Characterization of the products of the genes
SNO1
and
SNZ1
involved
in pyridoxine synthesis in
Saccharomyces cerevisiae
Yi-Xin Dong, Shinji Sueda, Jun-Ichi Nikawa and Hiroki Kondo
Department of Biochemical Engineering and Science, Kyushu Institute of Technology, Iizuka, Japan
Genes SNO1 and SNZ1 are Saccharomyces cerevisiae
homologues of PDX2 and PDX1 which participate in pyri-
doxine synthesis in the fungus Cercospora nicotianae.In
order to clarify their function, the two genes SNO1 and
SNZ1 were expressed in Escherichia coli either individually
or simultaneously and with or without a His-tag. When
expressed simultaneously, the two protein products formed
a complex and showed glutaminase activity. When purified
to homogeneity, the complex exhibited a specific activity of
480 nmolÆmg
)1
Æmin
)1
as glutaminase, with a K
m
of 3.4 m
M
for glutamine. These values are comparable to those for
other glutamine amidotransferases. In addition, the gluta-
minase activity was impaired by 6-diazo-5-oxo-
L
-norleucine

[9,10]. Herein, we report that Sno1p and Snz1p serve as a
glutaminase to supply ammonia for the ring nitrogen of
pyridoxine in yeast. Based on these and other lines of
evidence, a putative synthetic pathway to pyridoxine is
presented in the Discussion in which ribulose 5-phosphate
and ammonia serve as the key starting or intermediary
material.
Experimental procedures
Materials
Inorganic salts and common organic chemicals including
amino acids, nucleic bases and vitamins were obtained from
commercial sources. Acetylpyridine adenine dinucleotide
(APAD) and 6-diazo-5-oxo-
L
-norleucine (DON) were from
Sigma (St. Louis, MO, USA). Glutamate dehydrogenase
from bovine liver was obtained from Oriental Yeast (Tokyo,
Japan). Reagents for genetic engineering such as restriction
enzymes were purchased from Takara (Kyoto, Japan) and
New England Biolabs (Beverly, MA, USA). Oligonucleo-
tides were custom synthesized by Hokkaido Science (Sap-
poro, Japan). Plasmid YEpM4 was a 2 lmDNA-based
shuttle vector with gene LEU2 as the selectable marker [11].
Plasmids pET21a, pET21d (both ampicillin resistant),
pET24a (kanamycin resistant) and His-bind columns were
from Novagen (Madison, WI, USA). The TOPO TA
cloning kit was the product of Invitrogen (Carlsbad, CA,
USA).
Strains and media
The S. cerevisiae strain used in this study was D373-1

2
O, 40 lg; MnSO
4
ÆH
2
O, 400 lg; FeCl
3
Æ
6H
2
O, 200 lg; ZnSO
4
Æ7H
2
O, 400 lg; Na
2
MoO
4
Æ2H
2
O,
Correspondence to H. Kondo, Department of Biochemical
Engineering and Science, Kyushu Institute of Technology,
Kawazu 680-4, Iizuka 820-8502, Japan.
Fax: + 81 948 7801, Tel.: + 81 948 29 7814,
E-mail: [email protected]
Abbreviations: APAD, acetylpyridine adenine dinucleotide; APADH,
reduced form of APAD; DON, 6-diazo-5-oxo-
L
-norleucine.

from Shimadzu (Kyoto, Japan). Protein sequences were
determined on an Applied Biosystems 491 A sequencer.
Mutant construction
The pyridoxine auxotrophic mutants of yeast were pro-
duced by ethyl methanesulfonate mutagenesis [15]. In brief,
cells were treated with 3% ethyl methanesulfonate for
50 min and then spread over the entire surface of synthetic
medium plates supplemented with pyridoxine and the
growth requirements, and the plates were incubated at
30 °C. The colonies that appeared on the supplemented
plates were then transferred by replica plating first to a
minimal plate without pyridoxine (–PN) and then to one
supplemented with pyridoxine (+PN). The plates were
scored for colonies that appeared on +PN media but not
on the –PN media. This selection process was repeated
several times to ensure that the colonies that were unable to
grow in –PN medium are indeed pyridoxine auxotrophs.
Transformation of yeast cells
Pyridoxine auxotrophic mutants were transformed with a
YEpM4-based genomic library [11] for complementation
by the lithium acetate protocol [16]. Transformants were
selected on synthetic media lacking pyridoxine and leucine.
The plasmids were isolated from S. cerevisiae as described
[14].
Construction of over-expression plasmids for the
SNO1
and
SNZ1
genes
The coding regions of the SNO1 (672 bp) and SNZ1

