Tài liệu Báo cáo khóa học: Characterization of the products of the genes SNO1 and SNZ1 involved in pyridoxine synthesis in Saccharomyces cerevisiae pptx

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

in a time- and dose-dependent manner and the enzyme was
protected from deactivation by glutamine. These data sug-
gest strongly that the complex of Sno1p and Snz1p is a
glutamine amidotransferase with the former serving as the
glutaminase, although the activity was barely detectable with
Sno1p alone. The function of Snz1p and the amido acceptor
for ammonia remain to be identified.
Keywords: glutamine amidotransferase; pyridoxine biosyn-
thesis; Saccharomyces cerevisiae; SNO1; SNZ1.
Pyridoxal phosphate plays a crucial role in amino acid
metabolism. Pyridoxine and its phosphate are the precur-
sors of pyridoxal phosphate and the major forms of vitamin
B6. Pyridoxine biosynthesis in Escherichia coli has been
studied extensively but only recently has the whole synthetic
pathway been finally established [1–3]. There are organisms
such as budding yeast, Saccharomyces cerevisiae,which
also synthesize pyridoxine but in a different pathway. This
notion is based in part on an observation that the nitrogen
of pyridoxine is derived from the amide group of glutamine
in yeast [4], while glutamate is the source of the ring nitrogen
in E. coli [5]. Recently, two independent groups identified
pyroA and SOR1 (PDX1) as participating in pyridoxine
synthesis in fungi Cercospora nicotianae and Aspergillus
nidulans, respectively [6,7]. They are homologous genes and
their homologues are distributed widely in various organ-
isms, but nothing of their function is known except that
SNZ1, the yeast homologue, works in the stationary phase
of yeast cells together with SNO1 [8]. In addition to these
observations, it was shown recently that a pentose or
pentulose constitutes the skeleton of pyridoxine in yeast

[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

(MATa, leu2, his3, trp1) [12]. The following medium was
used to grow yeast: YPD [1% yeast extract, 2% polypep-
tone, 2% glucose (v/v/v)] and synthetic medium [glucose,
20 g; (NH
4
)
2
SO
4
,1.02g;KH
2
PO
4
, 0.875 g; K
2
HPO
4
,
0.125 g; CaCl
2
ÆH
2
O, 0.02 g; NaCl, 0.01 g; MgSO
4
Æ7H
2
O,
0.05 g; CuSO
4
Æ5H

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:
Abbreviations: APAD, acetylpyridine adenine dinucleotide; APADH,
reduced form of APAD; DON, 6-diazo-5-oxo-
L
-norleucine.

(Received 7 October 2003, revised 2 December 2003,
accepted 23 December 2003)
Eur. J. Biochem. 271, 745–752 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2003.03973.x
200 lg; KI, 100 lg; H
3
BO
3
,500lg; biotin, 2 lg; inositol,
10 mg; nicotinic acid, 0.2 mg and calcium pantothenate,
0.2 mg, per litre]. Where indicated, the following supple-
ments were added to the synthetic media at a final
concentration of 20 mgÆL
)1
: histidine, leucine, tryptophan
and pyridoxine. Solid media contained 2% agar. For
each liquid culture experiment, yeast cells were shaken at
250 r.p.m. in 10 mL of medium at 30 °Cfortheperiodof
time specified. LB medium [1% polypeptone, 0.5% NaCl
and 0.5% yeast extract (v/v/v)] was used to grow E. coli.
General procedures
All of the gene manipulations of S. cerevisiae and E. coli
were carried out using standard methods [13]. Chromo-
somal DNA from S. cerevisiae was prepared according to
the literature [14]. Yeast genomic DNA sequences were
retrieved from a database by using the website (http://
genome-www.stanford.edu/Saccharomyces/). PCR was run
on an Astec PC-700 (Tokyo, Japan) or Biometra T-gradient
thermoblock (Gottingen, Germany). Sequencing of plas-
mids was carried out by the dideoxy chain termination
method on an automatic DNA sequencer model DSQ1000

