Protein purification and function assignment of the epoxidase catalyzing the formation of fosfomycin [10]

Full text of DOI:10.1021/ja004153y

J. Am. Chem. Soc. 2001, 123, 4619-4620
4619
Protein Purification and Function Assignment of the
Epoxidase Catalyzing the Formation of Fosfomycin
Scheme 1
†
†
‡
Pinghua Liu, Kazuo Murakami, Takayuki Seki,
†
†
‡
Xuemei He, Siu-Man Yeung, Tomohisa Kuzuyama,
Haruo Seto, and Hung-wen Liu*
‡
,†
DiVision of Medicinal Chemistry
College of Pharmacy, and
Department of Chemistry and Biochemistry
UniVersity of Texas, Austin, Texas 78712
Department of Chemistry, UniVersity of Minnesota
Minneapolis, Minnesota 55455
Institute of Molecular and Cellular Biosciences
UniVersity of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan
Scheme 2
ReceiVed December 4, 2000
(
1R,2S)-Epoxypropylphosphonic acid (1), also known as fos-
1
fomycin, is a clinically useful antibiotic. Its biological target has
been identified as UDP-GlcNAc-O-enolpyruvoyl transferase,2
which catalyzes the attachment of phosphoenolpyruvate (PEP)
3
to UDP-GlcNAc, a key step in the assembly of the peptidoglycan
layer in bacterial cell wall. Early studies had shown that the
biosynthesis of fosfomycin begins with isomerization of PEP (2)
4
to phosphonopyruvate (3) (Scheme 1). Feeding experiments also
established that the immediate precursor of fosfomycin is (S)-2-
hydroxypropylphosphonic acid (HPP, 4). On the basis of these
The gene for Fom4 was amplified by polymerase chain reaction
(PCR) and cloned into a pQE30 expression vector. The ensuing
plasmid, pQE5110, was used to transform E. coli M15 host cells.
Growth of the recombinant strain in LB medium led to overpro-
duction of Fom4. This N-terminal His-tagged enzyme was purified
and exhibited no apparent absorption above 300 nm. Inductive
coupled plasma (ICP) analysis of this purified protein also failed
to detect any redox-active metal ions in significant quantities.
5
findings, a minimum of four enzymatic steps had been proposed
6
for the biogenesis of fosfomycin (Scheme 1). Recently, Seto and
co-workers had cloned the entire fosfomycin biosynthetic gene
7
cluster from Streptomyces wedmorensis, and part of the cluster
8
from Pseudomonas syringae PB-5123. Expression of orf3 of P.
syringae in Escherichia coli and a preliminary activity assay led
to the tentative assignment of orf3 as encoding for the HPP
To determine whether Fom4 is the desired epoxidase, (S)-HPP
8
9,10
epoxidase. In a joint effort, we have expressed the orf3 equivalent
(4) was synthesized,
and tested for its competence as the
in S. wedmorensis (fom4) and purified the encoded protein
substrate for Fom4. To our disappointment, no fosfomycin could
8
(Fom4). In addition, we have also developed an efficient assay
be detected by NMR in this incubation. A bioassay was then
for Fom4, which now allows the first unambiguous assignment
of Fom4 as the desired HPP epoxidase. Reported herein are the
initial characterization of this enzyme and the implications for
its mode of catalysis.
used to evaluate whether various metal ions and common
biological reducing agents are required for Fom4 to function as
an epoxidase.11 In this experiment, a paper disk soaked with the
assay mixture was placed in direct contact with a lawn of E. coli
K12 HW8235 grown on nutrient (LB) agar. When fosfomycin
was present in the assay mixture, an inhibition zone was visible
after a few hours of incubation. Using this sensitive bioautography
*
To whom correspondence should be addressed. Fax: 512-471-2746.
E-mail: h.w.liu@mail.utexas.edu.
†
University of Texas and University of Minnesota.
University of Tokyo.
‡
4 2 4 2
method, it was found that both Fe(NH ) (SO ) and NAD(P)H
(
1) (a) Hendlin, D.; Stapley, D. O.; Jackson, M.; Wallick, H.; Miller, A.
are essential for Fom4 to convert 4 to fosfomycin. These findings
provided initial evidence revealing that Fom4 is the desired
epoxidase, and its catalysis is iron-dependent. However, no
fosfomycin production was discernible by NMR even when all
K.; Wolf, F. J.; Miller, T. W.; Chaiet, L.; Kahan, F. M.; Foltz, E. L.; Woodruff,
H. B.; Mata, J. M.; Hernandez, S.; Mochales, S. Science 1969, 166, 122-
1
23. (b) Christensen, B. G.; Leanza, W. J.; Beattie, T. R.; Patchett, A. A.;
Arison, B. H.; Ormond, R. E.; Kuehl, F. A., Jr.; Albers-Schonberg, G.;
Jardetzky, O. Science 1969, 166, 123-125.
