uji,

Kyoto 611-0011, Japan

Fax: +81 774 38 3767

Tel: +81 774 38 3766

E-mail: kmurata@kais.kyoto-u.ac.jp

(Received 23 August 2004, revised 25 April

2005, accepted 3 May 2005)

doi:10.1111/j.1742-4658.2005.04749.x

ATP-NAD kinase phosphorylates NAD to produce NADP by using ATP, whereas ATP-NADH kinase phosphorylates both NAD and NADH Three NAD kinase homologues, namely, ATP-NAD kinase (Utr1p), ATP-NADH kinase (Pos5p) and function-unknown Yel041wp (Yef1p), are found in the yeast Saccharomyces cerevisiae In this study, Yef1p was identified as an ATP-NADH kinase The ATP-NADH kinase activity of Utr1p was also confirmed Thus, the three NAD kinase homologues were biochemically identified as ATP-NADH kinases The phenotypic analysis of the single, double and triple mutants, which was unexpectedly found to be viable, for UTR1, YEF1 and POS5 demonstrated the critical contribution of Pos5p to mitochondrial function and survival at 37
C and the critical contribution

of Utr1p to growth in low iron medium The contributions of the other two enzymes were also demonstrated; however, these were observed only in the absence of the critical contributor, which was supported by comple-mentation for some pos5 phenotypes by the overexpression of UTR1 and YEF1 The viability of the triple mutant suggested that a ‘novel’ enzyme, whose primary structure is different from those of all known NAD and NADH kinases, probably catalyses the formation of cytosolic NADP in

S cerevisiae Finally, we found that LEU2 of Candida glabrata, encoding b-isopropylmalate dehydrogenase and being used to construct the triple mutant, complemented some pos5 phenotypes; however, overexpression of LEU2 of S cerevisiae did not The complementation was putatively attri-buted to an ability of Leu2p of C glabrata to use NADP as a coenzyme and to supply NADPH

Abbreviations

CgLEU2, LEU2 of yeast Candida glabrata; FOA, 5-fluoroorotic acid; GFP, green fluorescent protein; KNDE, 10 m M potassium phosphate,

pH 7.0, containing 0.1 m M NAD, 0.5 m M dithiothreitol and 1.0 m M EDTA; poly(P), inorganic polyphosphate; ScLEU2, LEU2 of yeast

Saccharomyces cerevisiae; SD, synthetic dextrose; SG, synthetic glycerol; SD+FOA+Ura, synthetic dextrose ⁄ 5-fluoroorotic acid ⁄ uracil;

WT, wild type; YPD, yeast extract ⁄ peptone ⁄ dextrose; YPG, yeast extract ⁄ peptone ⁄ glycerol.

phosphoryl donor, the enzyme using ATP and poly(P)

is termed poly(P)⁄ ATP-NAD kinase [4] and that using

ATP, but not poly(P), is termed ATP-NAD kinase [6]

For the phosphoryl acceptor, the enzyme specific to

NAD is designated NAD kinase and that

phosphory-lating both NAD and NADH is NADH kinase (EC

2.7.1.86) [2,3,7]

Utr1p, which was initially identified as an

ATP-NAD kinase, was proposed to participate in the

ferri-reductase system [1,8] It was required for the

reduc-tion of extracellular ferric chelates to their ferrous

counterparts and for the uptake of extracellular iron

This system consists of three components, namely,

Fre1p, NADH dehydrogenase and Utr1p Utr1p was

proposed to contribute to the system by supplying

NADP [1,8] However, the NADH kinase activity of

Utr1p has not yet been reported [1] Pos5p was

iden-tified as an ATP-NADH kinase; it was shown to be

localized in the mitochondrial matrix and to be

important to several NADPH-requisite mitochondrial

processes, e.g resistance to a broad range of

oxida-tive stress, respiration, arginine biosynthesis,

mito-chondrial iron homeostasis and mitomito-chondrial DNA

stability [2,3] The pos5 cell exhibits poor growth in

the presence of oxidative damage, glycerol as the sole

carbon source and in a medium lacking arginine [2]

This mutant accumulates high mitochondrial iron and

is defective in the mitochondrial Fe–S

cluster-contain-ing enzymes [2] The disruption of POS5 increases

frame-shift mutations in the mitochondrial DNA [3]

However, the function of Yel041wp remains

unidenti-fied

Although Pos5p is believed to play a significant role

in NADPH biosynthesis in mitochondria [2,3], the

con-tribution of Yel041wp and Utr1p to cellular function

and the precise function of Yel041wp is yet to be

clar-ified In the cytosol, NADPH is mainly supplied from

NADP by NADP-dependent glucose-6-phosphate

dehydrogenase (EC 1.1.1.49) (Zwf1p) [9–11] Cytosolic

NADPH is required for methionine biosynthesis [2,10]

