Biosynthesis of vitamin B6: Enzymatic conversion of 1-deoxy-D-xylulose- 5-phosphate to pyridoxol phosphate [14]

Full text of DOI:10.1021/ja9914947

7722
J. Am. Chem. Soc. 1999, 121, 7722-7723
Biosynthesis of Vitamin B6: Enzymatic Conversion
of 1-Deoxy-D-xylulose-5-phosphate to Pyridoxol
Phosphate
Scheme 1
,
†
†
†
David E. Cane,* Shoucheng Du, J. Kenneth Robinson,
†
‡
Yuju Hsiung, and Ian D. Spenser
Department of Chemistry, Box H, Brown UniVersity,
ProVidence, Rhode Island, 02912-9108, and
Department of Chemistry, McMaster UniVersity, Hamilton,
Ontario, L8S 4M1, Canada
Scheme 2
ReceiVed May 6, 1999
Pyridoxal 5′-phosphate (1, PLP) is a central coenzyme in amino
1a
acid metabolism, while its congener pyridoxamine 5′-phosphate
2, PMP) also plays an important role in the biosynthesis of
(
1b
deoxysugars. PMP is generated from PLP by transamination,
while the latter is derived from pyridoxine 5′-phosphate (3, PNP)
by an O
known as PdxH. (Scheme 1). Pyridoxine itself (4, PN, vitamin
) can be converted to PNP by an ATP-dependent kinase, PdxK,3
2
-dependent oxidation catalyzed by the flavoenzyme
2
B
6
which can also mediate the formation of PLP and PMP from the
corresponding free alcohols.4
In Escherichia coli PNP is formed from two building blocks
derived from 4-hydroxythreonine (5, 4-HT) and 1-deoxy-D-
xylulose (6, dX). (Scheme 2). The latter metabolite has attracted
The E. coli proteins PdxA and PdxJ were cloned, overex-
pressed, and purified as previously described. The cosubstrate
-phosphohydroxy-L-threonine (8) was also prepared as previously
8
5
4
intensive interest with the recognition that 1-deoxy-D-xylulose-
6
a
described by phosphorylation of synthetic 4-hydroxythreonine
using homoserine kinase8 (Scheme 2). Recombinant E. coli
deoxyxylulose-5-phosphate synthase was used to prepare the
5
1
-phosphate (7, dXP) is also a precursor of thiamin (vitamin B )
,10
6b
as well as many bacterial and plant isoprenoids. Genetic studies
have implicated two gene products, PdxA and PdxJ, as being
responsible for the formation of PNP from 4-HT and dX or their
derivatives. Recently, we reported that PdxA is responsible for
the NAD -dependent oxidation of 4-phosphohydroxy-L-threonine
14
requisite samples of dXP (7) and [2- C]dXP from pyruvate and
14
7
[2- C]pyruvate, respectively, and glyceraldehyde-3-phosphate in
the presence of thiamin diphosphate.11 The corresponding samples
+
14
of unlabeled deoxyxylulose (6) and [2- C]dX were readily
obtained by treatment of the respective samples of dXP (7) with
acid phosphatase.
(
8, 4-PHT) to a product tentatively identified as 3-phosphohy-
8,9
droxy-1-aminoacetone (9). In the absence of additional sub-
strates, 9 is unstable and undergoes dimerization and loss of
phosphate to give the pyrazine derivative 10. We now report that
PdxJ catalyzes the condensation of 1-deoxy-D-xylulose-5-
phosphate (7) and 3-phosphohydroxy-1-aminoacetone (9) to yield
PNP.
14
A mixture of [2- C]-1-deoxy-D-xylulose-5-phosphate (7) and
4
-phosphohydroxy-L-threonine (8) was incubated with desalted
+
PdxA and PdxJ in the presence of NAD in 100 mM Tris, pH
7
.5, at 37 °C (Scheme 2). Aliquots of the mixture were
periodically withdrawn at times up to 90 min and analyzed by
silica gel TLC, visualizing the radioactivity on the plates by
phosphoimaging. Within 5 min, the spot corresponding to the
†
Brown University.
‡
McMaster University
(
6
1) (a) Dolphin, D.; Poulson, R.; Avramovic, O. Vitamin B Pyridoxal
14
substrate [2- C]dXP had disappeared and was replaced by a spot
with an R corresponding to pyridoxol-5′-phosphate (3, PNP).
