Farnesyl diphosphate synthase: The art of compromise between substrate selectivity and stereoselectivity

Article abstract of DOI:10.1021/ja065573b

Farnesyl diphosphate (FPP) synthase catalyzes the consecutive head-to-tail condensations of isopentenyl diphosphate (IPP, C5) with dimethylallyl diphosphate (DMAPP, C₅) and geranyl diphosphate (GPP, C₁₀) to give (E,E)-FPP (C₁₅). The enzyme belongs to a genetically distinct family of chain elongation enzymes that install E-double bonds during each addition of a five-carbon isoprene unit. Analysis of the C₁₀ and C₁₅ products from incubations with avian FPP synthase reveals that small amounts of neryl diphosphate (Z-C₁₀) and (Z,E)-FPP are formed along with the E-isomers during the C₅ → C₁₀ and C₁₀ → C₁₅ reactions. Similar results were obtained for FPP synthase from Escherichia coli, Artemisia tridentata (sage brush), Pyrococcus furiosus, and Methanobacter thermautotrophicus and for GPP and FPP synthesized in vivo by E. coli FPP synthase. When (R)-[2-²H]IPP was a substrate for chain elongation, no deuterium was found in the chain elongation products. In contrast, the deuterium in (S)-[2-²H]IPP was incorporated into all of the products. Thus, the pro-R hydrogen at C2 of IPP is lost when the E- and Z-double bond isomers are formed. The synthesis of Z-double bond isomers by FPP synthase during chain elongation is unexpected for a highly evolved enzyme and probably reflects a compromise between optimizing double bond stereoselectivity and the need to exclude DMAPP from the IPP binding site.

Full text of DOI:10.1021/ja065573b

Published on Web 11/17/2006
Farnesyl Diphosphate Synthase: The Art of Compromise
between Substrate Selectivity and Stereoselectivity
Hirekodathakallu V. Thulasiram and C. Dale Poulter*
Contribution from the Department of Chemistry, UniVersity of Utah, 315 South 1400 East,
RM 2020, Salt Lake City, Utah 84112
Received August 2, 2006; E-mail: poulter@chemistry.utah.edu
Abstract: Farnesyl diphosphate (FPP) synthase catalyzes the consecutive head-to-tail condensations of
isopentenyl diphosphate (IPP, C5) with dimethylallyl diphosphate (DMAPP, C5) and geranyl diphosphate
(GPP, C10) to give (E,E)-FPP (C15). The enzyme belongs to a genetically distinct family of chain elongation
enzymes that install E-double bonds during each addition of a five-carbon isoprene unit. Analysis of the
C10 and C15 products from incubations with avian FPP synthase reveals that small amounts of neryl
diphosphate (Z-C10) and (Z,E)-FPP are formed along with the E-isomers during the C5 f C10 and C10
f
C15 reactions. Similar results were obtained for FPP synthase from Escherichia coli, Artemisia tridentata
(sage brush), Pyrococcus furiosus, and Methanobacter thermautotrophicus and for GPP and FPP
synthesized in vivo by E. coli FPP synthase. When (R)-[2-2H]IPP was a substrate for chain elongation, no
deuterium was found in the chain elongation products. In contrast, the deuterium in (S)-[2-2H]IPP was
incorporated into all of the products. Thus, the pro-R hydrogen at C2 of IPP is lost when the E- and Z-double
bond isomers are formed. The synthesis of Z-double bond isomers by FPP synthase during chain elongation
is unexpected for a highly evolved enzyme and probably reflects a compromise between optimizing double
bond stereoselectivity and the need to exclude DMAPP from the IPP binding site.