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

(891 bp) genes were amplified by PCR in one step with the
oligonucleotides shown in Table 1 as primers. Both of the
genes were constructed either with or without a His-tag at
the 3¢ terminus of the coding region. Thus, the stop codon of
SNO1 and SNZ1 was replaced with bases coding for the
sequence LEHHHHH as a C-terminal extension. The PCR
was run as follows: After heating at 94 °C for 5 min, the
following cycle was repeated 25 times; 94 °C, 1 min; 55 °C,
1min;70°C, 1–1.5 min, and finally heated at 72 °Cfor
5 min. Each PCR product was purified by agarose gel
electrophoresis before ligation into pCR2.1-TOPO (Invi-
trogen). After confirming the correct DNA sequence, the
coding region of SNO1 and SNZ1 was excised from the
plasmid and, where the His-tag was present, recloned into
the NdeI/XhoI sites of pET21a or pET24a, respectively.
Where a His-tag was absent, the coding region of both
genes was recloned into the NcoI/BamHI sites of pET21d.
The resulting recombinant plasmids are termed pSNO1H,
pSNZ1H, pSNO1 and pSNZ1, respectively (Table 2).
Protein production and purification
The two genes were expressed in E. coli BL21(DE3)
(Novagen) separately or simultaneously following trans-
formation with one or two of the plasmids prepared
above. Transformants were grown in 1 L of LB medium
in the presence of ampicillin (50 lgÆmL
)1
), kanamycin
(30 lgÆmL
)1
) or both, to late logarithmic phase, whereupon

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

GGATCCTTAATTAGAAACAAACTGTCTGA
TAAAC
P2OH 5¢-
CTCGAGATTAGAAACAAACTGTCTGATAAACC
P1Z 5¢-ATA
CCATGGCTGGAGAAGACTTTAAGATC
P1ZH 5¢-
CATATGACTGGAGACTTTAAGATC
P2Z 5¢-TAT
GGATCCTCACCACCCAATTTCGGAAAG
P2ZH 5¢-
CTCGAGCCACCCAATTTCGGAAAGT
Table 2. Over-expression plasmids for Sno1p and Snz1p prepared in this
study. Symbols O, Z and H stand for Sno1p, Snz1p and His-tag,
respectively.
Name of
plasmid
Primer
VectorForward Backward
pSNO1 P1O P2O pET21d
pSNO1H P1OH P2OH pET24a
pSNZ1 P1Z P2Z pET21d
pSNZ1H P1ZH P2ZH pET24a
746 Y X. Dong et al.(Eur. J. Biochem. 271) Ó FEBS 2004
(buffer A) containing 5 m
M
glutamine, 0.4 m
M
phenyl-
methanesulfonyl fluoride and 0.2 mL of dimethylsulfoxide.

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

95% at this stage and the yield was 25 mg from a 1 L
culture. Sno1p and Snz1p with a His-tag were purified
analogously with a yield of 30 mg each from 1 L of culture.
Protein concentration was determined on the basis of the
molar extinction coefficient of 15.9 and 12.3 m
M
)1
Æcm
)1
at
280 nm for Sno1p and Snz1p, respectively, deduced from
their amino acid compositions.
Sno1p without a His-tag. The washed cells were resus-
pended in 50 mL of 10 m
M
potassium phosphate buffer and
1m
M
EDTA, pH 7.0. The suspension was sonicated and
then centrifuged at 17 000 g for 20 min. The pellet was
taken up in 50 mL of the same buffer and 2 mL of Triton
X-100 and shaken at room temperature for 30 min. The
suspension was centrifuged (17 000 g) for 30 min and the
supernatant discarded. This process was repeated twice
more for the pellet. The washed pellet was mixed with
10 mL of 50 m
M
Tris/HCl containing 10 m
M
dithiothreitol

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

0–500 m
M
NaCl in 50 m
M
potassium phosphate, pH 7.0.
Snz1p was eluted at about 200 m
M
salt. The pooled
fractions (25 mL) were concentrated and then subjected to
gel filtration chromatography as described above to give
14 mg of the desired protein.
Glutaminase assay
The glutaminase activity of the complex of Sno1p with a
His-tag and Snz1p was determined in two steps. Thus,
glutamate formed was converted to 2-oxoglutarate by
glutamate dehydrogenase with acetylpyridine adenine dinu-
cleotide (APAD) as cosubstrate [17]. The glutaminase
reaction was carried out in 0.3 mL of 50 m
M
Tris/HCl,
pH 8.0, in the presence of 1–10 m
M
glutamine and 30 lgof
the complex at 30 °C for 10 min. The sample was then
boiledfor1minandkeptfrozenat)80 °C for subsequent
analysis. To quantitate the amount of glutamate formed,
0.3 mL of the sample 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 (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

