Báo cáo khoa học: The product chain length determination mechanism of type II geranylgeranyl diphosphate synthase requires subunit interaction pptx

The product chain length determination mechanism of
type II geranylgeranyl diphosphate synthase requires
subunit interaction
Motoyoshi Noike
1,2
, Takashi Katagiri
1
, Toru Nakayama
1
, Tanetoshi Koyama
2
, Tokuzo Nishino
1
and
Hisashi Hemmi
1
1 Department of Biochemistry and Engineering, Graduate School of Engineering, Tohoku University, Miyagi, Japan
2 Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Miyagi, Japan
(All-E) prenyl diphosphate synthase catalyzes the con-
secutive condensation of isopentenyl diphosphates
(IPP) with allylic prenyl diphosphates to yield the final
product with a specific prenyl chain length [1,2]. The
chain length of the product must be tightly controlled
because polymerization of isoprene units is the key
reaction responsible for the tremendous variety of
naturally occurring isoprenoid compounds (> 50 000)
[3]. For example, many important compounds, such as
carotenoids, tocopherols, diterpenes, chrolophyll and
archaeal membrane lipids, are synthesized from gera-
nylgeranyl diphosphate (GGPP; C
20

). On the other
Keywords
farnesyl diphosphate synthase;
geranylgeranyl diphosphate synthase;
isoprenoid; mutagenesis; prenyltransferase
Correspondence
H. Hemmi, Department of Applied
Molecular Bioscience, Graduate School of
Bioagricultural Sciences, Nagoya University,
Furo-cho, Chikusa-ku, Nagoya, Aichi
464-8601, Japan
Fax: +81 52 789 4120
Tel: +81 52 789 4134
E-mail:
(Received 21 February 2008, revised 2 June
2008, accepted 4 June 2008)
doi:10.1111/j.1742-4658.2008.06538.x
The product chain length determination mechanism of type II geranyl-
geranyl diphosphate synthase from the bacterium, Pantoea ananatis, was
studied. In most types of short-chain (all-E) prenyl diphosphate synthases,
bulky amino acids at the fourth and/or fifth positions upstream from the
first aspartate-rich motif play a primary role in the product determination
mechanism. However, type II geranylgeranyl diphosphate synthase lacks
such bulky amino acids at these positions. The second position upstream
from the G(Q/E) motif has recently been shown to participate in the
mechaism of chain length determination in type III geranylgeranyl diphos-
phate synthase. Amino acid substitutions adjacent to the residues upstream
from the first aspartate-rich motif and from the G(Q/E) motif did not
affect the chain length of the final product. Two amino acid insertion in
the first aspartate-rich motif, which is typically found in bacterial enzymes,

is thought to be involved in the product determination mechanism. How-
ever, deletion mutation of the insertion had no effect on product chain
length. Thus, based on the structures of homologous enzymes, a new line
of mutants was constructed in which bulky amino acids in the a-helix
located at the expected subunit interface were replaced with alanine. Two
mutants gave products with longer chain lengths, suggesting that type II
geranylgeranyl diphosphate synthase utilizes an unexpected mechanism of
chain length determination, which requires subunit interaction in the
homooligomeric enzyme. This possibility is strongly supported by the
recently determined crystal structure of plant type II geranylgeranyl
diphosphate synthase.
Abbreviations
DMAPP, dimethylallyl diphosphate; FARM, first aspartate-rich motif; FPP, farnesyl diphosphate; FPS, farnesyl diphosphate synthase; GGPP,
geranylgeranyl diphosphate; GGPS, geranylgeranyl diphosphate synthase; GPP, geranyl diphosphate; GPS, geranyl diphosphate synthase;
IPP, isopentenyl diphosphate; SARM, second aspartate-rich motif.
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3921
hand, farnesyl diphosphate (FPP; C
15
) is the precursor
of steroids, sesquiterpenes and heme a. Moreover, FPP
is the usual allyic primer substrate for prenyl elonga-
tion reactions, which yield longer-chain prenyl diphos-
phates as the precursors of respiratory quinones,
dolichol and natural rubber, although some organisms
also use GGPP for the same purpose. Longer-chain
(all-E) prenyl diphosphates (up to C
60
) are utilized for
the biosynthesis of various respiratory quinones, which
have been used to classify the microorganisms. GGPP

and FPP are also utilized for protein modification,
although they modify very different classes of acceptor
proteins. Rab family G proteins are geranylgeranyl-
ated, whereas farnesylation typically occurs on Ras
proteins. In addition, geranyl diphosphate (GPP; C
10
)
is the precursor of volatile monoterpenes and also is
used to modify secondary metabolites.
The mechanism of prenyl-chain elongation, and
therefore of product determination in (all-E) prenyl
diphosphate synthases, which share many conserved
sequences in spite of their different reaction products,
has been investigated previously. The enzymes are con-
structed mainly of a-helices, which form a large reac-
tion cavity per a subunit [4–12]. Most of the enzymes
are homodimeric proteins, although some enzymes
consist of heterodimers with little homology between
the subunits [1]. A few mammalian enzymes are known
to form oligomers [13,14]. The highly conserved motifs
of the enzymes [i.e. the first aspartate-rich motif
(FARM) and the second aspartate-rich motif (SARM)]
are thought to bind the diphosphate group of the allylic
substrate via magnesium ions. FARM and SARM are
located on a-helices D and H. (Note that the present
study follows the helix designation first reported for the
crystal structure of avian farnesyl diphosphate synthase
(FPS) by Tarsis et al. [4].) Departure of the diphos-
phate group forms an allylic carbocation, which is
attacked by the p-electron at the double-bond of IPP,

forming a new bond between the fourth carbon of IPP
and the first carbon of the allylic substrate. Thus, pre-
nyl diphosphate is elongated by one C
5
prenyl unit.
The condensation reaction is repeated, elongating the
prenyl chain. As the chain elongates, the hydrocarbon
moiety becomes located deep within the reaction cavity
formed by a-helices C, D, E, F, G and H. Enzyme-
specific termination of prenyl-chain elongation results
in final products unique to each enzyme.
Mutational and structural studies have revealed that,
in general, bulky amino acids at the bottom of the cav-
ity block prenyl-elongation. In particular, our research
group has shown that, in (all-E) prenyl diphosphate
synthases yielding short-chain products such as GGPP
and FPP, the bulky amino acids are found in two
regions: upstream from FARM [15] and from the
highly-conserved G(Q/E) motif [16], respectively.
FARM exists on a-helix D and the G(Q/E) motif is
located on a-helix F. Based on the characteristic
sequences upstream from FARM, the short-chain
enzymes have been classified into five types [15]: three
types of geranylgeranyl diphosphate synthase (GGPS)
and two types of FPS (Fig. 1). Type I GGPS from
archaea has a bulky aromatic amino acid residue,
which plays a primary role in the chain length determi-
nation mechanism at the fifth position upstream from
FARM. The importance of the residue was shown by
mutational studies on GGPS from a thermoacidophilic