Chain elongation is the fundamental building reaction in the
isoprenoid pathway.1 During this process, the growing hydro-
carbon chain in an allylic isoprenoid diphosphate is added to
isopentenyl diphosphate (IPP). These reactions are catalyzed
by polyprenyl diphosphate synthases, a group of prenyltrans-
ferases that can be further subdivided into two families according
to the stereochemistry, E or Z, of the double bond in the newly
added isoprene unit. Members of the two families utilize the
same chemical mechanism for chain elongation but are geneti-
cally unrelated.1 Subgroups within each family are selective for
the size of the allylic substrate selected for chain elongation
and the length of the isoprenoid chain in the final product. Thus,
the “trunk” of the isoprenoid biosynthetic pathway is in reality
a complex set of trunks that varies from organism to organism
rather than a set of linear chain elongation reactions. Members
of each family share distinctive highly conserved motifs
characteristic of proteins that have evolved from a common
ancestor.2,3 Members of the E-family typically synthesize shorter
chain isoprenoid diphosphates found early in the pathway, while
members of the Z-family synthesize longer chain diphosphates.1
Farnesyl diphosphate (FPP) synthase is the prototypal rep-
resentative of enzymes in the E-family. FPP synthase catalyzes
two reactions: the sequential addition of the hydrocarbon
moieties of dimethylallyl diphosphate (C5) and geranyl diphos-
phate (C10) to IPP (C5) to give FPP (C15). Metabolites derived
from FPP are apparently required by all living organisms, and
FPP synthase activity appears to be ubiquitous. Because of its
central role in isoprenoid metabolism, FPP synthase has served
as the platform for studying the structure2,4,5 and mechanism6,7
of E-family enzymes. Structural and genetic studies indicate
that the all R-helical “isoprenoid” fold of FPP synthase2 is found
in all of the E-family chain elongation enzymes4 and in more
distant relatives that catalyze cyclopropanation reactions in the
sterol8 and carotenoid9 branches of the pathway and terpenoid
cyclization reactions.10 An ancestral FPP synthase or a closely
related relative is probably the protein from which the other
enzymes evolved.
The stereochemistry of the two reactions catalyzed by FPP
synthase was determined by Cornforth and Popjak as part of
their classical work on cholesterol biosynthesis.11 They deter-
mined the stereochemistry of four distinct events during the
chain elongation reaction (see Scheme 1): (a) the new bond
between C1 of the allylic substrate and C4 of IPP is formed
with inversion at C1 of the allylic substrate, (b) C1 of the allylic
substrate adds to the si-face of the double bond in IPP, (c) a
new E-double bond is formed between C2 and C3 in the product,
(4) Tarshis, L. C.; Proteau, P. J.; Kellogg, B. A.; Sacchettini, J. C.; Poulter, C.
D. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 15018-15023.
(5) Hosfield, D. J.; Zhang, Y.; Dougan, D. R.; Broun, A.; Tari, L. W.; Swanson,
R. V.; Finn, J. J. Biol. Chem. 2004, 279, 8526-8529.
(6) Poulter, C. D.; Rilling, H. C. Acc. Chem. Res. 1978, 11, 307-313.
(7) Poulter, C. D.; Wiggins, P. L.; Le, A. T. J. Am. Chem. Soc. 1981, 103,
3926-3927.
(8) Pandit, J.; Danley, D. E.; Schulte, G. K.; Mazzalupo, S.; Pauly, T. A.;
Hayward, C. M.; Hamanaka, E. S.; Thompson, J. F.; Harwood, H. J. J.
Biol. Chem. 2000, 275, 30610-30617.
(1) Poulter, C. D. Phytochem. ReV. 2006, 5, 17-26.
(2) Tarshis, L. C.; Yan, M.; Poulter, C. D.; Sacchettini, J. C. Biochemistry
1994, 33, 10871-10877.
(3) Fujuhashi, M.; Zhang, Y.-W.; Higuchi, Y.; Li, X.-Y.; Koyama, T.; Miki,
K. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 4337-4342.
(9) Blagg, B. S. J.; Jarstfer, M. B.; Rogers, D. H.; Poulter, C. D. J. Am. Chem.
Soc. 2002, 124, 8846-8853.
(10) Christianson, D. W. Chem. ReV. 2006, 106, 3412-3442.
(11) Popjak, G.; Cornforth, J. W. Biochem. J. 1966, 101, 553-568.