isopropyl thio-b-
D
-galactoside was added to 0.4 m
M
,except
for the expression of Sno1p (0.1 m
M
). Eight hours after
induction the cells were collected by centrifugation and
washed with phosphate buffered saline. Subsequent steps of
protein purification were carried out at 4 °C, unless
otherwise stated.
The complex of Sno1p with a His-tag and Snz1p without a
tag. The washed cells were resuspended in 100 mL of
5m
M
imidazole, 0.5
M
NaCl, 20 m
M
Tris/HCl, pH 8.0
Table 1. Sequences of oligonucleotides used as PCR primers. Symbols
O, Z and H stand for Sno1p, Snz1p and His-tag, respectively.
Underlined are the restriction enzyme sites.
Primer Sequence
P1O 5¢-ATA
CCATGGACAAAACCCACAGTACAATG
P1OH 5¢-
CATATGCACAAAACCCACAGTAC
P2O 5¢-TAT

The suspension was sonicated and then centrifuged at
17 000 g for 30 min. The supernatant containing 120 mg of
protein was filtered through a 0.45 lm filter and applied to
a His-Bind affinity column (1 · 10 cm), pretreated with
50 m
M
NiSO
4
and equilibrated with buffer A. The column
was washed successively with buffer A and buffer B
(50 m
M
imidazole, 0.5
M
NaCl, 20 m
M
Tris/HCl, pH 8.0)
and protein was finally eluted with a gradient of
100–500 m
M
imidazole in buffer B. The protein-containing
fractions were pooled and dialyzed against 35 m
M
potas-
sium phosphate, 1 m
M
EDTA and 0.1 m
M
dithiothreitol,
pH 7.5. The purity of the desired protein was greater than

and 8
M
urea, pH 9.0, shaken at 37 °Cfor1hand
centrifuged for 30 min. The supernatant was mixed with
30mLofthesamebufferwith6
M
urea and dialyzed
against the same buffer with 4
M
urea. Dialysis was
continued against the buffers containing 2, 1 and 0
M
urea,
successively. The solution was concentrated to one half the
volume by dialysis against 50 m
M
Tris/HCl, 1 m
M
dithio-
threitol and 0.1 m
M
EDTA, pH 9.0, containing 15%
polyethylene glycol 20 000 and subjected to gel filtration
chromatography on Superdex 200 (2.6 · 60 cm, Pharma-
cia) to give 30 mg of virtually pure protein.
Snz1p without a His-tag. Harvested cells were processed
in a way identical to that for Sno1p up to the cell
disruption step. The supernatant was subjected to DEAE-
cellulose chromatography (2 · 10 cm, Whatman, Maid-
stone, UK) and protein was eluted with a linear gradient of

M
EDTA, 0.5 m
M
APAD and
7 units of glutamate dehydrogenase at 37 °C for 90 min.
After centrifugation (14 000 g) for 1 min, the absorbance of
the supernatant was read at 363 nm. The absorbance of
the reduced form of APAD (APADH) was linear over the
2–100 l
M
range of glutamate with a molar extinction
coefficient of 8900
M
)1
Æcm
)1
[17].
Detection of glutamate
Glutamate formed from glutamine by the complex of Sno1p
with a His-tag and Snz1p, was further detected by TLC
following dansylation. The reaction mixture (0.3 mL)
contained 10 m
M
glutamine, 30 lgofthecomplexin
50 m
M
Tris/HCl, pH 8.0, and it was incubated at 30 °Cfor
10 min. Ten microliter aliquots of this solution together
with 10 lLeachof10m
M