archaeon Sulfolobus acidocaldarius [17–19]. Two
GGPSs with known crystal structures [i.e. those from
a hyperthermophilic archaeon Pyrococcus horikoshii
(Protein Data Bank code 1WY0) and from a thermo-
philic bacterium Thermus thermophilus (1WMW)] also
fall into this type. The bulky residue at the fifth posi-
tion upstream from FARM is in the center of the
cavity and likely to act as the bottom of it in these
structures; however, the structural information is inde-
cisive with respect to the role of the residue because
the structures do not contain allylic substrates or their
analogues bound in the active site. Such characteristic
Fig. 1. Alignment of amino acid sequences around FARM of vari-
ous (all-E) prenyl diphosphate synthases. The partial amino acid
sequences of the enzymes classified into two types of FPSs synth-
ases and three types of GGPSs are aligned. Sce FPS, S. cerevisiae
FPS; Gga FPS, Gallus gallus (avian) FPS; Eco FPS, E. coli FPS; Gst
FPS, G. stearothermophilus FPS; Sac GGPS, S. acidocaldarius
GGPS; Mth GGPS, Methanobacterium thermoautotrophicum GGPS;
Pan GGPS, P. ananatis GGPS; Sal GGPS, S. alba (mustard) GGPS;
SceGGPS, S. cerevisiae GGPS; Hsa GGPS, Homo sapiens GGPS.
The characteristic amino acid residues suggested to be involved
product determination for each type of enzyme are shaded.
Product determination mechanism of type II GGPS M. Noike et al.
3922 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
sequences also are evident in FPSs. Eukaryotic type I
FPS has bulky amino acids at both the fourth and fifth
positions upstream from FARM. Mutational and
structural studies on avian FPS clarified the role of the
positions [20]. In the structure of mutated avian FPS

(1UBX), in which phenylalanine residues at the fourth
and fifth positions upstream from FARM are replaced
with serine and alanine, respectively, the x-end of the
hydrocarbon chain of FPP bound in the cavity passes
through the hole formed by the mutagenesis. Broad
structural studies using inhibitor substrate analogues
were made with type 1 FPSs from human [11] and pro-
tozoa [6,12], and some of the structures [e.g. FPS from
human (2F94 and 3B7L) and Trypanosoma brucei
(2P1C and 2I19) binding bisphosphonate inhibitors
with hydrocarbon chains as long as that of GPP] sug-
gest the importance of the bulky amino acids because
they are in contact with the inhibitors. However, no
mutational study that supports the hypothesis has been
made with the enzymes. Bacterial type II FPS has a
bulky amino acid only at the fifth upstream position,
but also has a two amino acid insertion in FARM.
Mutational studies of FPS from Geobacillus stearother-
mophilus showed that the bulky amino acid upstream
from FARM is involved in the chain length determina-
tion mechanism [21,22]. The crystal structure of this
type of FPS was elucidated using the enzymes from
several bacteria, such as Staphylococcus aureus and
Escherichia coli [7]. The structures of E. coli FPS bind-
ing substrate analogues (1RQI and 1RQJ) also suggest,
but do not ensure, the role of tyrosine at the fifth posi-
tion upstream from FARM, which is still distant from
the analogues with short hydrocarbon chains used in
that study [7]. By contrast, eukaryotic type III GGPS,
which lacks bulky amino acids at the fourth or fifth

positions upstream from FARM, was shown to utilize
bulky amino acids at the second position upstream
from the G(Q/E) motif to terminate chain elongation
by our mutational work using GGPS from Saccharo-
myces cerevisiae [16]. This information was later
supported by a structural and mutational study on the
same enzyme (2DH4) [9].
In the present study, mutational studies of GGPS
from a bacterial plant pathogen, P. ananatis, were
performed to investigate the mechanism of chain
length determination in type II GGPS from bacteria
and plants, which has not been identified to date.
This type of GGPS lacks bulky aromatic amino acids
at the fourth and fifth positions upstream from
FARM, similar to type III GGPS, whereas it has a
two amino acid insertion in FARM, as does type II
FPS. Unexpectedly, mutations at the fourth and fifth
positions upstream from FARM and at the second
position upstream from the G(Q/E) motif did not
affect the chain length of the final product. In addi-
tion, deletion of the insertion sequence in FARM,
which is thought to be involved in the chain length
determination mechanism [18], also had no effect on
the chain length of the final product. An additional
mutational study with type II FPS from G. stearother-
mophilus confirmed that the insertion in FARM does
not play a role in the mechanism of chain length
determination in type II enzymes. These results sug-
gest that chain length determination is controlled by
another region of the enzymes. Thus, a new line of

mutants was created based on the crystal structures
of other short-chain enzymes and on the results from
previous mutational studies. Accordingly, a-helix E,
which would be located at the subunit interface of
the enzyme, was identified as playing a role in the
product chain length determination mechanism of
type II GGPS. Moreover, this result suggests that the
other subunit of the homooligomeric enzyme is
involved in the product chain length determination
mechanism. This conclusion is supported by the
recently-solved crystal structure of type II GGPS
from mustard [23]. The mechanism of product chain
length determination of type II GGPS identified in
the present study may also explain the participation
of noncatalytic subunits in the product determination
mechanisms of some heteromeric enzymes, such as
geranyl diphosphate synthase (GPS) and longer-chain
prenyl diphosphate synthases.
Results
Refolding and purification of recombinant
P. ananatis GGPS
P. ananatis GGPS and the mutant enzymes were
expressed in E. coli as inclusion bodies. To obtain sol-
uble enzymes, inclusion bodies prepared from the
insoluble fraction were denatured by guanidine hydro-
chloride and then purified by refolding on a HisTrap
column. The purified proteins gave almost single, iden-
tical bands by SDS/PAGE (data not shown). Only the
mutant L128A was completely inactive. All other
mutant GGPSs exhibited enzyme activity comparable

to that of the wild-type enzyme, whereas L127A
showed only approximately 20% activity of wild-type.
Analysis of the quaternary structure of the refolded
enzyme using blue native PAGE showed that the
molecular mass of P. ananatis GGPS is approximately
130 or 240 kDa, suggesting that the main part of the
enzyme exists as a homotetramer or a homooctamer
(Fig. 2).
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3923
Mutation in the region upstream from FARM
In type II GGPS, bulky aromatic amino acids at the
fourth and/or fifth positions upstream from FARM,
which are characteristic of many short-type (all-E) pre-
nyl diphosphate synthases, are not present. Previous
studies, mainly conducted by our research group, indi-
cated that bulky amino acids on a-helix D, which
includes FARM, block prenyl-chain elongation,
thereby controlling chain length [17–20,22]. To deter-
mine whether this mechanism of chain length deter-
mination also operates in type II GGPS from
P. ananatis, alanine 89 at the fifth position upstream
from FARM was replaced with bulky amino acids
(Fig. 3A). These mutations were designed to mimic
bacterial type II FPS. The mutants, A89F, A89L and
A89H, yielded shorter products than the wild-type
enzyme (Fig. 3B). In particular, the substrate specifici-
ties of A89F and A89H were almost identical to that
of FPS: product yield was minimal when GGPP was
used as the substrate. This result suggested that, in