9
10.1021/ja065573b CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 15819-15823
15819

A R T I C L E S
Thulasiram and Poulter
Scheme 1. Stereochemistry of Chain Elongation by FPP
Synthase
The volume of the extract was reduced to ∼50 µL with a stream of
dry nitrogen. A 1 µL portion of the extract was analyzed by gas
chromatography (GC) on a column 30 m × 0.25 mm (bore size) ×
0.25 µm (film thickness) DB-5 capillary column (J & W Scientific)
with a temperature gradient from 60 to 120 °C at 2 °C per min followed
by a temperature gradient of 120 to 230 °C at 10 °C per min with a He
with a flow rate of 1 mL/min. Samples were also analyzed by GC-
MS using the same conditions.
and (d) the pro-R proton is removed from C2 of IPP. Because
of the long evolutionary history of chain elongation and its
central role in isoprenoid biosynthesis, it has been assumed that
FPP synthase catalyzes the reaction with a high degree of
stereoselectivity. We now report that a significant amount of
Z-double bond isomers are formed during each of the chain
elongation steps and present evidence for the conformations of
IPP in the active site that lead to the double bond isomers.
In Vivo Studies. E. coli BL21(DE3)pLysS and E. coli JM101were
transformed with pET15B-His6-FDS and pKEN-His6-IDI, respectively.
LB containing ampicillin (100 µg/mL) was inoculated with the
transformed cells and incubated on a rotary shaker for 12 h at 37 °C.
The overnight culture was diluted 100-fold in the same medium and
incubated at 37 °C on a rotary shaker. At OD600 ≈ 0.6, cells were
induced with 1 mM IPTG and incubated at 30 °C for 4 h. The cells
were collected by centrifugation at 7000g, and the pellet was washed
three times with double distilled water. The cell pellet was suspended
in 35 mM HEPES buffer, pH 7.6 buffer (∼3 mL/g of cell paste), and
disrupted by sonication. The suspension was centrifuged at 10 000g
for 30 min, and the supernatant was centrifuged at 100 000g for an
hour. To a 5 mL portion of the 100 000g supernatant was added 1 mL
of 0.5 M glycine buffer, pH 10.5, containing 200 units of alkaline
phosphatase, 0.8 mL of 100 mM MgCl2, and 1.6 mL of 50 mM ZnCl2.
The mixture was incubated at 37 °C for 2 h and then extracted three
times with 1 mL of TBME. The combined organic extracts were
concentrated to ∼50 µL and analyzed by GC and GC-MS.
Experimental Section
Overproduction and Purification of Recombinant Enzymes.
Stock samples of E. coli strains containing the appropriate plasmid
(Table 1) were allowed to recover in Luria Bertani (LB) for 90 min
Table 1. Enzymes, Plasmids, and E. coli Host Strains
antibiotics
enzyme and source
plasmid
host strain
(
µ
g/mL)
ref
avian FPP synthase
Escherichia coli
FPP synthase
Artemisia tridentata
FPP synthase
pMJY189
XA-90
Amp (100)
Amp (100)
2
12
pET15BEFDS
pMPM3BFDS
pMPM3BCDS
BL-21
XA-90
XA-90
Results
Amp (100)
Amp (100)
Amp (100)
13
13
14
Analysis of C10 and C15 Products. The stereochemistry of
the C5 f C10 and C10 f C15 steps of chain elongation was
determined for seven E-selective enzymes (see Scheme 2). They
Artemisia tridentata
CPP synthase
Artemisia tridentata
pMPM3BBglCV XA-90
GPP synthase
Scheme 2. C5 f C10 and C10 f C15 Chain Elongation Reactions
(engineered)
Pyrococcus furiosus
FPP/GGPP synthase
Methanobacter
thermautotrophicus
FPP/GGPP synthase
pPet-pf1 102
pq-mth50
Rosetta (DE3) Amp (100),
15
16
pLysS
Cam(34)
Amp (100),
Tet (12.5),
Cam (34)
RosettaBlue
and plated on LB containing the appropriate antibiotic. Overnight
cultures grown from single colonies (24 h old) were used to inoculate
1 L of LB containing appropriate antibiotics. Cultures were grown at
37 °C and induced with isopropyl â-D-thiogalactoside (IPTG) to a final
concentration of 1 mM when the OD600 reached 0.6-0.8. After 4 h,
the cells were harvested by centrifugation at 7000g and stored at -80
°C until needed. Frozen cells were disrupted by sonication in 50 mM
phosphate buffer, pH 8.0, containing 300 mM NaCl and 10 mM
imidazole, and the resulting suspension was clarified by centrifugation.