1
2
(
2) Kahan, F. M.; Kahan J. S.; Cassidy, P. J.; Kropp, H. Ann. N.Y. Acad.
of the above components were included in the incubation.
Since this epoxidase is iron-dependent, to avert any complica-
tion that may be associated with the fused His -tag, the fom4 gene
Sci. 1974, 235, 364-386.
(
3) (a) Marquardt, J. L.; Brown, E. D.; Lane, W. S.; Haley, T. M.; Ichikawa,
Y.; Wong, C. H.; Walsh, C. T. Biochemistry 1994, 33, 10646-10651. (b)
6
Kim, D. H.; Lees, W. J.; Haley, T. M.; Walsh, C. T. J. Am. Chem. Soc. 1995,
was cloned into a pET24b vector to express the epoxidase in its
1
17, 1494-1502. (c) Schonbrum, E.; Sack, S.; Eschenburg, S.; Perrakis, A.;
13
wild-type form. The purified protein was eluted as a single peak
Krekel, F.; Amrhein, N.; Mandelkow, E. Structure 1996, 4, 1065-1075. (d)
with an apparent M
was calculated based on a monomeric M
r
of 89 kDa from an FPLC S200 column, and
of 21210 Da to be
Schonbrum, E.; Svergun, D. I.; Amrhein, N.; Koch, M. H. J. Eur. J. Biochem.
1
998, 253, 406-412.
(
r
4) (a) Seidel, H. M.; Freeman, S.; Seto, H.; Knowles, J. Nature 1988,
3
1
1
35, 457-458. (b) Hidaka, T.; Iwakura, H.; Imai, S.; Seto, H. J. Antibiot.
992, 45, 1008-1010. (c) Kim, J. B.; Dunnaway-Mariano, D. Biochemistry
996, 35, 5, 4628-4635.
(9) Hammerschmidt, F. Monatsh. Chem. 1991, 122, 389-398.
1
(10) Spectral data of 4: H NMR (500 MHz, D
2
O) δ 0.97 (3H, d, J ) 6.5
2
2-H); C NMR (75.5 MHz, D O) δ 22.1 (C-3), 37.9 (d, J ) 132 Hz, C-1),
31
Hz, 3-H), 1.39 (2H, ddd, J ) 18.0, 6.6, 15.3, 1-H), 3.80 (1H, dm, J ) 6.5,
1
3
(
5) (a) Seto, H.; Hidaka, T.; Kuzuyama, T.; Shibahara, S.; Usui, T.;
Sakanaka, O.; Imai, S. J. Antibiot. 1991, 44, 1286-1288. (b) Hammerschmidt,
F. J. Chem. Soc., Perkin Trans. 1 1991, 1993-1996.
2
65.0 (d, J ) 2, C-2); P NMR (121 MHz, D O) δ 19.9 (s).
(11) The redox active metals that had been tested include Fe, Cu, Co, Ni,
and Mn. Ascorbate, NADH or NADPH were used as reducing agents in the
incubation mixture.
(
6) (a) Kuzuyama, T.; Hidaka, T.; Kamigiri, K.; Imai, S.; Seto, H. J.
Antibiot. 1992, 45, 1812-1814. (b) Kuzuyama, T.; Hidaka, T.; Imai, S.; Seto,
H. J. Antibiot. 1993, 46, 1478-1480.
4 2
(12) The reaction mixture contained 21.8 mM NADH, 1.0 mM Fe(NH ) -
(7) Hidaka, T.; Goda, M.; Kuzuyama, T.; Takei, N.; Hidaka, M.; Seto, H.
4 2
(SO ) , 10.5 mM HPP (4), and 0.8 mM purified enzyme in 100 µL of 20 mM
Mol. Gen. Genet. 1995, 249, 274-280.
8) Kuzuyama, T.; Seki T.; Kobayashi, S.; Hidaka, T.; Seto, H. Biosci.
Biotechnol. Biochem. 1999, 63, 2222-2224.
Tris‚HCl buffer (pH 7.5). The reaction was run at room temperature for 4 h,
(
and then the mixture was lyophilized. The resulting powder was redissolved
1
31
2
in D O and analyzed by H and P NMR.