The zwf1 cell is methionine auxotrophic due to the

depletion of NADPH [2,9–11], but is not arginine

auxotrophic, whereas the pos5 cell exhibits arginine

auxotrophy, but not methionine auxotrophy [2],

thereby suggesting that NADPH is synthesized in the

cytosol, separate from the mitochondria [2]

In this study, we identified the functions of

Yel041wp (designated as Yef1p) and Utr1p as

ATP-NADH kinases We also examined the phenotypes of

single and double mutants, as well as the triple

mutant, which was unexpectedly viable, for UTR1,

YEF1 and POS5 and attempted to clarify the roles of

these three enzymes

Results

Identification of Yel041wp (Yef1p) as ATP-NAD kinase

First, we attempted to identify the function of Yel041wp We referred to Yel041wp as ‘Yef1p’ based

on the designation of this protein in the SWISS-PROT database (http://www.genome.ad.jp/dbget-bin/www_ bfind?sptrembl) YEF1 consists of 1488 nucleotides encoding a polypeptide of 496 amino acid residues with a calculated molecular mass of 55.9 kDa and a calculated pI of 5.46 The YEF1 locus on genomic DNA does not contain introns

YEF1 was expressed in Escherichia coli as a fusion recombinant protein with a V5 epitope and His6 tag

at the C terminus The fusion protein, referred to as Yef1p, consisted of 533 amino acid residues and exhib-ited the calculated molecular mass of 60.1 kDa The cell extract of E coli MK746 expressing YEF1 showed 0.078 UÆmg)1 ATP-NAD kinase activity, while that of control strain SK45 carrying vector alone exhibited an activity of 0.00086 UÆmg)1 When metaphosphate and polyphosphate were used at 1.0 mgÆmL)1 as poly(P) instead of ATP, no NAD kinase activity was detected, thereby suggesting that Yef1p is indeed an ATP-NAD kinase Yef1p was purified to homogeneity from the cell extract of MK746 (Table 1) The purified enzyme migrated as a single protein band corresponding to

60 kDa on SDS⁄ PAGE (Fig 1A) and was eluted as a single peak, consisting of a protein of 480 kDa on gel

filtration chromatography (Fig 1B), thereby indicating that the enzyme was a homooctamer composed of 60-kDa subunits ATP and other nucleoside triphosphates (especially dATP) at 5 mm were utilized by Yef1p as phosphoryl donors as follows: nucleoside triphosphates, relative activity: ATP, 100%; dATP, 91%; CTP, 43%; UTP, 14%; GTP, 13%; TTP, 6% Poly(P)s (pyrophos-phate, tripolyphosphate and trimetaphosphate at 5 mm and polyphosphate, metaphosphate and hexametaphos-phate at 1 mgÆmL)1) and phosphorylated compounds (phosphocreatine, glucose-6-phosphate and

phospho-Table 1 Purification of Yef1p.

Total protein (mg)

Total activity (U) Yield (%)

Specific activity (UÆmg)1)

Purification (fold)

Butyl-Toyopearl 15.1 13.4 6.1 0.886 11.4 Ni–chelate AF

Toyopearl

enolpyruvate at 5 mm) were not utilized, thereby

indi-cating that Yef1p was an ATP-NAD kinase Km for

NAD and ATP were 1.9 mm and 0.17 mm, respectively

Properties of Yef1p and identification of Yef1p

and Utr1p as ATP-NADH kinases

The enzyme had an optimum pH of 8.5 in Tris⁄ HCl

(Fig 2A), the optimal temperature was 45
C (Fig 2B)

and half of its activity was lost on treatment at 54
C

for 10 min (Fig 2C) Bivalent metal ions such as Mg2+,

Mn2+, Co2+ and Ca2+ were indispensable for

ATP-NAD kinase activity In the presence of 1 mm metal

ions, the ATP-NAD kinase activity was as follows

Metal ions, relative activity: Mg2+, 100%; Mn2+, 77%;

Co2+, 32%, Ca2+, 26% On the other hand, no activity was detected in the presence of 1 mm Zn2+, Fe2+,

Cu2+, and monovalent metal ions (Na+ and Li+) NADPH and NADH slightly inhibited Yef1p; however, NADP and intermediates involved in NAD biosynthesis (nicotinic acid mononucleotide, nicotinic acid adenine dinucleotide, nicotinic acid and quinolinic acid) did not inhibit Yef1p (Table 2) HgCl2inhibited enzyme activity (Table 2), thereby indicating the importance of the SH group of the enzyme for its catalytic activity

We also found that Yef1p exhibited NADH kinase activity in the presence of ATP, but not poly(P) (1 mgÆmL)1 metaphosphate) On assaying the NADH kinase activity of purified Utr1p [1], a similar result was obtained Km for NADH of Yef1p was 2.0 mm

Fig 1 Molecular mass of Yef1p (A) SDS ⁄

PAGE of Yef1p Lane 1, protein markers

(Bio-Rad); lane 2, purified enzyme (1.5 lg).