Soluble His6-tagged proteins were purified by chromatography on Ni2+
-
NTA agarose (Qiagen, Chatsworth, CA), according to the manufac-
turer’s protocol. Fractions containing pure His6-tagged protein were
combined, and solid NH4SO4 (47 g/100 mL) was added. Precipitated
protein was collected by centrifugation and dissolved in a minimum
volume of 20 mM Tris-HCl buffer, pH 8.0, containing 20% glycerol,
and 4 mM DTT. The solubilized protein was dialyzed against the same
buffer, flash frozen, and stored at -80 °C until needed.
Product Studies. Enzyme (200 µg) was added to 35 mM HEPES
buffer, pH 7.6, containing 10 mM MgCl2, 5.0 mM â-mercaptoethanol,
200 µM allylic substrate, and 200 µM IPP to a final volume of 400
µL. The mixture was incubated for 2 h at 30 °C (60 °C for the
Pyrococcus furiosus and Methanobacterium thermautotrophicus gera-
nylgeranyl diphosphate (GGPP) synthases). After the incubation, 80
µL of 500 mM glycine buffer, pH 10.5, containing 5 mM ZnCl2 and
80 units of calf mucosa alkaline phosphatase (Sigma) was added,
followed by incubation at 37 °C for 1 h to hydrolyze the diphosphate
esters. Solid NaCl (∼0.2 g) was added, and the aqueous phase was
extracted three times with 400 µL of tert-butyl methyl ether (TBME).
include two eukaryotic FPP synthases (avian2 and sagebrush13),
a eubacterial FPP synthase (E. coli12), two archaeal bifunctional
FPP/GGPP synthases (P. furiosus15 and M. thermautotrophi-
cus16), and two proteins with GPP synthase activity (chrysan-
themyl diphosphate (CPP) synthase from Artemisia tridentata
speciformis13 and a chimeric enzyme (BglCV) constructed from
A. tridentata speciformis FPP synthase and CPP synthase14).
The chain elongation enzymes were incubated with equimolar
concentrations of IPP and allylic diphosphate followed by
treatment with alkaline phosphatase to hydrolyze the diphosphate
(12) Chen, M.; Poulter, C. D. Unpublished results.
(13) Hemmerlin, A.; Rivera, S. B.; Erickson, H.; Poulter, C. D. J. Biol. Chem.
2003, 34, 32132-32140.
(14) Erickson, H.; Poulter, C. D. J. Am. Chem. Soc. 2003, 23, 6885-6888.
(15) Kidd, M. T.; Poulter, C. D. Unpublished results.
(16) Chen, A.; Poulter, C. D. Arch. Biochem. Biophys. 1994, 314, 399-404.
9
15820 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Farnesyl Diphosphate Synthase
A R T I C L E S
E,E-FOH comprised 96-97% of the mixture of C15 alcohols
from bacterial and eukaryotic FPP synthases, and 87-88% for
the archaeal enzymes. The relative amounts of GOH and NOH
for enzymes that synthesized FPP did not accurately reflect the
stereoselectivity of the C5 f C10 step because GPP, the normal
allylic substrate for the C10 f C15 step, is consumed in pref-
erence to neryl diphosphate (NPP). In control experiments where
IPP was incubated with GPP and NPP, the rate for conversion
of NPP to E,Z-FPP was <5% of that for conversion of GPP to
E,E-FPP. Thus, analysis of the peak intensities for NOH and
GOH overestimates the amount of NPP generated during chain
elongation because GPP is consumed preferentially. A peak for
E,Z-FOH was not seen in incubations with DMAPP, presumably
because of the small amount of NPP formed in the first step of
chain elongation and the inefficiency of elongating NPP relative
to GPP in the second step. The stereochemistry of the C5 f
C10 step was determined for BglCV, an engineered version of
sagebrush FPP synthase that did not catalyze the second step
of chain elongation, and CPP synthase, an enzyme with C5 f
C10 chain elongation activity. The E/Z stereoselectivity measured
for BglCV was similar to those seen for FPP synthases. A
substantially higher proportion of NPP, the Z-C10 stereoisomer,
was seen for the chain elongation catalyzed by CPP synthase.