glutamine and glutamate were
separately allowed to react with dansyl chloride under
standard conditions [18]. Aliquots (1.5 lLeach)were
spotted on silica gel 60 F
254
(Merck, Darmstadt, Germany)
and developed in chloroform-t-amyl alcohol-glacial acetic
acid (70 : 30 : 3, v/v/v). The dansylated products were
visualized under ultraviolet light.
Inhibition of Sno1p with DON
Inhibition of the complex of Sno1p with a His-tag and
Snz1p by 6-diazo-5-oxo-L-norleucine (DON) was studied
according to the literature [19]. The reaction mixture
contained 0.15 mg of the complex and various concentra-
tions (1–10 m
M
)ofDONin1mLof50m
M
Tris/HCl,
pH 8.0, and was incubated at 30 °C. Aliquots (200 lL)
removed at a specified time were mixed with other
ingredients for glutaminase assay and the remaining activity
of the complex was determined as described above.
Results
Yeast pyridoxine auxotrophs
Pyridoxine auxotrophs of S. cerevisiae were prepared by
ethyl methanesulfonate mutagenesis. Several strains were
tested including D373-1 and in all the cases auxotrophs
showing clear pyridoxine requirements were obtained (data
not shown). After several rounds of screening, around 20

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

either one of the genes was disrupted separately (data not
shown). This observation is consistent with those reported
previously [21]. Presumably, their homologues SNO2,
SNO3, SNZ2 and SNZ3 complement the defect. The
relationship of all of these genes remains to be clarified.
Expression and purification of Sno1p and Snz1p
In light of the similarity in the amino acid sequence of
Sno1p to that of the glutaminase subunit of imidazole
glycerol phosphate synthase [22], Sno1p may possess the
ability to hydrolyze glutamine. In addition, in reference to
the observation that Sno1p and Snz1p act together in the
stationary phase of yeast cells [8] and the result of two-
hybrid analysis [23], they may form a complex. To assess
these possibilities, we embarked on a characterization of
the two proteins.
Over-expression plasmids were prepared for both genes
with or without a C-terminal His-tag (Table 2), as detailed
under Experimental procedures. E. coli cells were trans-
formed with one or two of the recombinant plasmids. The
resulting cells over-expressed the desired protein(s) up to
15% of the total cellular proteins. When Sno1p was
expressed alone, it was kept in the inclusion body,
irrespective of the occurrence of a His-tag (data not shown).
Whereas, Snz1p, expressed either alone or coexpressed
with Sno1p, was found in the soluble fractions. Although
purification procedures differed slightly from sample to
sample, depending on the occurrence of a tag and the
location of the protein in question, Sno1p, Snz1p and their
complex were eventually purified to near homogeneity
(Fig. 3). Typically, 30, 30 and 25 mg of protein were

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

a more rigorous assessment is necessary to prove this
assertion. The glutaminase activity of this complex was
assessed in two ways. First, the glutamate formed was
oxidized to 2-oxoglutarate by APAD and glutamate
dehydrogenase, and the generated APADH was determined
spectroscopically. Second, the glutamate formed was dansy-
lated and analyzed by TLC (Fig. 4). These experiments
consistently pointed to the formation of glutamate, proving
unambiguously that the complex did indeed hydrolyze
glutamine efficiently.
Kinetics of the glutaminase reaction
The kinetics of glutamine hydrolysis mediated by the
complex of Sno1p with a His-tag and Snz1p was determined
at 30 °C. The reaction rate was proportional to protein
concentration over a 5–50 lg range and it was constant up
to at least 1 h (data not shown). The reaction followed the
simple Michaelis–Menten formulation as illustrated in
Fig. 5. The specific activity (V
max
)andK
m
for glutamine
obtained from analysis of these data were 0.48 lmolÆ
min
)1
Æmg
)1
and 3.4 m
M
, respectively. These values seem

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.

Several additives such as ATP were tested as a stabilizer,
but thus far we have not been successful in stabilizing the
enzyme activity. Incidentally, no glutaminase activity was
found for Snz1p alone.
The function of Sno1p
The data shown above revealed that the complex of Sno1p
and Snz1p is a glutamine amidotransferase, with the former
serving as the glutamine-hydrolyzing machinery. This latter
notion is based solely on the sequence homology of Sno1p
with the glutaminase subunit of imidazole glycerol phos-
phate synthase. To test this hypothesis directly, Sno1p and
Snz1p were separately expressed, purified and characterized
as detailed under Experimental procedures. As described
above, Sno1p expressed in E. coli resided in the inclusion
body and hence use of Triton X-100 and urea was necessary
for its solubilization. Even after renaturation, however,
Sno1p exhibited no glutaminase activity, nor did the
addition of 0.5 to 3-fold Snz1p have any effect. It was only
when Sno1p and Snz1p were partially denatured together
and then renatured, that glutaminase activity was observed.
Thus, the two proteins were mixed in 8
M
urea in various
Fig. 4. TLC analysis of glutamate formed from glutamine mediated by
the complex of Sno1p with a His-tag and Snz1p. Glutamate, glutamine
and the reaction mixture were dansylated separately and aliquots
(2.5 nmol each) were applied on silica gel 60 F
254
and developed in
chloroform-t-amyl alcohol-glacial acetic acid (70 : 30 : 3, v/v/v). The

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.