type II GGPS, the prenyl-chain of the product elon-
gates along a-helix D and that the amino acid residues
on a-helix D further upstream from FARM are
involved in chain length determination. Thus, new
mutations were introduced further upstream from
FARM. It was expected that substituting the smaller
amino acid, alanine, for the bulky residues would
increase the chain length of the final products
(Fig. 4A). However, these mutants (i.e., H87A, V86A
and M85A) did not yield longer products than the
wild-type (Fig. 4B). H87A activity using GGPP as the
substrate was undetectable, probably because the
mutation significantly decreased overall enzyme activ-
ity. These results clearly indicated that bulky amino
acids relative to FARM do not contribute to the prod-
uct determination mechanism of type II GGPS.
Mutation at the second position upstream from
the G(Q/E) motif
Because the mechanism of chain length determination
for type II GGPS was shown to be independent from
the region upstream from FARM, the other region
known to play a role in chain length determination
was expected to play a critical role. The second posi-
tion upstream from the conserved G(Q/E) motif was
first identified as an important residue in the chain
length determination mechanism of type III GGPS
from S. cerevisiae in a previous study conducted in our
laboratory [16]. The relatively bulky residue, histidine
139, rather than those upstream from FARM, was
found to block chain-elongation. The role of the resi-

due was later supported by Chang et al. [9]: the crystal
structure of S. cerevisiae GGPS that these authors
determined demonstrated that histidine 139 forms the
bottom of the reaction cavity. Therefore, mutants of
type II GGPS from P. ananatis, in which the residue
at the second position upstream from the G(Q/E)
Fig. 2. Blue native PAGE of refolded P. ananatis GGPS. The refold-
ing procedure is described in the Experimental procedures. Lane 1,
molecular mass standard; lane 2, wild-type; lane 3, I121A; lane 4,
V125A.
A
B
Fig. 3. Introduction of substitutive mutations into the fifth position
upstream from FARM of P. ananatis GGPS. (A) Partial amino acid
sequences around FARM of wild-type and mutated enzymes are
aligned. The substituted amino acid residues are shaded. (B) TLC
autoradiochromatograms of the reaction products of wild-type and
mutated enzymes. The products were analyzed as described in the
Experimental procedures. The allylic substrate used is indicated at
the top of each autoradiochromatogram. Lane 1, A89F; lane 2,
A89L; lane 3, A89H; lane 4, wild-type. Under all assay conditions,
< 30% of each substrate reacted. Ori., origin; S.F., solvent front.
Product determination mechanism of type II GGPS M. Noike et al.
3924 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
motif was replaced with smaller amino acids, were con-
structed (Fig. 5A). However, the mutants, V163A,
V163G and V163S, did not yield products with a chain
length longer than those produced by the wild-type
enzyme (Fig. 5B). Moreover, some mutants did not
give the C

25
side-product and exhibited decreased spec-
ificity for GGPP. This unexpected result indicated that
the second position upstream from the G(Q/E) motif
does not contribute to the mechanism of chain length
determination in type II GGPS. In addition, substi-
tution of V163 with a bulky amino acid, phenylala-
nine, resulted in loss of activity (data not shown).
Deletion of the insertion sequence in FARM
Ohnuma et al. [18] performed a detailed investigation
of the mechanism of chain length determination in
short-chain (all-E) prenyl diphosphate synthase, mainly
using type I GGPS from S. acidocaldarius to construct
various mutants. In their study, two amino acids were
inserted in FARM of type I GGPS to mimic type II
FPS because this two amino acid insertion, which is
specifically observed in type II FPS and type II GGPS,
was expected to affect the product chain length. Type
I GGPS with the insertion mutation yielded larger
amounts of the reaction intermediate, FPP, whereas
GGPP remained the final product. Although Ohnuma
et al. [18] did not confirm the effect of the insertion by
performing the converse mutation (i.e. deletion of the
insertion from type I FPS or type II GGPS), the inser-
tion sequence was thought to play a role in the mecha-
nism of chain length determination in type II GGPS.
Thus, in the present study, the two amino acids inser-
tion was deleted from FARM of type II GGPS from
P. ananatis to confirm the effect of the deletion on
product chain length (Fig. 6A). However, the mutant,

GGPS-DFARM, did not yield a final product with a
chain length longer than that of the product resulting
from the wild-type enzyme, which gave a small amount
of the C
25
side-product (Fig. 6C). The mutant enzyme
appeared to exhibit reduced activity toward GGPP,
although this reduction may have been due to a
decrease in overall enzyme activity. This result indi-
cates that the insertion does not have a significant
effect on the mechanism of chain length determination
in type II GGPS. In addition, the two amino acid
insertion in FARM was deleted from type II FPS
of G. stearothermophilus (Fig. 6B). The mutant
FPS-DFARM also showed product specificity similar
A
B
Fig. 4. Introduction of substitution mutations into the region
upstream from FARM of P. ananatis GGPS. (A) Partial amino acid
sequences around FARM of wild-type and mutated enzymes are
aligned. The substituted amino acid residues are shaded. (B) TLC
autoradiochromatograms of the reaction products of wild-type and
mutated enzymes. The products were analyzed as described in the
Experimental procedures. The allylic substrate used is indicated at
the top of each autoradiochromatogram. Lane 1, H87A; lane 2,
V86A; lane 3, M85A; lane 4, wild-type. Under all assay conditions,
< 30% of each substrate reacted. Ori., origin; S.F., solvent front.
A
B
Fig. 5. Introduction of substitution mutations into the second posi-