Control experiments showed that no PNP was formed in the
Phosphate. Chemical, Biochemical, and Medical Aspects; Wiley-Inter-
science: New York, 1986. (b) Burns, K. D.; Pieper, P. A.; Liu, H.-W.;
Stankovich, M. T. Biochemistry 1996, 35, 7879-7889.
F
(
2) Zhao, G.; Winkler, M. E. J. Bacteriol. 1995, 177, 883-891. Notheis,
1
4
absence of either enzyme or 4-PHT. By contrast, when [2- C]-
C.; Drewke, C.; Leistner, E. Biochim. Biophys. Acta 1995, 1247, 265-271.
14
(
3) Yang, Y.; Zhao, G.; Winkler, M. E. FEMS Microbiol. Lett. 1996, 141,
deoxyxylulose was used in place of [2- C]dXP, no pyridoxol
phosphate or pyridoxol could be detected. To gain insight into
the mechanism of the reaction, we monitored a preparative-scale
incubation by 31P NMR over a period of 100 min at room
8
9-95.
(
4) For comprehensive reviews see: Spenser, I. D. Nat. Prod. Rep. 1995,
1
2, 555-565. Hill, R. E.; Spenser, I. D. In Escherichia coli and Salmonella
typhimurium. Cellular and Molecular Biology, 2nd ed.; Neidhardt, F. C., Ed.;
American Society of Microbiology: Washington D.C., 1996. Vol. 1, pp 695-
7
03.
(
(9) Interestingly, although use of PdxA as a query and either the BLAST
or TFASTA algorithms to search the combined Genbank and EMBL databases
turned up no significant matches to known proteins, use of the iterative PSI-
BLAST algorithm (Altschul, S. F.; Madden, T. F.; Sch a¨ ffer, A. A.; Zhang, J.;
Zhang, Z.; Miller, W.; Lipman, D. J. Nucleic Acids Res. 1997, 25, 3389-
3402; http://www.ncbi.nlm.nih.gov/cgi-bin/BLAST/nph-psi•blast) to search
the NCBI nonredundant (nr) peptide sequence database revealed after 2
iterations ∼30-35% similarity over 150-200 amino acids both to isocitrate
dehydrogenases and to 3-isopropylmalate dehydrogenases, enzymes that also
catalyze nicotinamide-dependent decarboxylative oxidations of â-hydroxy
acids.
5) (a) Kennedy, I. A.; Hill, R. E.; Pauloski, R. M.; Sayer, B. G.; Spenser,
I. D. J. Am. Chem. Soc. 1995, 117, 1661-1662. Hill, R. E.; Himmeldirk, K.;
Kennedy, I. A.; Pauloski, R. M.; Sayer, B. G.; Wolf, E.; Spenser, I. D. J.
Biol. Chem. 1996, 271, 30426-30435. (b) Zhao, G.; Winkler, M. E. FEMS
Microbiol. Lett. 1996, 135, 275-280. (c) Drewke, C.; Klein, M.; Clade, D.;
Arenz, A.; M u¨ ller, R.; Leistner, E. FEBS Lett. 1996, 390, 179-182.
(6) (a) Begley, T. P. Nat. Prod. Rep. 1996, 13, 177-185. (b) Rohmer, M.
In ComprehensiVe Natural Products Chemistry. Isoprenoids Including Caro-
tenoids and Steroids; Cane, D. E., Ed.; Elsevier: Oxford, U.K., 1999; Vol. 2,
pp 45-67. Schwarz, M.; Arigoni, D. In ComprehensiVe Natural Products
Chemistry. Isoprenoids Including Carotenoids and Steroids; Cane, D. E., Ed.;
Elsevier: Oxford, U.K., 1999; Vol. 2, pp 367-400.
(10) Laber, B.; Gerbling, K.-P.; Harde, C.; Neff, K.-H.; Norhoff, E.;
Pohlenz, H. D. Biochemistry 1994, 33, 3413-3423.
(
7) (a) Lam, H.-M.; Tancula, E.; Dempsey, W. B.; Winkler, M. E. J.
(11) Sprenger, G. A.; Sch o¨ rken, U.; Wiegert, T.; Grolle, S.; de Graaf, A.
A.; Taylor, S. V.; Begley, T. P.; Bringer-Meyer, S.; Sahm, H. Proc. Natl.