(B) Gel filtration of Yef1p Purified Yef1p

was loaded on a Superdex 200 pg column

and was eluted as described in

Experimen-tal procedures The arrow (s) indicates the

elution volume (Ve) of the purified Yef1p.

Protein standards (d) were as follows:

(a) blue dextran 2000 (2000 kDa); (b)

tyroglobulin (669 kDa); (c) ferritin (440 kDa);

(d) catalase (232 kDa); (e) BSA (67 kDa);

(f) ovalbumin (43 kDa); and (g)

chymotry-psinogen A (25 kDa).

Fig 2 Effect of pH and temperature on Yef1p activity and stability (A) Effect of pH on ATP-NAD kinase activity NAD kinase activity was assayed by the stop method as described in Experimental procedures by using potassium phosphate (r), Tris ⁄ HCl ( ) and glycine ⁄ NaOH (m) Activity in the presence of Tris ⁄ HCl (pH 8.5) was taken relatively as 100% (B) Effect of temperature on ATP-NAD kinase activity NAD kinase activity was assayed by the stop method as described in Experimental procedures at indicated temperatures The activity at 45
C was taken relatively as 100% (C) Effect of temperature on the stability of Yef1p Purified Yef1p was incubated for 10 min at indicated tem-peratures in KNDE, cooled in an ice-water bath and the residual activity was determined by the stop method as described in Experimental procedures The activity after incubation at 30
C was taken relatively as 100%.

and that of Utr1p was 3.9 mm; this was similar to and

higher than the Km value of the NAD of Yef1p

(1.9 mm) and Utr1p (0.5 mm) [1], respectively Vmax

for the NADH of Yef1p was 1.9 mmÆmin)1ÆU)1 and

that of Utr1p was 3.5 mmÆmin)1ÆU)1; this was also similar to and higher than the Vmax value of the NAD

of Yef1p (1.7 mmÆmin)1ÆU)1) and Utr1p (1.2 mmÆ min)1ÆU)1), respectively Km and Vmax for NAD and NADH of Pos5p have not been reported [2,3]

Constructions of double and triple mutants for UTR1, YEF1 and POS5

To examine the roles of Yef1p, Utr1p and Pos5p, we attempted to construct double and triple mutants for UTR1, YEF1 and POS5 Tables 3, 4 and 5 list the yeast strains, plasmids and primers, respectively, used in this study We hypothesized that the triple mutant (utr1yef1-pos5) may be lethal due to the proposed significance of intracellular NADP and NADPH; therefore, we con-structed a triple mutant carrying UTR1 on YCplac33 (MK1208, utr1yef1pos5 YCp-UTR1) by replacing POS5 in MK933 (utr1yef1 YCp-UTR1) with CgLEU2 (LEU2 of Candida glabrata, GenBank ID CGU90626) and examined the viability of the triple mutant after the loss of YCp-UTR1 by using synthetic dextrose⁄ 5-fluoro-orotic acid⁄ uracil (SD+FOA+Ura) medium [12] The MK1208 (utr1yef1pos5 YCp-UTR1) was able to grow

in SD+FOA+Ura liquid medium as well as on the SD+FOA+Ura solid medium (data not shown) The resultant triple mutant (MK1219, utr1yef1pos5) that grew on the SD+FOA+Ura media was believed to lose YCp-UTR1 [12] The loss of YCp-UTR1 was con-firmed by the Ura– phenotype of MK1219

(require-Table 2 Effect of compounds on ATP-NAD kinase activity of

Yef1p The effect of compounds on the activity of Yef1p was

stud-ied by assaying ATP-NAD kinase activity in a reaction mixture

con-taining compounds at the indicated concentrations as described in

Experimental procedures The effect of NADP and HgCl2 was

examined by the stop method and others, by the continuous

method Activity in the absence of each compound was taken

relat-ively as 100%.

Compound

Concentration (m M )

Relative activity (%)

Nicotinic acid adenine dinucleotide 1.0 100

Table 3 S cerevisiae strains used in this study.