tion upstream from the G(Q/E) motif of P. ananatis GGPS. (A) Par-
tial amino acid sequences around the G(Q/E) motif of wild-type and
mutated enzymes are aligned. The substituted amino acid residues
are shaded. (B) TLC autoradiochromatograms of the reaction prod-
ucts of the wild-type and mutated enzymes. The products were
analyzed as described in the Experimental procedures. The allylic
substrate used is indicated at the top of each autoradiochromato-
gram. Lane 1, V163A; lane 2, V163G; lane 3, V163S; lane 4, wild-
type. Under all assay conditions, < 30% of each substrate reacted.
Ori., origin; S.F., solvent front.
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3925
to that of the wild-type FPS, although the characteris-
tics around FARM mimicked those of type I GGPS
(Fig. 6D). Therefore, it was concluded that the two
amino acid insertion does not play an important role
in the chain length determination mechanism in either
type II GGPS or type II FPS.
Mutation in a-helix E
In the mutational study on type III GGPS from S. ce-
revisiae conducted by Chang et al. [9], the bottom of
the reaction cavity was suggested to be comprised not
only of histidine 139 at the second position upstream
from the G(Q/E) motif, but also of tyrosine 107 and
phenylalanine 108. Substituting alanine for tyrosine
107 and phenylalanine 108 increased the chain length
of the final products, as did histidine 139. These bulky
residues exist in proximity in the structure of the
enzyme, which was also reported in the same study.
Tyrosine 107 and phenylalanine 108 are located in

a-helix E (Chang et al. [9] referred to a-helix E as
a-helix F), whereas FARM and the G(Q/E) motif are
located in a-helices D and F, respectively (D and G
according to the designation of Chang et al. [9]).
Moreover, the structure of S. cerevisiae GGPS binding
GGPP recently reported (2E8V) revealed that tyrosine
107 directly touches the x-end of GGPP bound in the
same subunit, whereas phenylalanine 108 supplied
from the other subunit exists in the proximity of the
x-end [24].
These results led to the hypothesis that the prenyl-
chain of the product elongates in the space enclosed by
a-helices D, E and F, and that the bulky amino acid
residues on at least one of the a-helices block chain-
elongation. If this hypothesis is correct, type II GGPS
should use residues on a-helix E to terminate chain-
elongation. Thus, alanine substitution mutations were
introduced at each position on a-helix E where a bulky
amino acid was located (Fig. 7A). These bulky amino
acids on a-helix E can form the bottom of a reaction
cavity similar to those residues located at the key posi-
tions upstream from FARM and from the G(Q/E)
A
B
CD
Fig. 6. Deletion of the insertion sequences in FARM of P. ananatis GGPS and G. stearothermophilus FPS. (A) Partial amino acid sequences
around FARM of P. ananatis GGPS and mutated enzymes are aligned. The deleted positions are shaded. (B) Partial amino acid sequences
around FARM of G. stearothermophilus FPS and mutated enzymes are aligned. The deleted positions are shaded. (C) TLC autoradiochroma-
tograms of the reaction products of the P. ananatis GGPS and mutated enzymes. The products were analyzed as described in the Experi-
mental procedures. The allylic substrate used is indicated at the top of each autoradiochromatogram. Lane 1, GGPS-DFARM; lane 2,

wild-type GGPS. Under all assay conditions, < 30% of each substrate reacted. Ori., origin; S.F., solvent front. (D) TLC autoradiochromato-
grams of the reaction products of the G. stearothermophilus FPS and mutated enzymes. The products were analyzed as described in the
Experimental procedures. The allylic substrate used is indicated at the top of each autoradiochromatogram. Lane 1, FPS-DFARM; lane 2,
wild-type FPS. Under all assay conditions, < 30% of each substrate reacted. Ori., origin; S.F., solvent front.
Product determination mechanism of type II GGPS M. Noike et al.
3926 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
motif in the other types of the enzyme. Among the
constructed mutants, I121A and V125A yielded longer
products than the wild-type enzyme (Fig. 7B,C). I121A
gave a C
35
product when GGPP was used as the sub-
strate, whereas the final product of the wild-type
GGPS was C
25
prenyl diphosphate. On the other
hand, V125A yielded a series of products whose maxi-
mum chain length reached over C
40
when GPP or
GGPP was used as the primer substrate. The other
mutants showed product specificity similar to that of
the wild-type enzyme, although some mutants exhib-
ited negligible substrate specificity for GGPP.
A
B
C
Fig. 7. Introduction of substitutive mutations into the predicted a-helix E of P. ananatis GGPS. (A) Partial amino acid sequences around
a-helix E of wild-type and mutated enzymes are aligned. The substituted amino acid residues are shaded. (B) TLC autoradiochromatograms
of the reaction products of wild-type and mutated enzymes. The products were analyzed as described in the Experimental procedures. The

allylic substrate used is indicated at the top of each autoradiochromatogram. Lane 1, L122A; lane 2, I121A; lane 3, H118A; lane 4, E117A;
Lane 5, Y115A; lane 6, H114A; lane 7, wild-type. Under all assay conditions, < 30% of each substrate reacted. Ori., origin; S.F., solvent
front. (C) TLC autoradiochromatograms of the reaction products of wild-type and mutated enzymes. Lane 1, V125A; lane 2, L127A; lane 3,
wild-type.
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3927
Discussion
In the present study, type II GGPS from P. ananatis
was recombinantly expressed and purified by refolding
on a column. Blue native PAGE suggested that the
enzyme has a homotetrameric or homooctermeric
structure. Crystal structural analysis of human GGPS
reveals that three homodimers, which comprise the
same quaternary structure observed for most short-
chain (all-E) prenyl diphosphate synthases, join
together to form homohexamer [9]. Thus, in the case
of P. ananatis GGPS, it also is likely that two or four
homodimers join together to form a homotetramer or
a homooctamer, respectively. The recombinant enzyme
and its mutants were used to identify amino acid resi-
dues that contribute to the mechanism of chain length
determination. Unexpectedly, the two regions that are
known to play important roles in the mechanism in
some types of short-chain (all-E) prenyl diphosphate
synthases [i.e. the fourth and fifth positions upstream
from FARM and the second position upstream from
the G(Q/E) motif] were not involved in chain length
determination in type II GGPS. Moreover, a two
amino acid insertion in FARM, which was thought to
be involved in the mechanism of chain length determi-

nation, had no significant effect on the product chain
length in either type II GGPS or type II FPS. Alterna-
tively, alanine substitution mutations in a-helix E
revealed that isoleucine 121 and valine 125 are the resi-
dues involved in the mechanism of chain length deter-
mination. To the best of our knowledge, this is the
first report to describe mutations in type II GGPS that
change the chain length of the final product of the
enzyme.
Although it was apparent that a-helix E was
involved in the mechanism of chain length determina-
tion in type II GGPS, an additional question was
raised. The crystal structures of some of the homodi-
meric (all-E) prenyl diphosphate synthases indicated
that a-helix E exists at the dimer interface. Thus, the
question arose as to whether the critical residues (i.e.
I121 and V125) provided for the reaction cavity are
from the same catalytic subunit or from the other pair-
ing subunit? Fortunately, a crystal structure that was
recently solved has provided a clear answer to this
question. Kloer et al. [23] reported the crystal structure
of type II GGPS from mustard (Sinapis alba), binding
GGPP. In the homodimeric structure (2J1P), the gera-
nylgeranyl chain of GGPP elongates in the cavity
formed by four a-helices, D, E, F (from the catalytic
subunit) and E¢ (from the pairing subunit). V178¢ and
D182¢ in a-helix E¢, which correspond to I121 and
V125 of P. ananatis GGPS, respectively, exist much
closer to the geranylgeranyl chain than do V178 and
D182 in a-helix E (Fig. 8A). Especially, D182¢ directly