Acad. Sci. U.S.A. 1997, 94, 12857-12862. Lois, L. M.; Campos, N.; Rosa
Putra, S.; Danielsen, K.; Rohmer, M.; Boronat, A. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 2105-2110. Glyceraldehyde-3-phosphate was generated in situ from
fructose-1,6-diphosphate using aldolase and triose phosphate isomerase.
Bacteriol. 1992, 174, 1554-1567. (b) Takiff, H. E.; Baker, T.; Copeland, T.;
Chen, S.-M.; Court, D. L. J. Bacteriol. 1992, 174, 1544-1553. (c) Roa, B.
B.; Connolly, D. M.; Winkler, M. E. J. Bacteriol. 1989, 171, 4767-4777.
(
8) Cane, D. E.; Hsiung, Y.; Cornish, J. A.; Robinson, J. K.; Spenser, I. D.
J. Am. Chem. Soc. 1998, 120, 1936-1937.
1
0.1021/ja9914947 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/06/1999

Communications to the Editor
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7723
Scheme 3
Figure 1. Time course of 31P NMR spectra showing the conversion of
DXP and PHT to PNP and P
The reaction was carried out at room temperature in 0.7 mL of Tris/HCl
buffer (0.1 M, pH 7.5) which contained D O (0.15 mL), DXP (1.21 mg,
.19 µmol), PHT (1.05 mg, 4.91 µmol), NAD (9.85 µmol), desalted PdxA
0.6 mg), and PdxJ (0.6 mg).
i
by PdxA and PdxJ in the presence of NAD.
2
4
(
in a total vol of 1 mL at 37 °C in the presence of varying
concentrations of 1-deoxy-D-xylulose-5-phosphate from 8 to 160
µM. Lactate dehydrogenase-catalyzed reduction of pyruvate
consumed NADH as it was formed and allowed monitoring of
the progress of the reaction by direct detection of pyridoxol
temperature using dXP (4.14 µmol), 4-PHT (4.91 µmol), and
+
NAD (9.85 µmol) in the presence of 600 µg of desalted PdxA
and 600 µg of desalted PdxJ in 0.1 M Tris, pH 7.5, plus 0.15 mL
of D O in a total volume of 0.7 mL. As illustrated in Figure 1,
2
consumption of 4-PHT and dXP was accompanied by the
simultaneous appearance of peaks of equal intensity at 4.79 and
-1
-1
12
phosphate (λmax 324 nm, ꢀ 7330 M cm ; 0.1 M Tris, pH 7.5).
The observed initial rate data were fitted directly to the Michae-
3
.42 ppm, corresponding to the stoichiometric formation of
lis-Menten equation, and the analysis was carried out in duplicate,
pyridoxol phosphate (3, PNP) and inorganic phosphate (P ), the
I
giving a K
.17 ( 0.12 min .
The PdxA-catalyzed oxidation of 4-phosphohydroxythreonine8
was inhibited by the addition of 1.0 mM EDTA. Addition of 1.0
m
(app) for DXP (7) of 26.9 ( 2.3 µM and a kcat of
identities of both of which were verified by comparison with
authentic samples.
To confirm further the identity of the enzymatic reaction
-
1
4
product, we applied the preparative scale incubation mixture from
2
+
2+
2+
2+
the 31P NMR experiment to an AG1 × 8 ion-exchange column,
mM Mn , Co , Mg , or Ca fully restored activity, while 1
mM Ni2 or 2 mM Zn restored only half of the original PdxA
+
2+
which was washed with water and then 6 mM HCl. The fractions
+
activity.