mutants were established from D373-1. Among these was
mutant K64 (Fig. 1). To identify the gene(s) affected by
complementation, the mutant was transformed with a
library of yeast chromosomal DNA. It was found that 4.6
and 5.1 kb overlapping fragments of chromosome XIII
carrying SNO1 and SNZ1 were capable of complementing
the defect of mutant K64 (Figs 1 and 2). They are
homologues of PDX2 and PDX1 of fungus C. nicotianae,
respectively [7,20], strongly suggesting that they are the
genes responsible for the pyridoxine auxotrophy observed
for that mutant.
Ó FEBS 2004 Pyridoxine biosynthesis in yeast (Eur. J. Biochem. 271) 747
Site of mutation in K64
In order to identify the site of mutation, the two genes were
amplified by PCR with the chromosomal DNA of mutant
K64 as template. Sequencing of the mutated genes revealed
that there was indeed a mutation on both of the genes. In
SNO1, the 199th G from the 5¢ terminus of the open reading
frame was deleted to result in a frame-shift and appearance
of stop codons in the downstream. As a result, a protein as
small as 70 residues is generated, a size too small for any
protein to be functional as an enzyme (see below). In SNZ1,
the 709th G was converted to A to result in replacement of
Gly237 with Arg. It is noted that this residue and the
surrounding regions are well conserved among the homo-
logues of SNZ1 including pyroA and PDX1 [6,7], suggesting
that the residue plays an important role. It should be
emphasized that the dual mutation of the two genes was
necessary to make yeast cells pyridoxine auxotrophic; cells
were still viable in the absence of pyridoxine even when

Fig. 1. Growth of S. cerevisiae strains D373-1, K64 and K64t. Yeast
cells were treated with ethyl methanesulfonate as detailed under
Experimental procedures. After several rounds of screening on media
containing pyridoxine (+PN) and not containing pyridoxine (–PN),
auxotrophic mutants were established, one of which was K64. K64t
represents the transformant harboring plasmid pDYX11 (Fig. 2) and
is capable of growing in –PN medium.
Fig. 2. Partial map of S. cerevisiae chromosome XIII carrying SNO1
and SNZ1. Dark grey bars represent 4.6 and 5.1 kb DNA fragments
capable of complementing the pyridoxine auxotrophy of mutant K64.
Fig. 3. SDS/PAGE of purified Sno1p, Snz1p and their complex with or
without a C-terminal His-tag. Lane 1, protein markers; lane 2, Sno1p
without tag; lane 3, Sno1p with tag; lane 4, complex of Sno1p with tag
and Snz1p without tag; lane 5, Snz1p without tag; lane 6, Snz1p with
tag. About 10–20 lg of protein was loaded and stained with Coo-
massie Brilliant blue.
748 Y X. Dong et al.(Eur. J. Biochem. 271) Ó FEBS 2004
obtained from a 1 L culture for Sno1p, Snz1p and the
complex, respectively. Identity of each protein was con-
firmed by N-terminal sequencing; the sequences were
correct at least to the 9th cycle from the N-terminus
including the initiating Met: MHKTHSTMS for Sno1p and
MTGEDFKIKS for Snz1p.
Properties of the complex of Sno1p and Snz1p
When Sno1p with a His-tag and Snz1p without a tag were
coexpressed and then applied to a His-Bind affinity column,
not only Sno1p but also Snz1p bound to the column and
was eluted simultaneously by imidazole, strongly suggesting
that they form a complex. The ratio of the two proteins,
assessed by SDS/PAGE, seemed to be equimolar, although

to be reasonable for a glutaminase, though the specific
activity varies drastically from enzyme to enzyme and
depends on whether a proper synthetase partner and
cosubstrate are present or not. For example, the V
max
and
K
m
values of imidazole glycerol phosphate synthase in
the absence of substrate are 0.084 lmolÆmin
)1
Æmg
)1
and
4.8 m
M
, respectively [22]. The V
max
was enhanced 39-fold
and K
m
lowered 20-fold in the presence of substrate N
1
-[(5¢-
phosphoribulosyl)formimino]-5-aminoimidazole 4-carbox-
amide ribonucleotide. It may be reasonable therefore to
assume that once the unknown substrate or ligand for
Snz1p was added, the glutaminase activity could have been
even higher. It should be noted that the enzyme activity is
gradually lost over time with a half life of 2–3 days at 4 °C.