touches the x-end of the geranylgeranyl-chain.
Although V178¢ is not in direct contact with GGPP, it
appears to support D182¢ or L179¢ at the next position
in a-helix E¢, which touches the center of the geranyl-
geranyl-chain and bends it towards the bottom of the
cavity formed by L185 in a-helix E, I216 in a-helix F,
and D182¢ and S186¢ in a-helix E¢ (Fig. 8B, left). The
fourth and fifth residues upstream from FARM [i.e.
S147 and MSE (selenomethionine)146, respectively]
also are in contact with the geranylgeranyl chain, but
these residues appear to act only as part of the cavity
wall, as does the second residue upstream from the
GQ motif (i.e. V222). An almost similar spatial
arrangement was observed in the model structure of
A
B
Fig. 8. Structural information on the product determination mecha-
nism of type II GGPS. (A) The direction of the geranylgeranyl chain
of GGPP bound in a subunit (blue) of S. alba GGPS. The x-end of
the chain touches a-helix E¢ supplied from the other subunit (pink).
GGPP is indicated by a cylinder model and some equivalent resi-
dues on a-helices E and E¢ are shown as sphere models. (B) Close
view of the substrate pocket of S. alba GGPS (left) and that of the
modeled dimeric structure of P. ananatis GGPS (right). The model
of P. ananatis GGPS was constructed based on the crystal struc-
ture of S. alba GGPS as the template. Some of the amino acid
residues surrounding the geranylgeranyl chain of GGPP bound in
S. alba GGPS and the corresponding residues in P. ananatis GGPS
are indicated as sphere models. The geranylgeranyl chain is
indicated by green cylinders.

Product determination mechanism of type II GGPS M. Noike et al.
3928 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
P. ananatis GGPS, which was constructed using the
crystal structure of S. alba GGPS as the template for
molecular modeling (Fig. 8B, right). From this struc-
tural information, it is readily apparent that replace-
ment of V178 or D182 of S. alba GGPS with a smaller
amino acid would create a new chain-elongation path
through which the prenyl chain can lengthen across
the subunit interface. It is conceivable that a similar
scenario also occurred in the I121A and V125A
mutants of P. ananatis GGPS.
The role of a-helix E in the mechanism of chain length
determination also has been reported for a few short-
chain (all-E) prenyl diphosphate synthases of other
types. As mentioned above, type III GGPS from S. cere-
visiae utilizes tyrosine 107 and phenylalanine 108 to
terminate elongation of the product [9], although these
residues do not correspond with the positions 121 and
125 of P. ananatis GGPS. The recently reported struc-
ture of the complex of S. cerevisiae GGPS and GGPP
revealed that tyrosine 107 was supplied from the other
subunit [24], as is the case for I121 and V125 in type II
GGPS from P. ananatis. On the other hand, Lee et al.
[25] reported that, in type II FPS from E. coli, glycine
substitution of aspartate 115, which also exists in a-helix
E, allows the enzyme to produce GGPP. The position of
the aspartate residue corresponds to V125 in P. ananatis
GGPS. Lee et al. [25] hypothesized that the destruction
of the hydrogen bond between aspartate 115 and histi-

dine 83 increases the flexibility of the a-helices, expand-
ing the reaction cavity. In their paper it is likely that
only the arrangement of the residues in a monomer sub-
unit was considered as an explanation of the phenome-
non. However, in the crystal structure of E. coli FPS
binding IPP and the analogue of dimethylallyl diphos-
phate (DMAPP) (1RQI) [7], aspartate 115 exists in close
proximity to the tyrosine 79¢ at the fifth position
upstream from FARM of the other subunit, which is
considered to play a significant role in chain length
determination. This observation strongly suggests that
the aspartate residue might be supplied to offer struc-
tural support of tyrosine 79¢ or to block chain-elonga-
tion, probably in part, in the other subunit.
The results of the present study, which suggest the
requirement of subunit interaction for chain length
determination in type II GGPS, are reminiscent of an
intriguing study by Burke and Croteau [26]. These
authors reported that a subunit of homodimeric type
II GGPS from Taxus candensis and a small subunit
of heterotetrameric GPS from Mentha piperita can
form a hybrid heterodimer, which yields GPP when
DMAPP is used as the substrate. The large subunit of
M. piperita GPS is very similar to T. candensis GGPS,
whereas the small subunit is not. Thus, the small,
probably noncatalytic subunit was shown to influence
the product specificity of type II GGPS. It is conceiv-
able that the mechanism that acts in P. ananatis GGPS
is similar to that observed for the hybrid heteromeric
enzyme and to the mechanism that likely occurs in het-

eromeric GPSs from plants. Heteromeric longer-chain
(all-E) prenyl diphosphate synthases have been identi-
fied, including heptaprenyl diphosphate (C
35
) synthases
from bacilli [27–29]; hexaprenyl diphosphate (C
30
) syn-
thase from Micrococcus luteus B-P 26 [30]; solanesyl
diphosphate (C
45
) synthase from mouse [31]; and deca-
prenyl diphosphate (C
50
) synthase from human [31]
and Schizosaccharomyces pombe [32]. These heteromer-
ic enzymes may provide more definitive evidence for
subunit interaction in the mechanism of chain length
determination. Indeed, Zhang et al. [33] reported that
mutation in the small subunit of heterodimeric hepta-
prenyl diphsophate synthase from Bacillus subtilis,
which shows only slight similarity with homodimeric
(all-E) prenyl diphosphate synthases, affects the chain
length of the final product.
As noted above, the crystal structures of avian FPS
and human GGPS as the complexes with their final
products, FPP and GGPP, respectively, have been
solved [10,20]. However, the direction of prenyl-chain
elongation differs between these enzymes. In the struc-
ture of mutated avian type I FPS binding FPP