MK1208 MATa leu2D0 lys2D0 ura3D0 his3D1 utr1::kanMX4 yef1::HIS3 pos5::CgLEU2 YCp-UTR1 This study

ment of uracil for growth; data not shown) Thus, the

triple mutant was unexpectedly viable

Growth phenotypes of mutants for UTR1, YEF1

and POS5

We examined the growth phenotypes of single, double

and triple mutants, i.e utr1, yef1, pos5, utr1yef1,

utr1pos5, yef1pos5 and utr1yef1pos5 cells These mutants did not exhibit any severe growth defects at

30
C in SD, YPD, YPD high dextrose (20% glucose), YPD low dextrose (0.2% glucose) liquid media (Fig 3) and on SD solid medium (Fig 4A, control) However,

in yeast extract⁄ peptone ⁄ glycerol (YPG; 3% glycerol) medium, pos5 mutants (pos5, yef1pos5 and utr1pos5) showed a longer doubling time than the other single

Table 4 Plasmids used in this study YGRC, Yeast Genetic Resource Centre, Osaka University, Japan.

YEplac195 E coli ⁄ S cerevisiae shuttle vector, URA3, 2 lm, Ap r

[13]

R80Q, V163A) -His3MX6

a LEU2 of C glabrata b LEU2 of S cerevisiae.

Table 5 Primers used in this study The start and stop codons are specified in bold The Shine–Dalgarno sequence is indicated by double underlining The sequence corresponding to the genomic DNA sequence of S cerevisiae is underlined.

ATGAAAACTGATAGATTACTG

ATGCGTACGCTGCAGGTCGAC

TTAATCGATGAATTCGAGCTCG

ATGCGTACGCTGCAGGTCGAC

TTAATCGATGAATTCGAGCTCG

ATGCCAATTCTGTGTTTCCCGGAAATG

TTAGTAAAGTTCGTTTGCCGATACATG

and double mutants and the wild-type (WT, BY4742)

cell, although the triple mutant (utr1yef1pos5) did not

(Fig 3) The growth defect of pos5 mutants probably

reflected the mitochondrial dysfunction caused by the

deletion of POS5 [2,3] The absence of growth defects

in the triple mutant suggested that CgLEU2, which

was used for the disruption of POS5 in the utr1yef1

cell to construct triple mutant, can complement the

growth defect of pos5 mutants

All mutants exhibited proper growth on solid ium lacking methionine (data not shown), the med-ium on which we confirmed that the zwf1 cell exhibited growth defect as reported elsewhere [2,9– 11] (data not shown) Growth defects of the pos5 cell on medium lacking arginine, on medium contain-ing oxidative stress (2 mm hydrogen peroxide) and

on synthetic glycerol (SG) medium were previously reported [2] and confirmed in this study (Fig 4A), thereby indicating that Pos5p is a critical contributor

to mitochondrial functions [2,3] We found that utr1-pos5 and yef1pos5 cells appeared to grow somewhat less than pos5 cells on solid medium lacking argi-nine, on solid medium containing hydrogen peroxide and on solid SG medium (Fig 4A); this was con-firmed using liquid media lacking arginine (Table 6) However, utr1yef1 and other single mutants showed

no growth defects on these solid media (Fig 4A) These growth defects indicate that Utr1p or Yef1p can partially contribute to the mitochondrial function only in the absence of the critical contributor (Pos5p), i.e partial contribution was observed only

in the absence of the critical contributor However, the utr1yef1pos5 cell exhibited no growth defects on the solid and liquid media, unlike the other pos5 mutants (Fig 4A and Table 6), thereby suggesting that CgLEU2 can complement the growth defects of pos5 mutants In liquid medium lacking arginine, leu-cine slightly inhibited the growth of the triple mutant (Table 6)

Fig 3 Doubling times for the growth of single, double and triple

mutants for UTR1, YEF1 and POS5 The mutants and BY4742 (WT)

cells that were cultivated in YPD liquid medium to saturation were

washed three times in sterilized water and inoculated into 3 mL SD

(2% glucose), YPD (2% glucose), YPD high dextrose (20%

glu-cose), YPD low dextrose (0.2% glucose) and YPG (3% glycerol)

liquid media until D 600 of 0.05 The cells were cultivated aerobically

at 30
C and their growth was monitored by following D 600 every

4 h Averages in two independent experiments are provided.

A

B

Fig 4 Growth phenotypes of the mutants for UTR1, YEF1 and POS5 (A) The mutants and WT cells that were cultivated in SD liquid medium to saturation were washed three times in sterilized water and spotted

as described in Experimental procedures on

SD solid medium (control), SD solid media without arginine (–Arg), with 2 m M hydrogen peroxide (+H 2 O 2 ) and SG solid medium (SG) (B) pos5 mutants lacking ScLEU2 (pos5), carrying ScLEU2 on high copy vector (pos5 YEp13), low copy vector (pos5 pRS415) and carrying CgLEU2 on chromo-some (pos5::CgLEU2) were treated and spotted as in (A).