Ó FEBS 2004 Pyridoxine biosynthesis in yeast (Eur. J. Biochem. 271) 749
molar ratios (0.2 : 4 of Sno1p : Snz1p). The urea concen-
tration was lowered gradually to 6, 4, 2, and 1
M
by dialysis.
Finally, the enzyme solution was dialyzed against buffer
without urea, and glutaminase activity was determined. The
mixtures of Sno1p and Snz1p with a molar ratio of the
former greater than 0.5 exhibited partial activity; at a
maximum of 5% of that of the intact complex described
above. This suggests that although Sno1p and Snz1p
associate spontaneously, renaturation of either one or both
of the proteins was incomplete and the complex formation is
slow and/or incomplete under the conditions employed.
Inhibition of the glutaminase activity by DON
Glutamine amidotransferases are inhibited irreversibly by
6-diazo-5-oxo-
L
-norleucine (DON) [24]. Hence, the suscep-
tibility of the complex of Sno1p and Snz1p to DON was
studied under conditions detailed under Experimental
procedures. As shown in Fig. 6, DON deactivated the
enzyme in a time- and dose-dependent manner. Typically,
the half-life of deactivation was 10 min at 10 m
M
DON at
30 °C. It is noted that the inhibition did not go to completion
even at the highest concentration of DON employed but
halted at a point where about half of the enzyme activity is
lost. The reason for this phenomenon is not known but a

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-

ferases, Snz1p may be a synthetase that is mediating the
coupling of ammonia released from glutamine with an
unknown acceptor substrate. In reactions of many amido-
transferases, ammonia is delivered from the site of forma-
tion to the site of coupling through a tunnel that is present
in the synthetase subunit [25,26]. It is hence tempting to
assume that there is such a tunnel in Snz1p as well, although
this notion awaits experimental verification of the structure
Fig. 6. Inhibition of the glutaminase activity of the complex of Sno1p
with a His-tag and Snz1p by 6-diazo-5-oxo-
L
-norleucine (DON). The
reaction mixture contained 150 lgofthecomplexandno(s), 1 (e), 5
(h,) and 10 m
M
(n) DON in the absence (open symbols) and presence
(j)of10 m
M
glutamine in 1 mL of 50 m
M
Tris/HCl, pH 8.0, at 30 °C.
Aliquots (200 lL)werewithdrawnatspecifiedtimesandtheremaining
activity of the complex was determined.
Scheme 1. Putative synthetic pathway to pyridoxine in yeast. Ammonia released from glutamine by Sno1p condenses with an unknown acceptor (?)
or a derivative of ribulose 5-phosphate. This process is mediated by Snz1p but the subsequent steps remain obscure.
750 Y X. Dong et al.(Eur. J. Biochem. 271) Ó FEBS 2004
by means of X-ray crystallography. Incidentally, the amino
acid sequence of Snz1p does not show homology to any
other known proteins and hence its three-dimensional
structure could be unique. In addition, once the amido-

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

is not the direct acceptor of ammonia.
Based on these arguments, the following scheme is
proposed for the synthetic route to pyridoxine in yeast and
related organisms (Scheme 1). Ammonia released from
glutamine undergoes condensation with either a derivative
of ribulose 5-phosphate or an unknown acceptor with a
three-carbon unit to give an aminated intermediate. These
two components eventually undergo coupling and cycliza-
tion to form 2¢-hydroxypyridoxine. If this scheme holds, the
synthetic pathways to pyridoxine of yeast and related
organisms are remarkably similar to that of E. coli, as both
utilize a ketopentose or its derivative as the starting material.
Although it seems certain that 3-amino-1-phosphohydroxy-
propan-2-one, a key component in the E. coli pathway and
generated by the product of gene pdxA, is not involved,
compounds related to it structurally could be the coupling
partner of the ketopentose in yeast.
One characteristic feature of the genes SNO1 and SNZ1
is that they act simultaneously in the stationary phase of
yeast cells [8]. They are assumed to participate in the
metabolism of nucleotides. In this context, the assertion that
pyridoxine may be the precursor of the pyrimidine moiety of
thiamin in yeast is intriguing [27]. In fungi, pyroA and SOR1
(PDX1), homologues of yeast SNZ1, play a defensive role
against reactive oxygen species such as singlet oxygen [6,7].
Taken together, pyridoxine synthesis in various organisms
ranging from Bacillus to plants seems to play a broader
physiological role than simply supplying the cofactor that is
essential for amino acid metabolism.
Acknowledgements