1
0.1021/ja004153y CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/21/2001

4620 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Communications to the Editor
Scheme 3
for the reductase has greatly facilitated the initial investigation
of HPP epoxidase. It should be noted that no fosfomycin was
produced when the reaction was conducted anaerobically, and
no H
thus concluded that the oxidant of this reaction is O
2
O
2
formation could be detected during the reaction. It is
, which is
2
reduced by four electrons to water during the catalysis.
To test whether the oxiranyl ring formation in the biosynthesis
of fosfomyin proceeds via a net dehydrogenation as implicated
by earlier feeding experiments,5 the O-labeled HPP (5) was
b,19
18
1
8
prepared (Scheme 2). A 1:1 mixture of unlabeled 4 and O-
labeled 5 was incubated with purified epoxidase in the presence
of FMN. The product was purified by DEAE Sepharose eluted
with a linear gradient of 0-0.5 M NH
4
HCO
3
, and analyzed by
and C
13
C NMR. Unlike the unlabeled fosfomycin whose C
â
R
signals appear as a singlet and doublet (JC-P ) 176 Hz),
respectively,20 the corresponding signals in fosfomycin derived
from this mixed sample were further split. The complication
derives from the overlapping of two sets of signals, one for the
unlabeled fosfomycin and the other for the labeled fosfomycin.
The upfield shift by 0.018-0.025 ppm for the signals of the
tetrameric. N-terminal amino acid sequencing confirmed its
identity.14 Similar to its His-tagged homologue, the as-isolated
18
21
labeled species is due to the O isotope substitution effect. Thus,
these data firmly establish the retention of 18O label in the product.
Most enzymes capable of catalyzing epoxidation rely on a
mechanism of oxygen insertion to an olefinic substrate to form
the oxirane.22 The epoxidation catalyzed by HPP epoxidase is
unusual since it is effectively a dehydrogenation of a secondary
alcohol (4f1). Two closely related examples are the epoxide ring
formation catalyzed by hyoscyamine 6â-hydroxylase involved in
the biosynthesis of scopolamine,23 and the ring closure of
proclavaminic acid catalyzed by clavaminate synthase. However,
both enzymes belong to the family of R-ketoglutarate-dependent
non-heme iron oxygenases.25 In contrast to these two examples,
HPP epoxidase is R-ketoglutarate-independent, and its activity
is adversely sensitive to ascorbate.26 Thus, the cyclization of HPP
protein is transparent above 300 nm and requires NAD(P)H and
exogenous ferrous ion for activity. Upon treatment with Fe(NH
4 2
) -
(
SO , the apo-enzyme could be reconstituted to have one iron
4 2
)
per monomer. However, the catalytic efficiency of the reconsti-
tuted protein remained poor. Considering that additional electron
mediators may be required, the effect of a variety of redox-active
cofactors on the catalysis of this protein was investigated. Inter-
estingly, it was discovered that the addition of FMN or FAD
greatly enhanced the production of fosfomycin (100 nmol‚
24
-1
-1 15
min ‚mg ). The effect, however, was catalytic, and reconstitu-
tion failed to incorporate FMN/FAD into the epoxidase. Hence,
the flavin coenzyme is unlikely an integral part of the epoxidase
itself, but may act as a surrogate for an electron-transfer protein
involved in the overall catalysis in vivo.
(4) to fosfomycin (1) clearly represents an intriguing conversion
3
Indeed, E , an NADH-dependent [2Fe-2S]-containing flavopro-
beyond the scope entailed by common biological epoxidation and
C-O bond formation. One of several plausible mechanisms that
are consistent with the available data is shown in Scheme 3. The
reaction may be initiated with an electron transfer from the
putative reductase followed by oxygen activation to generate an
iron-coordinated superoxide radical 6. Abstraction of an R-H atom
of 6 in conjunction with a proton-coupled electron transfer may
lead to a high valent iron species 7.27 Radical-induced homolytic
cleavage of the Fe-O bond in 7 will produce 1 and also regenerate
the iron core. Future research on this enzyme is aimed at further
elucidation of the mechanism of this unique epoxidase and
isolation as well as characterization of its putative reductase.
tein reductase from the 3,6-dideoxyhexose biosynthetic pathway
in Yersinia pseudotuberculosis,16 was found to be a better electron
17
mediator than FMN (activity increased >20-fold). This result
clearly indicated that inclusion of a reductase is necessary to fully
reconstitute the epoxidase activity. Hence, conversion of 4 to
fosfomycin is most likely catalyzed by a complex of HPP
epoxidase and its reductase with NAD(P)H as the reductant (see
Scheme 3). Interestingly, OrfD, the encoded product of another
7
gene (orfD) just downstream of fom4 in this cluster, was found
by the yeast two-hybrid system18 to form a complex with Fom4
in vivo. Since all other open reading frames with unknown
function in this cluster gave negative results in the yeast two-
hybrid experiment, the physiological reductase for Fom4 is likely
encoded by orfD. Although study of the catalytic details of
epoxidation must await the isolation and characterization of the
Acknowledgment. We thank Professors Lawrence Que, Jr. and John
Lipscomb and Dr. Aimin Liu for helpful discussion.This work was
supported in part by the National Institutes of Health grant GM40541.