Complementing abilities of LEU2, UTR1 and YEF1

for the growth defects of the pos5 mutant

To confirm the complementing ability of LEU2 for the

growth defects of the pos5 mutant, we examined the

growth of pos5 cells containing CgLEU2 instead of

POS5on the chromosome, ScLEU2 (LEU2 of S

cere-visiae) on a high copy vector (YEp13) [13] and on a

low copy vector (pRS415) [13], i.e MK1224 (pos5::

CgLEU2), MK751 (pos5 YEp13) and MK1223 (pos5

pRS415), on several media on which pos5 mutants,

except for the triple mutant, showed growth defects

(Fig 4A) The pos5::CgLEU2 cell was able to grow on

these media, whereas pos5 YEp13, pos5 pRS415 and

pos5cells were unable to grow (Fig 4B), thereby

indi-cating that CgLEU2 on the chromosome, but not

ScLEU2 on YEp13 and pRS415, could complement

the growth defect of the pos5 cell The effect of leucine

on the growth of pos5::CgLEU2, pos5 YEp13 and

pos5 pRS415 cells were not detected on these solid media (data not shown)

The expression of POS5, UTR1 and YEF1 via a high-copy vector complemented the poor growth of the pos5 cell on SG solid medium and on and in solid and liquid media lacking arginine (Fig 5 and Table 6), thereby supporting the partial contribution of Utr1p and Yef1p to mitochondrial functions (Fig 4A)

Temperature sensitivity of the mutants for UTR1, YEF1 and POS5

At higher temperature (37
C) on SD solid medium, pos5 single mutant showed a slight growth defect, and the deletion of UTR1 or YEF1 and particularly of both UTR1 and YEF1 from the pos5 cell enhanced the growth defect (Fig 6) However, utr1yef1 and the other single mutants did not exhibit growth defects at

37
C (Fig 6), thereby indicating that Pos5p is a crit-ical contributor to the survival of the cells at 37
C on

SD solid medium; Utr1p or Yef1p and in particular, both Utr1p and Yef1p can contribute significantly to the survival only in the absence of main contributor (Pos5p) On YPD solid medium, the growth defect was alleviated (Fig 6)

Growth phenotypes of mutants for UTR1, YEF1 and POS5 in low iron medium

Because Utr1p is proposed to participate in the ferri-reductase system required for low iron uptake, the utr1 cell was expected to exhibit growth defect on the low iron medium [1,8] As expected, utr1 exhibited lower growth in the low iron medium than the yef1 and pos5 single mutants (Fig 7) The deletion of YEF1 or POS5 from utr1 further decreased the growth of utr1 to the same level as that of the ftr1 mutant, which lacks a high-affinity iron transporter and shows severe growth defects in the low iron medium [14] (Fig 7) Further-more, the deletion of both YEF1 and POS5 from utr1

Table 6 Doubling time of WT (BY4742) and pos5 mutants Means

of two independent experiments are provided Arginine

concentra-tions are specified in parentheses in mgÆL)1 In this study, 20

mgÆL)1arginine was usually added NG, No growth; ND, not

deter-mined.

Strains

Doubling time (h) Arg (0) Arg (0.2) Arg (2) Arg (20)

a This strain was grown in media lacking leucine The other strains

were grown in media containing leucine.

Fig 5 Complementation of pos5 cell

Indi-cated pos5 and WT cells carrying each gene

on a high-copy vector or high-copy vector

alone were treated as in Fig 4A and spotted

on SD solid media (glucose) with (control)

and without (–Arg) arginine and SG solid

media (glycerol) with (+Arg) and without

(–Arg) arginine.

decreased the growth to a level that was much lower

than that of the ftr1 mutant (Fig 7) It should be

noted that in the presence of Utr1p, the mutants (yef1,

pos5and yef1pos5 cells) did not exhibit growth defects

(Fig 7), thereby indicating that Utr1p is a critical

con-tributor to growth in the low iron medium and that

Yef1p or Pos5p and, in particular, both Yef1p and

Pos5p can contribute significantly to this kind of

growth only in the absence of the critical contributor

(Utr1p)