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

between pyridoxine (vitamin B6) and serine biosynthesis in
Escherichia coli K-12. J. Bacteriol. 172, 6518–6528.
6. Osmani, A.H., May, G.S. & Osmani, S.A. (1999) The extremely
conserved pyroA gene of Aspergillus nidulans is required for
pyridoxine synthesis and is required indirectly for resistance to
photosensitizers. J. Biol. Chem. 274, 23565–23569.
7. Ehrenshaft, M., Bilski, P., Li, M.Y., Chignell, C.F. & Daub, M.E.
(1999) A highly conserved sequence is a novel gene involved in
de novo vitamin B6 biosynthesis. Proc. Natl Acad. Sci. USA 96,
9374–9378.
8. Padilla, P.A., Fuge, E.K., Crawford, M.E., Errett, A. & Werner-
Washburne, M. (1998) The highly conserved, coregulated SNO
and SNZ gene families in Saccharomyces cerevisiae respond to
nutrient limitation. J. Bacteriol. 180, 5718–5726.
9. Gupta,R.N.,Hemscheidt,T.,Sayer,B.G.&Spenser,I.D.(2001)
Biosynthesis of vitamin B6 in yeast: incorporation pattern of
glucose. J. Am. Chem. Soc. 123, 11353–11359.
10. Zeidler,J.,Ullah,N.,Gupta,R.N.,Pauloski,R.M.,Sayer,B.G.&
Spenser, I.D. (2002) 2¢-hydroxypyridoxol, a biosynthetic precur-
sor of vitamins B6 and B1 in yeast. J. Am. Chem. Soc. 124, 4542–
4543.
11. Nikawa, J., Sass, P. & Wigler, M. (1987) Cloning and character-
ization of the low-affinity cyclic AMP phosphodiesterase gene of
Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 3629–3636.
12. Kodaki, T. & Yamashita, S. (1989) Characterization of the
methyltransferases in the yeast phosphatidylethanolamine
methylation pathway by selective gene disruption. Eur. J.
Biochem. 185, 243–251.
13. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) In Molecular
Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor

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

genome. Nature 418, 387–391.
22. Klem, T.J. & Davisson, V.J. (1993) Imidazole glycerol phosphate
synthase: the glutamine amidotransferase in histidine biosynthesis.
Biochemistry 32, 5177–5186.
23. Uetz, P., Giot, L., Cagney, G., Mansfield, T.A., Judson, R.S.,
Knight, J.R., Lockshon, D., Narayan, V., Srinivasan, M.,
Pochart, P. et al. (2000) A comprehensive analysis of protein–
protein interactions in Saccharomyces cerevisiae. Nature 403,
623–627.
24. Mei, B. & Zalkin, H. (1989) A cysteine-histidine-aspartate cata-
lytic triad is involved in glutamine amide transfer function in
purF-type glutamine amidotransferases. J. Biol. Chem. 264,
16613–16619.
25. Omi, R., Mizuguchi, H., Goto, M., Miyahara, I., Hayashi, H.,
Kagamiyama, H. & Hirotsu, K. (2002) Structure of imidazole
glycerol phosphate synthase from Thermus thermophilus HB8:
Open-closed conformational change and ammonia tunneling.
J. Biochem. 132, 759–765.
26. Thoden, J.B., Holden, H.M., Wesenberg, G., Raushel, F.M. &
Rayment, I. (1997) Structure of carbamoyl phosphate synthetase:
a journey of 96 A
˚
from substrate to product. Biochemistry 36,
6305–6316.
27. Tazuya, K., Azumi, C., Yamada, K. & Kumaoka, H. (1995)
Pyrimidine moiety of thiamin is biosynthesized from pyridoxine
and histidine in Saccharomyces cerevisiae. Biochem. Mol. Biol. Int.
36, 883–888.
752 Y X. Dong et al.(Eur. J. Biochem. 271) Ó FEBS 2004