Stereochemistry at C2 of IPP. The stereochemistry for
proton elimination at C2 of IPP during formation of both double
bond stereoisomers was determined for the C5 f C10 step for
BglCV and the C5 f C10 and C10 f C15 steps for avian, E.
coli, and A. tridentata FPP synthases during incubations with
(R)-[2-2H]IPP and (S)-[2-2H]IPP. The results of GC-MS
analysis of the chain elongation alcohols following hydrolysis
of the diphosphate esters are shown in Table 3 (mass spectra
are shown in the Supporting Information). When unlabeled IPP
and DMAPP were used as substrate, the EI MS for GOH and
NOH gave peaks at m/z 154 (M+), 139 (M+ - CH3), 136 (M+
- H2O), 123 (M+ - CH2OH), and 121 (M+ - H2O - CH3).
Similarly, the EI MS of Z,E- and E,E-FOH had peaks at m/z
222 (M+), 207 (M+ - CH3), and 204 (M+ - H2O). All of the
alcohols gave a m/z 69 base peak for loss of the ω-isoprene
unit. For each of the enzymes, deuterium was lost from C2 of
(R)-[2-2H]IPP and retained at C2 of (S)-[2-2H]IPP during the
C5 f C10 step and the C10 f C15 step, regardless of which
double bond isomer was produced. When DMAPP was the
allylic substrate, incubation with (R)-[2-2H]IPP gave C10 chain
elongation alcohols with m/z at 154, 139, 136, 123, and 121
(BglCV) and C15 alcohols with m/z at 222, 207, and 204 (avian,
E. coli, A. tridentata). Similar incubations with (S)-[2-2H]IPP
gave peaks at m/z 155, 140, 137, 124, and 122 for the C10
alcohols and m/z 224, 209, and 206 (addition of two IPP units)
Figure 1. GC analysis of C10 and C15 alcohols after hydrolysis of chain
elongation products from avian FPP synthase. (a and b) Authentic samples
of NOH (Z-C10), GOH (Z-C10), Z,E-FOH, and Z,E-FOH. (c and d)
Incubation with IPP and DMAPP. (e and f) Incubation with IPP and GPP.
Minor peaks at ∼39.0 and 39.4 min are contaminants seen in control
incubations without substrate. Conditions are in the Experimental Section.
Table 2. C10 and C15 Chain Elongation Products
a
b
C10
C15
allylic
substrate
enzyme
Zc
Ec
Z,Ec
E,Ec
DMAPP
avian FPP synthase
E. coli FPP synthase
A. tridentata FPP synthase
P. furiosus FPP/GGPP synthase
M. thermautotrophicus
FPP/GGPP synthase
3
4
3
12
13
97
96
97
88
87
A. tridentata CPP synthase
BglCV GPP synthase
8
4
92
96
GPP
avian FPP synthase
E. coli FPP synthase
A. tridentata FPP synthase
P. furiosus FPP/GGPP synthase
3
3
5
97
97
95
86
14
a C10 alcohols were 45-73% and C15 alcohols were 35-50% of the total.
b C15 alcohols were 87-94% of the C10 + C15 total. c <(0.5% for three
independent determinations.
esters to their corresponding alcohols. The stereochemistries of
the newly formed double bonds in the C10 and C15 alcohols
were determined by GC and GC-MS analysis with authentic
samples for comparison. A typical set of GC traces is shown in
Figure 1, and the results are summarized in Table 2. In each
instance, the expected E-isomer was accompanied by a small,
but easily detectable amount of the Z-isomer.
for the C15 alcohols. Related behavior was seen for the C10
C
f
15 step when GPP was the allylic substrate where (R)-[2-2H]-
IPP gave C15 alcohols whose mass spectra had peaks at m/z
222, 207, and 204 (avian, E. coli, A. tridentata) and (S)-[2-2H]-
IPP gave alcohols with peaks at m/z 223, 208, and 205 (addition
of one IPP unit).