Discussion

The genomic sequence of the yeast S cerevisiae con-tains three NAD kinase homologues, i.e Utr1p, Pos5p and Yel041wp [1–3] In this study, we termed Yel041wp ‘Yef1p’ Among the three proteins, only the function of Yef1p was not identified biochemically; therefore, it was termed the ‘function-unknown’ pro-tein We identified that Yef1p functions as an ATP-NADH kinase by using recombinant protein expressed

in E coli We also confirmed that Utr1p, initially identified as an ATP-NAD kinase [1], was in fact an ATP-NADH kinase Thus, the three isozymes of NAD kinase, namely, Utr1p, Yef1p and Pos5p, were bio-chemically identified as ATP-NADH kinases [1–3] Yef1p exhibited a homooctameric structure consist-ing of 60-kDa subunits, while Utr1p exhibited a homo-hexameric structure consisting of 60-kDa subunits [1]; however, the structure of Pos5p has not been deter-mined [2,3] The homooctameric structure of Yef1p shows good agreement with that of the NADH kinase found in C utilis (a homooctamer consisting of 32-kDa subunits) [15] and pigeon liver NAD kinase (a homooctamer consisting of 34-kDa subunits) [16] However, it was not in agreement with the NAD kin-ase structure of humans (a homotetramer consisting

of 49-kDa subunits) [17] and that in Mycobacterium tuberculosis (a homodimer or homotetramer with 33–35-kDa subunits) [4,18–20] No regulators for Yef1p activity were found (Table 5) Intermediates of NAD biosynthesis, particularly quinolinic acid, did not affect Yef1p, although poly(P)⁄ ATP-NAD kinase of the Gram-positive bacterium Bacillus subtilis is acti-vated by this compound [21] NADPH, NADH, and NADP at 0.05 mm also exerted a slight effect on Yef1p activity, although these inhibited the ATP-NAD kinase activity of Utr1p; the residual activity of Utr1p was 0%, 59% and 61% in the presence of 0.05 mm

Fig 6 Temperature sensitivity of the mutants for UTR1, YEF1 and POS5 The mutants and WT cells were treated and spotted on SD and YPD solid media as des-cribed in Fig 4A and were grown at 30
C and 37
C as indicated.

Fig 7 Growth of mutants for UTR1, YEF1 and POS5 in low iron

medium In order to exhaust the intracellular iron content, the

mutants and WT cells were cultivated in a low iron liquid medium

to saturation and further cultivated for 24 h after a 100-fold dilution

of the saturated culture by the same fresh medium The cells were

washed three times in sterilized redistilled water and inoculated

into 3 mL SD (filled bar) and low iron (open bar) liquid media to give

D600of 0.05 (SD) and of 0.20 (low iron) The cells were cultivated

aerobically at 30
C, and growth was monitored by following the

D 600 Bars represent the relative D 600 (%) of the cultures in the

sta-tionary phase (SD, after 34 h; low iron, after 100 h), taking D600

(%) of the WT cell in each medium (SD, D600of 5.8; low iron, D600

of 2.4) as 100% Means of two independent experiments are

pro-vided.

NADPH, NADH and NADP [1], respectively, thereby

suggesting a difference in the regulation of Yef1p and

Utr1p by these compounds

The viability of the triple mutant for the three

NADH kinase genes (UTR1, YEF1 and POS5) at

30
C was unexpected NAD and NADH kinases have

been regarded as the sole enzymes producing NADP

and NADPH [5] Accordingly, NAD kinase was

recently reported to be essential to bacteria such as

B subtilis [22] and M tuberculosis [23] Taking into

account the fact that no NAD kinase homolog other

than Utr1p, Yef1p and Pos5p is found in the genome

sequence of S cerevisiae, we propose that a ‘novel’

enzyme, whose primary structure is different from

those of all known NAD and NADH kinases,

cata-lyses the formation of NADP or NADPH in S

cere-visiae Furthermore, we believe that the novel enzyme

was able to catalyse the formation of cytosolic NADP,

but not cytosolic NADPH and mitochondrial

NADP(H) for the following reasons: (a) methionine

auxotrophy of the zwf1 mutant [2,9–11] indicates that

cytosolic NADPH is not supplied by the novel enzyme

in this mutant; (b) viability and methionine

prototro-phy of the triple mutant (utr1yef1pos5) (data not

shown) supports the possibility that cytosolic NADP,

which is probably converted to NADPH by Zwf1p, is

supplied by the novel enzyme; and (c) the decreased

mitochondrial NADPH level in the pos5 mutant [2,3]

(Fig 4A) indicates that mitochondrial NADPH and⁄ or

NADP are not supplied by the novel enzyme in the

pos5mutant

The viability of the triple mutant (utr1yef1pos5)

might imply that the three NADH kinases are

dispen-sable (Utr1p, Yef1p and Pos5p) However, the

pheno-typic analysis of the single, double and triple mutants

for UTR1, YEF1 and POS5 and previous reports [2,3]

showed the critical contribution of Pos5p to

mitoch-ondrial functions and survival at 37
C, and the critical

contribution of Utr1p in supporting growth in a low

iron medium The contributions of the other two

enzymes were shown only in the absence of the critical

contributor, which was supported by the

complementa-tion of certain pos5 phenotypes through the

over-expression of UTR1 or YEF1 (Figs 4A,5,6,7; Table 6)