3
Fom4 reductase, the fact that FMN or E is a convenient substitute
JA004153Y
(
13) The ensuing construct, pPL001, was used to transform BL21 (DE3)
cells, and the recombinant strain was grown in LB medium at 37 °C under
(19) Hammerschmidt, F.; Bovermann, G.; Bayer, K. Liebigs Ann. Chem.
IPTG induction (0.1 mM). The overproduced Fom4 was purified to near
1990, 1055-1061.
1
homogeneity by a sequence of 40-60% (NH
4
)
2
SO
4
fractionation, DEAE-
(20) Spectral data of 1: H NMR (500 MHz, D
2
O) δ 1.21 (3H, d, J ) 5.5.Hz,
1
3
Sepharose, and FPLC Mono Q chromatography.
3-H), 2.57 (1H, dd, J ) 5.0, 14.0, 1-H), 3.03 (1H, dm, J ) 5.0, 2-H);
NMR (75.5 MHz, D O) δ 13.2 (C-3), 54.3 (C-1, d, J ) 176), 54.2 (C-2, s);
P NMR (121 MHz, D O) δ 10.9 (s).
(21) Risley, J. M.; Van Etten, R. L. J. Am. Chem. Soc. 1979, 101, 252-
253.
C
(
14) N-terminal amino acid sequencing confirmed that the first 10 residues,
2
3
1
SNTKTASTGF, of the purified protein are identical to those of the translated
2
fom4 sequence.
(
4 2
15) The reaction mixture contained 21.8 mM NADH, 140 µM Fe(NH ) -
(
SO
4
2
) , 10.5 mM HPP, 200 µM FMN, and 108 µM purified enzyme in 100
(22) Archelas, A.; Furstoss, R. Annu. ReV. Microbiol. 1997, 51, 491-525.
µL of 20 mM Tris‚HCl buffer (pH 7.5). The reaction was run at room
(23) (a) Yun, D.-J.; Hashimoto, T.; Yamada, Y. Biosci. Biotechnol.
Biochem. 1993, 57, 502. (b) Hashimoto, T.; Matsuda, J.; Yamada, Y.; FEBS
Lett. 1993, 329, 35-39.
temperature for 30 min, quenched by adding 100 mM EDTA, and frozen with
1
liquid N
2
. The sample was later thawed and analyzed by H NMR. The activity
-
1
-1
of HPP epoxidase was determined to be 100 nmol‚min ‚mg
16) Miller, V. P.; Thorson, J. S.; Ploux, O.; Lo, S. F.; Liu, H.-w.
Biochemistry 1993, 32, 11934-11942.
17) The contents of the reaction mixture were identical to those described
in ref 15 except for the amount of epoxidase (21.6 µM) and Fe(NH (SO
28 µM), and the addition of 60 µM E . The reaction was run at room
temperature for 5 min, quenched by adding 100 mM EDTA, and frozen with
liquid N . The analysis was also the same as described in ref 15.
18) (a) Fields, S.; Song, O.-k. Nature 1989, 340, 245-246. (b) Vojtek,
A. B.; Hollenberg, S. M.; Cooper, J. A. Cell 1993, 74, 205-214.
.
(24) (a) Baggaley, K. H.; Brown, A. G.; Schofield, C. J. Nat. Prod. Rep.
1997, 14, 309-333. (b) Iwata-Reuyl, D.; Basak, A.; Townsend, C. A. J. Am.
Chem. Soc. 1999, 121, 11356-11368.
(
(
(25) Prescott, A. G.; Lloyd, M. D. Nat. Prod. Rep. 2000, 17, 367-383.
(26) Addition of R-ketoglutarate up to 1 equiv of the substrate in the
incubation mixture showed no effect on the production of fosfomycin.
However, complete inactivation of Fom4 was observed when 5-10-fold of
ascorbate relative to enzyme concentration was added.
(27) Roach, P. L.; Clifton, I. J.; Hensgens, C. M. H.; Shibata, N.; Schofield,
C. J.; Hajdu, J.; Baldwin, J. E. Nature 1997, 387, 827-830.
)
4 2
4 2
)
(
3
2
(