(1UBX), reported as the monomeric form, the farnesyl
chain elongates toward the expected dimer interface
[20]. Thus, the cavity of avian type I FPS is thought to
be constructed by a-helices D, E, F and probably E¢,
as is that of type II GGPS from S. alba. By contrast,
in human GGPS binding GGPP (2Q80), the x-end of
the geranylgeranyl chain enters the space enclosed by
a-helices C, D and G [10]. In the structure, the resi-
dues known to be important in the chain length deter-
mination [i.e. the fourth and fifth positions upstream
from FARM and the second position upstream from
the G(Q/E) motif] just come into contact with the
product at the center of the prenyl chain, suggesting
that these residues do not act to form the bottom of
the cavity in human type III GGPS. A similar path of
prenyl-chain elongation was suggested for hexaprenyl
diphosphate synthase from Sulfolobus solfataricus.Ina
mutational study, alanine or glycine substitution for
leucine 164 in a-helix G increased the chain length of
the final product [8]. However, enzymes with chain-
elongation paths enclosed by a-helices C, D and G
might be exceptional because the structural studies of
the other (all-E) prenyl diphosphate synthases, includ-
ing type III GGPS from S. cerevisiae [24] and octapre-
nyl diphosphate synthase from Thermotoga maritima
[34], as well as a large number of mutational studies,
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3929
suggest that the majority of enzymes have paths
enclosed by a-helices D, E and F. In the enzymes pos-

sessing structures similar to human GGPS, it is possi-
ble that a different type of chain length determination
mechanism exists in which amino acid residues in
unknown regions play crucial roles.
Experimental procedures
Materials
Precoated reversed-phase TLC plates, LKC-18F were pur-
chased from Whatman (Maidstone, UK). (all-E) GGPP,
(all-E) FPP and (all-E) GPP were synthesized as previously
reported [35]. Nonlabeled IPP and DMAPP were donated
by C. Ohto (Toyota Motor Co., Japan). [1-
14
C]IPP was
purchased from GE Healthcare (Piscataway, NJ, USA). All
other chemicals were of analytical grade.
General procedures
Restriction enzyme digestions, transformations and other
standard molecular biology techniques were carried out as
previously described [36].
Plasmid construction and site-directed
mutagenesis
Using pACYC-IBE [37], which contains carotenoid biosyn-
thetic genes from P. ananatis, as the template, the crtE gene
encoding GGPS was amplified using PCR with KOD DNA
polymerase (Toyobo, Osaka, Japan) and the primers: PaG
GPS-Fw, 5¢-AAGAAA
CATATGACGGTCTGCGCAAA
AAAACACG-3¢, and PaGGPS-Rv, 5¢-TGCAGA
GGATCC
TTAACTGACGGCAGCGAGTTTTTTG-3¢. The sequen-

ces corresponding to the NdeI and BamHI sites that were
used in subsequent experiments are underlined in the pri-
mer sequences above. The amplified fragment was cleaved
with the restriction enzymes and then inserted into an
NdeI/BamHI-treated pET-15b vector (Novagen, Madison,
WI, USA) to construct pET-HisPaGGPS, a plasmid for the
recombinant expression of His
6
-tagged P. ananatis GGPS.
For the expression of His
6
-tagged G. stearothermophilus
FPS, the gene was amplified using PCR with KOD DNA
polymerase (Toyobo), a pFPS [21], and the primers: His-
BsFPS-Fw, 5¢-ACAG
CCATGGGACATCATCATCATCA
TCATGCGCAGCTTTCAGTTGAA-3¢, and HisBsFPS-
Rv, 5¢-TGAATTTA
AAGCTTAATGGTCGCGGGCG-3¢.
The sequences corresponding to the NcoI and HindIII sites
that were used in subsequent experiments are underlined in
the above sequences. The amplified fragment was cleaved
with the restriction enzymes and then inserted into the
NcoI/HindIII-treated pTV118N vector (TaKaRa, Shiga,
Japan) to construct pTV-HisBsFPS. Site-directed mutations
were introduced into each parental plasmid utilizing a
QuikChange Mutagenesis Kit (Stratagene, La Jolla, CA,
USA) according to the manufacturer’s instructions.
Expression and purification of wild-type and
mutated enzymes

For the expression of P. ananatis GGPS, E. coli
BL21(DE3) was transformed with pET-HisPaGGPS or the
mutated plasmids. The transformants were cultivated in
50 mL of M9 minimal broth supplemented with glycerol
(2 gÆL
)1
), yeast extract (2 gÆL
)1
) and ampicillin (50 mgÆL
)1
).
When D
600
of 0.5 was reached, the transformed bacteria in
the culture were induced with 1.0 mm isopropyl thio-ß-d-
galactoside. The cells were incubated overnight and then
harvested. The cells were disrupted in lysis buffer contain-
ing 20 mm sodium phosphate buffer (pH 8.0), 10 mm
imidazol and 0.5 m NaCl. The homogenate was centrifuged
at 6000 g for 15 min at 4 ° C and the precipitate containing
the inclusion body enzyme was recovered. The precipitate
was dissolved in lysis buffer supplemented with 4%
Triton X-100. After shaking for 30 min at 25 °C, the
mixture was centrifuged at 6000 g for 15 min at 4 °C and
the precipitate was recovered. The process was repeated
twice to remove bacterial membranous compounds. The
washed precipitate was lyophilized and used as the inclusion
body. Ten milligrams of the inclusion body was dissolved in
10 mL of denaturation buffer containing 50 mm sodium
phosphate buffer (pH 8.0), 10 mm imidazol, 6 m guanidine

hydrochloride, 10 mm dithiothreitol, 10 mm 2-mercapto-
ethanol and 0.5 m NaCl. After centrifugation at 9000 g for
15 min at 4 °C, the supernatant was recovered and then
applied to a HisTrap column (GE Healthcare) equilibrated
with equilibration buffer containing 50 mm sodium
phosphate buffer (pH 8.0), 10 mm imidazol, 6 m guanidine
hydrochloride, 10 mm dithiothreitol, 10 mm 2-mercapto-
ethanol and 0.5 m NaCl. The column was washed with
10 mL of equilibration buffer and then with start buffer con-
taining 50 mm sodium phosphate buffer (pH 8.0), 10 mm im-
idazol, 10 mm 2-mercaptoethanol and 0.5 m NaCl to remove
the denaturant. The protein renatured in the column was
eluted from the column with 10 mL of elution buffer contain-
ing 50 mm sodium phosphate buffer (pH 8.0), 500 mm imi-
dazol, 10 mm 2-mercaptoethanol and 0.5 m NaCl. The eluate
was fractioned, and the fraction with the highest enzyme
activity and purity was used as the partially purified enzyme
in the experiments described below. The purity of the enzyme
was determined by 15% SDS/PAGE.
For the expression of G. stearothermophilus FPS, E. coli
DH5a was used as the host. The transformants were culti-
vated and induced as described above. Disruption of
harvested cells and the purification of the tagged enzymes
were conducted utilizing a MagExtractor His-tag Kit
(Toyobo) according to the manufacturer’s instructions.
Product determination mechanism of type II GGPS M. Noike et al.
3930 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
Enzyme purify was determined by 15% SDS/PAGE (data
not shown).
Analysis of quaternary structure of refolded