Furthermore, the alleviated temperature sensitivity of

the pos5 mutants on YPD solid medium when

com-pared with that on SD solid medium (Fig 6) may be

indicative of the significance of NADP and NADPH

in biosynthetic reactions, which is in agreement with

the well-accepted concept that NADP and NADPH

are involved primarily in biosynthetic reactions, while

NAD and NADH are involved primarily in catabolic

reactions [24]

Although the critical contribution of Yef1p alone to specific cellular function was not observed in this study, a difference in the regulation of Yef1p and Utr1p by NADPH, NADH and NADP (Table 2) [1], and the different transcriptional patterns and protein– protein interactions of Yef1p, Utr1p and Pos5p [25– 27] may be indicative of a certain critical contribution

of Yef1p In brief, for example, transcriptions of YEF1 are repressed under anaerobic conditions and in the presence of ethanol stress; however, those of UTR1 and POS5 are not affected [25,26] Two-hybrid analysis indicated that Yef1p interacted with Utr1p and the ‘function-unknown’ proteins (Yor315wp, Yhr115cp and Ykl009wp) On the other hand, Utr1p interacted with Yef1p and Nup119p (nuclear pore complex involved in nucleocytoplasmic transport), and Pos5p interacted with Gts1p (putative transcription factor) [27] The interaction of Yef1p with Utr1p is of biological interest and may be related to the pro-nounced requirement of both Yef1p and Utr1p in the absence of Pos5p

The elucidation of the localization of Utr1p and Yef1p would be helpful in understanding the critical contribution of Yef1p as well as the molecular mech-anism underlying the findings described in this study The localizations of Yef1p and Utr1p were predicted

by computer program analysis using ipsort [28], which detects the mitochondrial targeting sequence and N-terminal signal sequence for targeting proteins to the ER ipsort did not show any positive sequence in Yef1p and Utr1p, although it detected a mitochondrial targeting sequence in Pos5p [2,3], thereby implying that Yef1p and Utr1p are not, at least, mitochondrial enzymes We attempted to examine the localization of Yef1p and Utr1p by inserting the green fluorescent protein (GFP) gene into the 3¢ terminus of YEF1 and UTR1 on the chromosome by using the pFA6a-GFP(F64A, S65T, R80Q, V163A)-His3MX6 [29] gene modification plasmid in order to express them as GFP-fusion proteins, and then observing them using fluorescence microscopy or detecting them by western blotting with anti-GFP Ig (Molecular probes, Eugene,

OR, USA) However, their localization could not

be confirmed, possibly due to the low expression of the GFP-fusion proteins and⁄ or the sensitivity of the detection system

Finally, we also found that CgLEU2 (LEU2 of

C glabrata), but not ScLEU2 (LEU2 of S cerevisiae), complemented certain pos5 phenotypes (Fig 4) LEU2 encodes b-isopropylmalate dehydrogenase that cata-lyses the oxidation of b-isopropylmalate by using NAD, but not NADP [30] ScLeu2p reportedly uses NAD, but not NADP (< 5% efficiency) [30]; it was

also reported to be localized in the cytosol [31] No

positive sequence was detected in ScLeu2p during the

computer program analysis using ipsort, thereby

sup-porting the cytosolic localization of ScLeu2p The

co-enzyme specificity and localization of CgLeu2p have

not been reported However, ipsort did not show any

positive sequence in CgLeu2p, possibly suggesting that

CgLeu2p was localized in the cytosol Collectively, we

assume that cytosolic CgLeu2p has the ability to utilize

NADP and that it supplies cytosolic NADPH, whereas

cytosolic ScLeu2p cannot provide NADPH due to its

specificity to NAD In the triple mutant (utr1yef1pos5),

cytosolic NADP might be supplied by the ‘novel’