containing pyridoxol phosphate contaminated with NAD were
The formation of pyridoxol phosphate from 1-deoxy-D-xylu-
lose-5-phosphate and 9 catalyzed by PdxJ can be explained by
the mechanism illustrated in Scheme 3, which is a variant of
mechanisms proposed earlier.4 On the basis of the observations
that pyridoxol phosphate and inorganic phosphate are formed in
stoichiometric proportions, that the free alcohol deoxyxylulose
is not a substrate for PdxJ, and that the enzyme does not release
pyridoxol-4′,5′-diphosphate as an initial product, we propose that
PdxJ has a phosphomutase activity which transfers the 5-phos-
phoryl group of dXP (7) to the 4-hydroxyl at some point in the
reaction. The resulting 4-phospho intermediate 11 would then be
able to eliminate phosphate rather than water to generate the enol
reapplied to the AG1 × 8 column in a volume of 20 mL and
eluted with 2 mM HCl, giving pure pyridoxol phosphate (3). The
1
4
00 MHz H NMR spectrum of 3 (HCl salt, D
2
O) showed a
), a singlet at 4.88 ppm (2 H, CH O),
a doublet at 4.90 ppm (2 H, JHP)7.55 Hz, CH OP), and a singlet
at 8.05 (1 H, H-6), and the 162 MHz 3 P NMR spectrum of 3
HCl salt, D O) showed a single peak at 1.20 ppm (t, J ) 7.55
,8
singlet at δ 2.49 (3 H, CH
3
2
2
1
(
2
1
31
Hz). Both the H and P spectra of enzymatically generated 3
were identical to those of an authentic sample of PNP. The TOF
electrospray MS of enzymatically generated 3 also matched that
8 12 6
of authentic PNP with peaks at m/z 249 (C H O NP), 248 (base
peak), and 230. Finally the two samples were identical by UV
12 that is thought to undergo cyclization. The roles of PdxJ and
(λ
max 324 nm) and TLC.
On the basis of the results of the 31P NMR experiment, it was
PdxA in the formation of the pyridoxine ring of vitamin B in E.
6
coli have now been established. The substrates for PdxA and PdxJ
are thus 4-phosphohydroxy-L-threonine (8) and 1-deoxy-D-xylu-
lose-5-phosphate (7), rather than the corresponding free alcohols
evident that the formation of pyridoxol phosphate catalyzed by
PdxJ does not involve initial formation of either pyridoxal-4′,5′-
diphosphate or the free alcohol deoxyxylulose, consistent with
the fact that deoxyxylulose is not a substrate for the PdxJ-
catalyzed reaction. In a separate 31P NMR experiment, 4-phos-
6
and 5, respectively, neither of which are involved in the
13
biosynthesis of pyridoxol phosphate. We suggest that the gene
products PdxA and PdxJ should henceforth be known as 4-phos-
phohydroxy threonine dehydrogenase and pyridoxol phosphate
synthase, respectively.
+
phohydroxy-L-threonine (8) was initially incubated with NAD
and PdxA under the conditions described above, resulting in the
appearance of a new 31P signal at 4.90 ppm assumed to be due to
3
-phosphohydroxy-1-aminoacetone (9), along with inorganic
Acknowledgment. This work was supported by NIH Grants GM22172
to D.E.C. and GM50788 to I.D.S. We thank Dr. Bernd Laber for kindly
informing us of his results in advance of publication.13
phosphate. Addition of 1-deoxyxylulose-5-phosphate (7) and PdxJ
then resulted in depletion of 9 and concomitant formation of
pyridoxol phosphate. Incubation of 1-deoxyxylulose-5-phosphate
JA9914947
(
7) alone with purified PdxJ for 5 h with monitoring by 31P NMR
(
12) Peterson, E. A.; Sober, H. A. J. Am. Chem. Soc. 1954, 76, 169-175:
did not lead to the formation of any detectable product. Taken
together, the latter results are completely consistent with the
postulated role of 9 as the cosubstrate for PdxJ and indicate that
pyridoxol phosphate formation is catalyzed by PdxJ, not PdxA.
To determine the steady-state kinetic parameters of the PdxJ-
catalyzed reaction, we used a coupled UV assay. Purified PdxA
-1 -1
ꢀ
) 7400 M cm (0.1 M sodium phosphate, pH 7.0). A control experiment
was carried out to ensure that no NADH (λmax 340 nm) would accumulate
under the conditions of the assay that could interfere with the detection of
+
PNP at 324 nm. Thus, when PdxA was added to a mixture of 4-PHT, NAD ,
pyruvate, and LDH in the same proportions used in the assay, no absorption
at 340 nm was detected.
(13) After completion of this work, we were informed by Dr. Bernd Laber
of Hoechst Schering AgrEvo Gmbh that he and his colleagues had also
independently observed the formation of PNP from DXP catalyzed by PdxJ
and that their results were in press (Laber, B.; Maurer, W.; Scharf, S.; Stepusin,
K.; Schmidt, F. S. FEBS Lett. 1999, 449, 45-48.)
(
400 µg, 11.4 nmol) and PdxJ (14 µg, 0.53 nmol) were incubated
+
with 4-phosphohydroxy-L-threonine (400 µM), NAD (500 µM)
and lactate dehydrogenase (2.5 units) in 0.1 M Tris‚HCl (pH 7.5)