dansylated (DANS) products were visualized under ultraviolet light.
Fig. 5. Michaelis–Menten kinetics for glutamine hydrolysis mediated by
the complex of Sno1p with a His-tag and Snz1p. The reaction was
carried out in 1 mL of 50 m
M
Tris/HCl, pH 8.0, in the presence of
1–10 m
M
glutamine and 50 lg of the complex, at 30 °C for 10 min.
The sample was then boiled for 1 min and a 0.3 mL aliquot was
incubated in 1 mL of 50 m
M
Tris/HCl, pH 8.0, containing 1 m
M
EDTA, 0.5 m
M
APAD and 7 units of glutamate dehydrogenase at
37 °C for 90 min. After centrifugation for 1 min, the absorbance of the
supernatant was read at 363 nm for APADH. The curve drawn is a
theoretical one based on 0.48 lmolÆmin
)1
Æmg
)1
for V
max
and 3.4 m
M
for K
m
for glutamine.

similar observation was made for phosphoribosylpyrophos-
phate amidotransferase [24]. Glutamine was effective in
protecting the enzyme from inactivation and its effect was
again dose-dependent, suggesting that inhibition by DON
occurs at the active site or the glutamine-binding site of
the enzyme (Sno1p). Although DON inhibition of Sno1p
was not pursued further, it is worth pointing out that the
cysteine serving as the key catalytic residue and modified
covalently by DON in other amidotransferases is also
conserved in Sno1p at position 100.
Discussion
As described above, the gene products of SNO1 and SNZ1
serve as a glutamine amidotransferase, which is needed
to supply ammonia as a source of the ring nitrogen of
pyridoxine [4]. Although Sno1p alone does not exhibit
detectable glutaminase activity, it seems certain that it is
responsible for the hydrolysis of glutamine mediated by the
complex with Snz1p. For example, the amino acid sequence
of Sno1p has 40% identity to that of the glutaminase
subunit of yeast imidazole glycerol phosphate synthase [22].
In addition, the key catalytic residues required for glutamine
hydrolysis by glutamine amidotransferases including imi-
dazole glycerol phosphate synthase, i.e. Cys100, His203 and
Glu205 (numbering based on Sno1p), are conserved in
Sno1p as well. Presumably, Sno1p hydrolyzes glutamine by
the same mechanism as those of other glutamine amido-
transferases such as imidazole glycerol phosphate synthase
and carbamoyl-phosphate synthetase, whose three-dimen-
sional structures are available [25,26].
In light of the function of these glutamine amidotrans-

acceptor substrate of Snz1p is identified, its addition to the
reaction system will enhance the glutaminase activity of
Sno1p significantly. In light of the fact that a ketopentose
seems to be a component of the skeleton of pyridoxine in
yeast and related organisms (see below), dihydroxyacetone
phosphate, glyceraldehyde 3-phosphate or related com-
pounds are probable candidates for the coupling partner.
These possibilities are presently under scrutiny in this
laboratory.
Recently, it was shown that a ketopentose is one of the
starting materials for pyridoxine in yeast and the initial form
of vitamin B6 produced is 2¢-hydroxypyridoxine [9,10]. Our
unpublished observation supports these findings; the gene
RKI1 coding for ribose 5-phosphate ketol-isomerase
(Rki1p), which interconverts ribose 5-phosphate and ribu-
lose 5-phosphate, dictates somehow pyridoxine synthesis in
yeast. In this light, ribulose 5-phosphate may be the more
probable candidate for the starting material, as its structure
fits the skeleton from positions 2¢ to 4¢ of 2¢-hydroxypyri-
doxine neatly. Hence, the possibility that ribulose 5-phos-
phate serves as the direct ammonia-acceptor was addressed.
It was found that this compound considerably inhibits the
glutaminase activity of the complex of Sno1p and Snz1p in a
competitive fashion; the activity decreased to 70% at 8 m
M
.
Ribose 5-phosphate was as equally effective as ribulose
5-phosphate but dihydroxyacetone phosphate was without
effect. These data seem to suggest that, although it does
interact with the glutaminase complex, ribulose 5-phosphate