P. ananatis GGPS
The quaternary structures of refolded wild-type and
mutated P. ananatis GGPSs were analyzed by blue native
PAGE [38] using a NativePAGE Ô NovexÒ 4–16% Bis-Tris
Gel (Invitrogen, Carlsbad, CA, USA). The cathode buffer
was composed of 50 mm BisTris, 50 mm Tricine (pH 6.8)
and 0.02% Coomassie Brilliant Blue G 250, whereas the
anode buffer contained 50 mm BisTris (pH 6.8).
Measurement of prenyltransferase activity
The assay mixture for wild-type and mutant P. ananatis
GGPSs contained, in a final volume of 200 lL, 0.5 nmol of
[1-
14
C]IPP (2 GBqÆmmol
)1
), 0.5 nmol of an allylic primer
(GPP, FPP or GGPP), 1 lmol of MgCl
2
,2lmol of a Tris–
HCl buffer (pH 8.0) and a suitable amount of each enzyme.
This mixture was incubated at 30 °C for 15 min and the
reaction was stopped by chilling the mixture in an ice bath.
The mixture was shaken with 600 lL of 1-butanol satu-
rated with H
2
O. The butanol layer was washed with water
saturated with NaCl, and the radioactivity in 100 lL of the
butanol layer was measured with a TRI-CARB 1600 liquid
scintilation counter (Packard Instrument Company, Meri-
den, CT, USA). The remaining butanol layer was used for

product analysis. For the assay of G. stearothrmophilus
FPS, the reaction mixture contained 5 nmol [1-
14
C]IPP
(2 GBqÆnmol
)1
), 5 nmol allylic primer (DMAPP or GPP),
1 lmol MgCl
2
,2lmol Tris/HCl buffer (pH 8.5), 10 lmol
ammonium chloride, 10 lmol 2-mercaptoethanol and a
suitable amount of each enzyme in a final volume of
200 lL. The mixture was incubated at 55 °C for 10 min
and then processed as described above.
Product analysis
Prenyl diphosphates in the residual 1-butanol layer were trea-
ted with acid phosphatase according to the method of Fujii
et al. [39]. The hydrolysates were extracted with pentane and
analyzed by reversed-phase TLC using a precoated plate,
LKC-18F, developed with acetone/H
2
O (9 : 1). Authentic
standard alcohols were visualized with iodine vapor, and the
distribution of radioactivity was detected using a Molecular
imager (Bio-Rad, Hercules, CA, USA) or a BAS 1000 Mac
bioimaging analyzer (Fujifilm, Tokyo, Japan).
Homology modeling
The 3D model of dimeric P. ananatis GGPS was con-
structed with MODELLER [40], using the structure of
dimeric S. alba GGPS binding a molecule of GGPP (2J1P)

as the template. pymol () and imol
( were used to generate figures
from the known structure and constructed model.
Acknowledgements
This work was supported in part by Grants-in-Aid
from the Ministry of Education, Culture, Sports, Sci-
ence, and Technology of Japan. We are grateful to Dr
Chikara Ohto, Toyota Motor Co., for providing iso-
pentenyl diphosphate and dimethylallyl diphosphate.
References
1 Ogura K & Koyama T (1998) Enzymatic aspects of iso-
prenoid chain elongation. Chemical Reviews 98, 1263–
1276.
2 Koyama T (1999) Molecular analysis of prenyl chain
elongating enzymes. Biosci Biotechnol Biochem 63,
1671–1676.
3 Christianson DW (2007) Chemistry. Roots of biosyn-
thetic diversity. Science 316, 60–61.
4 Tarshis LC, Yan M, Poulter CD & Sacchettini JC
(1994) Crystal structure of recombinant farnesyl diphos-
phate synthase at 2.6-A
˚
resolution. Biochemistry 33,
10871–10877.
5 Guo RT, Kuo CJ, Chou CC, Ko TP, Shr HL, Liang
PH & Wang AH (2004) Crystal structure of octaprenyl
pyrophosphate synthase from hyperthermophilic Ther-
motoga maritima and mechanism of product chain
length determination. J Biol Chem 279, 4903–4912.
6 Mao J, Gao YG, Odeh S, Robinson H, Montalvetti A,

Docampo R & Oldfield E (2004) Crystallization and
preliminary X-ray diffraction study of the farnesyl
diphosphate synthase from Trypanosoma brucei. Acta
Crystallogr D Biol Crystallogr 60, 1863–1866.
7 Hosfield DJ, Zhang Y, Dougan DR, Broun A, Tari
LW, Swanson RV & Finn J (2004) Structural basis for
bisphosphonate-mediated inhibition of isoprenoid bio-
synthesis. J Biol Chem 279, 8526–8529.
8 Sun HY, Ko TP, Kuo CJ, Guo RT, Chou CC, Liang
PH & Wang AH (2005) Homodimeric hexaprenyl pyro-
phosphate synthase from the thermoacidophilic crenar-
chaeon Sulfolobus solfataricus displays asymmetric
subunit structures. J Bacteriol 187, 8137–8148.
9 Chang TH, Guo RT, Ko TP, Wang AH & Liang PH
(2006) Crystal structure of type-III geranylgeranyl pyro-
phosphate synthase from Saccharomyces cerevisiae and
the mechanism of product chain length determination.
J Biol Chem 281, 14991–15000.
10 Kavanagh KL, Dunford JE, Bunkoczi G, Russell RG
& Oppermann U (2006) The crystal structure of human
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3931
geranylgeranyl pyrophosphate synthase reveals a novel
hexameric arrangement and inhibitory product binding.
J Biol Chem 281, 22004–22012.
11 Kavanagh KL, Guo K, Dunford JE, Wu X, Knapp S,
Ebetino FH, Rogers MJ, Russell RG & Oppermann U
(2006) The molecular mechanism of nitrogen-containing
bisphosphonates as antiosteoporosis drugs. Proc Natl
Acad Sci USA 103, 7829–7834.