Double Bond Stereochemistry in Vivo. E. coli strain BL21-
(DE3)pLysS, transformed with a plasmid that directs expression
of E. coli FPP synthase under control of the T7 promoter, was
incubated as described in the Experimental Section. The cell
free extract was treated with alkaline phosphatase, the resulting
alcohols were extracted with MTBE, and extract was analyzed
For incubations with DMAPP as the starting substrate, the
mixture of alcohols, following treatment with phosphatase,
included dimethylallyl alcohol from unreacted DMAPP, geraniol
(GOH), nerol (NOH), and the E,E- and Z,E-isomers of farnesol
(FOH). The stereoselectivity for the C10 f C15 step was deter-
mined from the relative amounts of E,E-FOH and Z,E-FOH.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15821

A R T I C L E S
Thulasiram and Poulter
E,Z-FPP
Table 3. Molecular Ions (M+) for the Chain Elongation Alcohols from Incubations with (R)- and (S)-[2-2H]IPP
GPP
NPP
E,E-FPP
enzyme
allylic substrate
R
S
R
S
R
S
R
S
avian
DMAPP
GPP
DMAPP
GPP
DMAPP
GPP
DMAPP
154
155
154
155
222
222
222
222
222
222
224
223
224
223
224
223
222
222
222
222
222
222
224
223
224
223
224
223
E. coli
154
154
154
155
155
155
154
154
154
155
155
155
A. tridentata
BglCV
glyceraldehyde phosphate and pyruvate by the methylerythrose
phosphate pathway.24 One might have expected that FPP
synthase, with a history that apparently stretches back to the
very beginning of cellular life, would have become highly
selective for E-chain elongation during the course of its
evolution. Our results with eukaryotic, bacterial, and archaeal
forms of the enzyme indicate that this is not the case.
A biosynthetic pathway that uses IPP and DMAPP in the
first step of chain elongation presents special challenges to FPP
synthase. For maximal efficiency, the regions of the active site
that bind IPP and DMAPP must distinguish between the two
C5 isomers. Although IPP and DMAPP can both bind to the
allylic site,25,26 this mode of IPP binding does not appear to
present a problem. At physiological concentrations of IPP and
DMAPP, substrate inhibition due to IPP binding in the allylic
site does not appear be important.25 In addition, the homoallylic
moiety in IPP cannot participate in chain elongation by the well-
established stepwise electrophilic alkylation mechanism shown
in Scheme 3.6,7 In contrast, the double bond in DMAPP is
Figure 2. GC trace of MTBE extracts of cell-free homogenate from E.
coli strain BL21(DE3)pLysS/pET15BEFDS after treatment with alkaline
phosphatase.
by GC-MS. The GC profile from single ion monitoring at m/z
69, the mass of base peak for E,E- and Z,E-FOH, is shown in
Figure 2. The peaks at 38.4 and 38.8 min gave mass spectra
that were virtually identical to those of Z,E-FOH and E,E-
farnesol, respectively. Other peaks gave mass spectra that were
not characteristic of mono- or sesquiterpenoid alcohols. Thus,
the minor stereoisomer of FPP is synthesized in vivo by E. coli
from plasmid encoded E. coli FPP synthase in a ratio similar
to that measured in vitro for the same enzyme.