The expert technical assistance of Ms. Miwa Kitamura is gratefully
acknowledged. This work was supported in part by a grant from the
Regional Science Promotion Program of Japan Science and Technol-
ogy Corporation.
References
1. Cane, D.V., Hsiung, Y., Cornish, J.A., Robinson, J.K. & Spenser,
I.D. (1998) Biosynthesis of vitamin B6: the oxidation of 4-(phos-
phohydroxy)-
L
-threonine by PdxA. J. Am. Chem. Soc. 120, 1936–
1937.
2. Cane, D.V., Du, S.C., Robinson, K., Hsiung, Y. & Spenser, I.D.
(1999) Biosynthesis of vitamin B6: enzymatic conversion of
1-deoxy-
D
-xylulose-5-phosphate to pyridoxol phosphate. J. Am.
Chem. Soc. 121, 7722–7723.
3. Laber, B., Maurer, W., Scharf, S., Stepusin, K. & Schmidt, F.S.
(1999) Vitamin B6 biosynthesis: formation of pyridoxine
5¢-phosphate from 4-(phosphohydroxy)-
L
-threonine and 1-deoxy-
D
-xylulose-5-phosphate by PdxA and PdxJ protein. FEBS Lett.
449, 45–48.
4.Tazuya,K.,Adachi,Y.,Masuda,K.,Yamada,K.&
Kumaoka, H. (1995) Origin of the nitrogen atom of pyridoxine
in Saccharomyces cerevisiae. Biochim. Biophys. Acta 1244,
113–116.
5. Lam, H.M. & Winkler, M.E. (1990) Metabolic relationships

Laboratory, Cold Spring Harbor, NY.
14. Treco, D.A. (1987) Preparation of yeast RNA, DNA, and pro-
teins. In Current Protocols in Molecular Biology (Ausubel, F.M.,
Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith,
J.A., & Struhl, K., eds) pp. 13.11.1–13.11.5. Current Protocols,
John Wiley & Sons, NY.
15. Fink, G.R. (1975) The biochemical genetics of yeast. Methods
Enzymol. 17A, 59–78.
Ó FEBS 2004 Pyridoxine biosynthesis in yeast (Eur. J. Biochem. 271) 751
16. Ito, H., Fukuda, Y., Murata, K. & Kimura, A. (1983) Transfor-
mation of intact yeast cells treated with alkali cations. J. Bacteriol.
153, 163–168.
17. Van Kuilenburg, A.B., Elzinga, L., Van den Berg, A.A., Slinger-
land, R.J. & Van Gennip, A.H. (1994) A fast and novel assay of
CTP synthetase. Evidence for hysteretic properties of the mam-
malian enzyme. Anticancer Res. 14, 411–415.
18. Niederwieser, A. (1972) Thin-layer chromatography of amino
acids and derivatives. Methods Enzymol. 25B, 60–99.
19. Chaparian, M.G. & Evans, D.R. (1991) The catalytic mechanism
of the amidotransferase domain of the Syrian hamster multi-
functional protein CAD. Evidence for a CAD-glutamyl covalent
intermediate in the formation of carbamyl phosphate. J. Boil.
Chem. 266, 3387–3395.
20. Ehrenshaft, M. & Daub, M.E. (2001) Isolation of PDX2,asecond
novel gene in the pyridoxine biosynthesis pathway of eukaryotes,
archaebacteria, and a subset of eubacteria. J. Bacteriol. 183,
3383–3390.
21. Giaever, G., Chu, A.M., Ni, L., Connelly, C., Riles, L., Veron-
neau,S.,Dow,S.,Lucau-Danila,A.,Anderson,K.,Andre,B.
et al. (2002) Functional profiling of the Saccharomyces cerevisiae