12 Gabelli SB, McLellan JS, Montalvetti A, Oldfield E,
Docampo R & Amzel LM (2006) Structure and mecha-
nism of the farnesyl diphosphate synthase from Trypan-
osoma cruzi: implications for drug design. Proteins 62,
80–88.
13 Sagami H, Morita Y & Ogura K (1994) Purification and
properties of geranylgeranyl-diphosphate synthase from
bovine brain. J Biol Chem 269, 20561–20566.
14 Kuzuguchi T, Morita Y, Sagami I, Sagami H & Ogura
K (1999) Human geranylgeranyl diphosphate synthase.
cDNA cloning and expression. J Biol Chem 274, 5888–
5894.
15 Wang K & Ohnuma S (1999) Chain-length determina-
tion mechanism of isoprenyl diphosphate synthases and
implications for molecular evolution. Trends Biochem
Sci 24, 445–451.
16 Hemmi H, Noike M, Nakayama T & Nishino T (2003)
An alternative mechanism of product chain-length
determination in type III geranylgeranyl diphosphate
synthase. Eur J Biochem 270, 2186–2194.
17 Ohnuma S-i, Hirooka K, Hemmi H, Ishida C, Ohto C
& Nishino T (1996) Conversion of product specificity of
archaebacterial geranylgeranyl-diphosphate synthase.
Identification of essential amino acid residues for chain
length determination of prenyltransferase reaction.
J Biol Chem 271, 18831–18837.
18 Ohnuma S-i, Hirooka K, Ohto C & Nishino T (1997)
Conversion from archaeal geranylgeranyl diphosphate
synthase to farnesyl diphosphate synthase. Two amino
acids before the first aspartate-rich motif solely deter-

mine eukaryotic farnesyl diphosphate synthase activity.
J Biol Chem 272, 5192–5198.
19 Ohnuma S-i, Hirooka K, Tsuruoka N, Yano M, Ohto
C, Nakane H & Nishino T (1998) A pathway where
polyprenyl diphosphate elongates in prenyltransferase.
Insight into a common mechanism of chain length
determination of prenyltransferases. J Biol Chem 273,
26705–26713.
20 Tarshis LC, Proteau PJ, Kellogg BA, Sacchettini JC &
Poulter CD (1996) Regulation of product chain length
by isoprenyl diphosphate synthases. Proc Natl Acad Sci
USA 93, 15018–15023.
21 Ohnuma S-i, Nakazawa T, Hemmi H, Hallberg AM,
Koyama T, Ogura K & Nishino T (1996) Conversion
from farnesyl diphosphate synthase to geranylgeranyl
diphosphate synthase by random chemical mutagenesis.
J Biol Chem 271, 10087–10095.
22 Ohnuma S-i, Narita K, Nakazawa T, Ishida C, Takeu-
chi Y, Ohto C & Nishino T (1996) A role of the amino
acid residue located on the fifth position before the first
aspartate-rich motif of farnesyl diphosphate synthase on
determination of the final product. J Biol Chem 271,
30748–30754.
23 Kloer DP, Welsch R, Beyer P & Schulz GE (2006)
Structure and reaction geometry of geranylgeranyl
diphosphate synthase from Sinapis alba. Biochemistry
45, 15197–15204.
24 Guo RT, Cao R, Liang PH, Ko TP, Chang TH, Hu-
dock MP, Jeng WY, Chen CK, Zhang Y, Song Y et al.
(2007) Bisphosphonates target multiple sites in both cis-

and trans-prenyltransferases. Proc Natl Acad Sci USA
104, 10022–10027.
25 Lee PC, Petri R, Mijts BN, Watts KT & Schmidt-
Dannert C (2005) Directed evolution of Escherichia coli
farnesyl diphosphate synthase (IspA) reveals novel
structural determinants of chain length specificity.
Metab Eng 7, 18–26.
26 Burke C & Croteau R (2002) Interaction with the small
subunit of geranyl diphosphate synthase modifies the
chain length specificity of geranylgeranyl diphosphate
synthase to produce geranyl diphosphate. J Biol Chem
277, 3141–3149.
27 Koike-Takeshita A, Koyama T, Obata S & Ogura K
(1995) Molecular cloning and nucleotide sequences of
the genes for two essential proteins constituting a novel
enzyme system for heptaprenyl diphosphate synthesis.
J Biol Chem 270, 18396–18400.
28 Zhang YW, Koyama T & Ogura K (1997) Two cistrons
of the gerC operon of Bacillus subtilis encode the two
subunits of heptaprenyl diphosphate synthase. J Bacte-
riol 179, 1417–1419.
29 Zhang YW, Koyama T, Marecak DM, Prestwich GD,
Maki Y & Ogura K (1998) Two subunits of heptapre-
nyl diphosphate synthase of Bacillus subtilis form a
catalytically active complex. Biochemistry 37,
13411–13420.
30 Shimizu N, Koyama T & Ogura K (1998) Molecular
cloning, expression, and characterization of the genes
encoding the two essential protein components of
Micrococcus luteus B-P 26 hexaprenyl diphosphate syn-

thase. J Bacteriol 180, 1578–1581.
31 Saiki R, Nagata A, Kainou T, Matsuda H & Kawamu-
kai M (2005) Characterization of solanesyl and decapre-
nyl diphosphate synthases in mice and humans. FEBS J
272, 5606–5622.
32 Saiki R, Nagata A, Uchida N, Kainou T, Matsuda H
& Kawamukai M (2003) Fission yeast decaprenyl
diphosphate synthase consists of Dps1 and the newly
characterized Dlp1 protein in a novel heterotetrameric
structure. Eur J Biochem 270, 4113–4121.
33 Zhang YW, Li XY & Koyama T (2000) Chain length
determination of prenyltransferases: both heteromeric
Product determination mechanism of type II GGPS M. Noike et al.
3932 FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS
subunits of medium-chain (E)-prenyl diphosphate
synthase are involved in the product chain length
determination. Biochemistry 39, 12717–12722.
34 Guo RT, Kuo CJ, Ko TP, Chou CC, Liang PH &
Wang AH (2004) A molecular ruler for chain elonga-
tion catalyzed by octaprenyl pyrophosphate synthase
and its structure-based engineering to produce unprece-
dented long chain trans-prenyl products. Biochemistry
43, 7678–7686.
35 Davisson VJ, Woodside AB & Poulter CD (1985)
Synthesis of allylic and homoallylic isoprenoid
pyrophosphates. Methods Enzymol 110,
130–144.
36 Sambrook J, Fritsch EF & Maniatis T (1989)
Molecular Cloning : A Laboratory Manual. 2nd edn.
Cold Spring Harbor Laboratory Press, Cold Spring

Harbor, NY.
37 Ohnuma S-i, Suzuki M & Nishino T (1994) Archaebac-
terial ether-linked lipid biosynthetic gene. Expression
cloning, sequencing, and characterization of geranylger-
anyl-diphosphate synthase. J Biol Chem 269,
14792–14797.
38 Wittig I, Braun HP & Schagger H (2006) Blue native
PAGE. Nat Protoc 1, 418–428.
39 Fujii H, Koyama T & Ogura K (1982) Efficient
enzymatic hydrolysis of polyprenyl pyrophosphates.
Biochim Biophys Acta 712, 716–718.
40 Sali A & Blundell TL (1993) Comparative protein
modelling by satisfaction of spatial restraints. J Mol
Biol 234, 779–815.
M. Noike et al. Product determination mechanism of type II GGPS
FEBS Journal 275 (2008) 3921–3933 ª 2008 The Authors Journal compilation ª 2008 FEBS 3933