Scheme 3. Mechanism for Chain Elongation
Discussion
Isoprenoid chain elongation is generally thought to proceed
with a high degree of stereoselectivity characteristic of most
enzyme-catalyzed reactions. The stereochemistry of the double
bonds, particularly the C2-C3 double bond, in the growing
hydrocarbon chain of the allylic substrate is important if the
intermediate is to serve as a substrate for the next addition of
IPP.1 E,E-FPP, which lies at an important multiple branch point
in the isoprenoid pathway, is the precursor for biosynthesis of
sterols,17 farnesylated proteins,18 and sesquiterpenes.10 FPP is
also the allylic primer for the chain elongation enzymes required
to synthesize geranylgeranylated proteins,19 dolichols,20 heme
a,21 and ubiquinones.22 All of these reactions require the E,E-
stereoisomer. IPP and DMAPP are “expensive” substrates for
FPP synthase. IPP is synthesized in six steps, and DMAPP in
seven steps, from three molecules acetyl-CoA by the mevalonate
pathway.23 Both molecules are synthesized in six steps from
susceptible to electrophilic alkylation. FPP synthase apparently
has the ability to exclude that molecule from the IPP binding
region to prevent a reaction with itself.14,26 The X-ray structure
of the E. coli enzyme complexed with IPP and an unreactive
thio analogue of DMAPP5 provides an explanation for the
discrimination against DMAPP at the IPP site based on
differences in the conformational flexibility of the two substrates.
The orientation of IPP and the thioDMAPP analogue in the E.
coli FPP synthase‚IPP‚thioDMAPP complex (see Figure 3)
would produce the E-stereochemistry during chain elongation
(21) Weinstein, J. D.; Branchaud, R.; Beele, S. I.; Bemet, W. J.; Sinclair, P. R.
Arch. Biochem. Biophys. 1986, 245, 44-50.
(17) Poulter, C. D.; Rilling, H. C. Conversion of farnesyl pyrophosphate to
squalene. In Biosynthesis of Isoprenoid Compounds; Porter, J. W., Ed.;
John Wiley & Sons: New York, 1981; Vol. 1, Chapter 8, pp 414-441.
(18) Mayer, M. P.; Prestwich, G. D.; Dolence, J. M.; Bond, P. D.; Wu, H.-y.;
Poulter, C. D. Gene 1993, 132, 41-47.
(22) Ashby, M. N.; Edwards, P. A. J. Biol. Chem. 1990, 265, 13157-13164.
(23) Qureshi, N.; Porter, J. W. Conversion of acetyl-coenzyme A to isopentenyl
pyrophosphate. In Biosynthesis of Isoprenoid Compounds; Porter, J. W.,
Spurgeon, S. L., Eds.; John Wiley and Sons: New York, 1981; Vol. 1, pp
47-94.
(19) Jiang, Y.; Proteau, P.; Poulter, C. D.; Ferro-Novick, S. J. Biol. Chem. 1995,
270, 21793-21799.
(24) Rohdich, F.; Kis, K.; Bacher, A.; Eisenreich, W. Curr. Opin. Chem. Biol.
2001, 5, 535-540.
(20) Matsuoka, S.; Sagami, H.; Kurisaki, A.; Ogura, K. J. Biol. Chem. 1991,
266, 3464-3468.
(25) Laskovics, F. M.; Poulter, C. D. Biochemistry 1981, 20, 1893-1901.
(26) King, H. L.; Rilling, H. C. Biochemistry 1977, 16, 3815-3819.
9
15822 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Farnesyl Diphosphate Synthase
A R T I C L E S
Figure 3. Orientation of IPP and DMAPP in the active site of E. coli FPP
synthase, highlighting stereochemical features important for E-selective chain
elongation.
Figure 4. Space-filling representations of IPP and DMAPP. (a and b) Both
molecules in conformations that allow maximum spatial overlap, viewed
from the top face of isoprenoid units. (c and d) Edge-on view of the
conformers shown in parts a and b. (e) Same orientation as part c except
the C1-C2/C3-C4 dihedral angle is 118°. (f) Same orientation as part c
except the C1-C2/C3-C4 dihedral angle is 62°.
by FPP synthase. The hydrocarbon unit of DMAPP is located
on the si-face of the C3-C4 double bond in IPP with C1
positioned for an inversion of configuration upon alkylation.