enzyme, being different from Utr1p, Yef1p, and Pos5p,

as discussed above In this context, we assume that the

adequate amount of cytosolic NADPH that is being

provided by CgLeu2p is possibly transported into the

mitochondria via an unidentified transporter localized

in the mitochondrial membrane This results in

comple-mentation of the pos5 phenotypes caused by low

mitochondrial NADPH levels [2,3], although an

NADPH supply of this kind is not adequate for

com-plementing the growth defects of the triple mutant at

37
C and in a low iron medium (Figs 6 and 7) This

assumption is also supported by the complementation

of pos5 phenotypes through the expression of UTR1

or YEF1 via a high-copy vector (Fig 5 and Table 6),

wherein it is implied that Utr1p and Yef1p are not

mitochondrial enzymes, as mentioned above The slight

growth inhibition of the triple mutant by leucine in

liquid medium lacking arginine (Table 6) might imply

that the expression and⁄ or activity of CgLeu2p are

sup-pressed by leucine

Experimental procedures

Materials

Yeast extract, tryptone, glucose-6-phosphate and NADH

were from Nacalai Tesque (Kyoto, Japan) Glutamate

dehydrogenase (EC 1.4.1.3), ATP, NAD, NADP and

NADPH were from Oriental Yeast (Tokyo, Japan)

Ferro-zine, pyrophosphate, tripolyphosphate, trimetaphosphate,

glucose-6-phosphate dehydrogenase and other nucleotides

were from Sigma (St Louis, MO, USA) Polyphosphate,

metaphosphate, hexametaphosphate, quinolinic acid and

5-fluoro-orotic acid (FOA) were from Wako Pure Chemical

Industries (Osaka, Japan) Yeast nitrogen base without

amino acids was from Difco (Sparks, MD, USA), and yeast

nitrogen base without ferric chloride and copper sulfate

was from Q-Bio Gene (Carlsbad, CA, USA) Purified Utr1p

was obtained as described elsewhere [1] Sources of other

materials are provided in the text

Strains

Strains of S cerevisiae were cultured at 30
C in nutrient-rich yeast extract⁄ peptone ⁄ dextrose (YPD) medium [1% (w⁄ v) yeast extract, 2% (w ⁄ v) peptone, 2% (w ⁄ v) glucose;

pH 5.0), if necessary, with 0.2 mgÆmL)1geneticin or in syn-thetic dextrose (SD) medium [0.67% (w⁄ v) yeast nitrogen base without amino acids, 2% (w⁄ v) glucose, and appropri-ate amino acids; pH 5.0] Glucose was replaced with 3% (v⁄ v) glycerol in the synthetic glycerol (SG) medium and the YPG medium The concentration of glucose was chan-ged to 0.2% and 20% (w⁄ v) for YPD low dextrose medium and YPD high dextrose medium, respectively The low iron medium was composed of 0.67% (w⁄ v) yeast nitrogen base without ferric chloride and copper sulfate, 40 lgÆmL)1 CuCl2, 2% (w⁄ v) glucose, 50 mm 2-morpholinoethanesulf-onic acid, pH 6.1, 1 mm ferrozine and appropriate amino acids [14] The SD+FOA+Ura medium was composed of 0.7% (w⁄ v) yeast nitrogen base without amino acid, 2% (w⁄ v) glucose, 0.1% FOA, 50 mgÆL)1 uracil and appropri-ate amino acids [12] The SD+FOA+Ura medium was similar; however, FOA and uracil were not included In order to prepare solid media, liquid media were solidified using 2% agar To check the growth on solid media, the cells were cultured to saturation at 30
C, collected, washed three times in sterilized water and diluted in water to yield

A600 of 2.0, 0.2 and 0.02 The diluted cell suspensions (5 lL) were spotted on appropriate solid media After

5 days, photographs were taken Culture conditions for derivative strains of E coli BL21(DE3) (Novagen, Madi-son, WI, USA) are given below In order to serve as a host for plasmid amplification, E coli DH5a was routinely cultured at 37
C in Luria–Bertani medium (1% tryptone, 0.5% yeast extract, 1% NaCl; pH 7.2) supplemented with

100 lgÆmL)1 ampicillin or 30 lgÆmL)1 kanamycin as required

Construction of YEF1 expression plasmid and strain

YEF1 was amplified from genomic DNA of S cerevisiae BY4742 with PfuUltra high-fidelity DNA polymerase (Stratagene, La Jolla, CA, USA) by using PCR and was cloned into pET-DEST42 (Invitrogen, Carlsbad, CA, USA) to produce pET-YEF1 in accordance with the manufacturer’s protocol The primers used were as follows: yef1-attB1FSD, yef1-attB2R, attB1 and attB2 A Shine– Dalgarno sequence (GAAGGAG) with optimal spacing (ATATAAAA) for appropriate translation initiation in

E coli was inserted upstream of the start codon of YEF1 (Table 5) The use of pET-DEST42 enabled us to fuse a V5 epitope and a His6 tag to the C-terminal of Yef1p E coli BL21(DE3) was transformed with YEF1 and pET-DEST42 to yield MK746 and SK45, respectively