The C1-C2/C3-C4 dihedral angle in IPP, Θ, is 118°. In this
orientation, a “least motion” elimination of a proton from C2
would give an E-double bond in the allylic product. In addition,
the HR proton at C2 in IPP is suitably positioned for elimination,
presumably assisted by a nonbridging oxygen atom in the nearby
molecule of inorganic pyrophosphate.1 DMAPP has a C2-C3
double bond, and as a consequence the C1-C2/C3-C4 dihedral
angle Θ is restricted to 180°. The inability of FPP synthase to
bind DMAPP at the IPP site can be explained by conformational
differences between the two substrates, whereas IPP cannot be
readily excluded from the DMAPP site because IPP is a more
flexible molecule that can adopt conformations whose topologies
are very similar to any of those accessible to DMAPP. This is
illustrated in Figure 4 using space-filling representations of IPP
and DMAPP. The shape of the conformer of IPP where the
five carbon atoms are in the same plane (a and c) is very similar
to that of DMAPP (b and d). In contrast, rotation about the
C2-C3 bond in IPP, which is not possible for DMAPP, gives
a substantially different shape (d and e).
Why is the degree of stereocontrol exerted by FPP synthase
during chain elongation not higher? The observation that the
HR in IPP is removed when either double bond isomer is formed
provides a clue. Presumably the electrophilic alkylation of IPP
proceeds with the initial formation of a π-complex between C1
of the allylic cation and the double bond, followed by collapse
to the C3 tertiary cation.27-29 The relative positions of C1 in
the analogue and the C3-C4 IPP double bond in the FPP
synthase‚IPP‚thioDMAPP complex are consistent with this
scenario. A 180° rotation about the C2-C3 bond in IPP gives
a conformer, Θ ) 62°, that is topologically similar to the Θ )
118° conformer (see Figure 4e and f) but which would produce
a Z-double bond in the chain elongation product. It might be
difficult to distinguish between the Θ ) 62° and Θ ) 118°
conformers of IPP during binding because of their similar
topologies. Thus, a relatively small movement of the dimethyl-
allyl cation during alkylation would suffice to alkylate either
conformer of IPP with subsequent elimination of HR, in
agreement with our observations.
Bacterial and eukaryotic FPP synthases “waste” approxi-
mately 4% of the available IPP and 2% of the DMAPP during
synthesis of FPP. These numbers increase substantially to 30%
and 15%, respectively, for the thermophilic archaeal enzymes.
With the possible exception of plants, the Z-C10 and Z,E-C15
double bond isomers do not appear to be utilized in subsequent
biosynthetic reactions and are presumably degraded. While this
amount of waste is burdensome for the synthesis of FPP, it is
unsustainable for the longer chain isoprenoid diphosphates
required for biosynthesis of ubiquinones or dolicohols. Given
the ancient origins of FPP synthase, it seems doubtful that the
enzyme can achieve a greater degree of discrimination for differ-
ent conformers of IPP through continued evolution. In this re-
gard, it is interesting to note that the long chain polyprenyl
diphosphate synthases use FPP or GGPP instead of DMAPP as
the allylic substrate to initiate elongation. Perhaps greater stereo-
control is possible in the long chain enzymes through better
immobilization of the carbocationic intermediate as a result of
binding interactions between the enzyme and the longer
isoprenoid chain.
Acknowledgment. We thank Professor John Vederas for
generous samples of (R)- and (S)-[2-2H]IPP and Ms. Mo Chen
for providing the plasmid for E. coli FDS. This work was
supported by National Institutes of Health grant GM 21328.
Supporting Information Available: General materials and
methods, synthesis of (E,E)- and (Z,E)-farnesol, and mass
spectra of (E,E)- and (Z,E)-farnesol (unlabeled, monodeuterated,
and dideuterated), geraniol (unlabeled and monodeuterated), and
nerol (unlabeled and monodeuterated). This material is available
free of charge via the Internet at http://pubs.acs.org.
JA065573B
(27) Dewar, M. J. S.; Reynolds, C. H. J. Am. Chem. Soc. 1984, 106, 1744-
1750.
(28) Farcasiu, D.; Norton, S. H. J. Org. Chem. 1997, 62, 5374-5379.
(29) Farcasiu, D.; Norton, S. H.; Hancu, D. J. Am. Chem. Soc. 2000, 122, 668-
676.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15823