Structure-Function Studies of Artemisia tridentata Farnesyl Diphosphate Synthase and Chrysanthemyl Diphosphate Synthase by Site-Directed Mutagenesis and Morphogenesis

Article abstract of DOI:10.1021/jacs.7b07608

The amino acid sequences of farnesyl diphosphate synthase (FPPase) and chrysanthemyl diphosphate synthase (CPPase) from Artemisia tridentata ssp. Spiciformis, minus their chloroplast targeting regions, are 71% identical and 90% similar. FPPase efficiently and selectively synthesizes the "regular" sesquiterpenoid farnesyl diphosphate (FPP) by coupling isopentenyl diphosphate (IPP) to dimethylallyl diphosphate (DMAPP) and then to geranyl diphosphate (GPP). In contrast, CPPase is an inefficient promiscuous enzyme, which synthesizes the "irregular" monoterpenes chrysanthemyl diphosphate (CPP), lavandulyl diphosphate (LPP), and trace quantities of maconelliyl diphosphate (MPP) from two molecules of DMAPP, and couples IPP to DMAPP to give GPP. A. tridentata FPPase and CPPase belong to the chain elongation protein family (PF00348), a subgroup of the terpenoid synthase superfamily (CL0613) whose members have a characteristic α terpene synthase α-helical fold. The active sites of A. tridentata FPPase and CPPase are located within a six-helix bundle containing amino acids 53 to 241. The two enzymes were metamorphosed into one another by sequentially replacing the loops and helices of the six-helix bundle from enzyme with those from the other. Chain elongation was the dominant activity during the N-terminal to C-terminal metamorphosis of FPPase to CPPase, with product selectivity gradually switching from FPP to GPP, until replacement of the final α-helix, whereupon cyclopropanation and branching activity competed with chain elongation. During the corresponding metamorphosis of CPPase to FPPase, cyclopropanation and branching activities were lost upon replacement of the first helix in the six-helix bundle. Mutations of active site residues in CPPase to the corresponding amino acids in FPPase enhanced chain-elongation activity, while similar mutations in the active site of FPPase failed to significantly promote formation of significant amounts of irregular monoterpenes. Our results indicate that CPPase, a promiscuous enzyme, is more plastic toward acquiring new activities, whereas FPPase is more resistant. Mutations of residues outside of the α terpene synthase fold are important for acquisition of FPPase activity for synthesis of CPP, LPP, and MPP.

Full text of DOI:10.1021/jacs.7b07608

Subscriber access provided by UNIV OF CAMBRIDGE
Article
Structure-Function Studies of Artemisia tridentata Farnesyl
Diphosphate Synthase and Chrysanthemyl Diphosphate
Synthase by Site-Directed Mutagenesis and Morphogenesis
J. Scott Lee, Jian-Jung Pan, Gurusankar Ramamoorthy, and C Dale Poulter
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b07608 • Publication Date (Web): 19 Sep 2017
Downloaded from http://pubs.acs.org on September 19, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Structure-Function Studies of Artemisia tridentata Farnesyl
Diphosphate Synthase and Chrysanthemyl Diphosphate
Synthase by Site-Directed Mutagenesis and Morphogenesis
J. Scott Lee,† Jian-Jung Pan, Gurusankar Ramamoorthy and C. Dale
Poulter*
Department of Chemistry
University of Utah
315 South 1400 East, RM 2020
Salt Lake City, UT 84112
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 41
1
2
3
4
5
6
7
8
9
Abstract. The amino acid sequences of farnesyl diphosphate synthase (FPPase) and
chrysanthemyl diphosphate synthase (CPPase) from Artemisia tridentata ssp. Spiciformis,
minus their chloroplast targeting regions, are 71% identical and 90% similar. FPPase
efficiently and selectively synthesizes the “regular” sesquiterpenoid farnesyl diphosphate
(FPP) by coupling isopentenyl diphosphate (IPP) to dimethylallyl diphosphate (DMAPP)
and then to geranyl diphosphate (GPP). In contrast, CPPase is an inefficient promiscuous
enzyme, which synthesizes the “irregular” monoterpenes chrysanthemyl diphosphate
(CPP), lavandulyl diphosphate (LPP), and trace quantities of maconelliyl diphosphate
(MPP) from two molecules of DMAPP, and couples IPP to DMAPP to give GPP. A. tridentata
FPPase and CPPase belong to the chain elongation protein family (PF00348), a subgroup of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
the terpenoid synthase superfamily (CL0613) whose members have a characteristic
α
terpene synthase α-helical fold. The active sites of A. tridentata FPPase and CPPase are
located within a six-helix bundle containing amino acids 53 to 241. The two enzymes were
metamorphosed into one another sequentially replacing the loops and helices of the six-
helix bundle from enzyme with those from the other. Chain elongation was the dominant
activity during the N-terminal to C-terminal metamorphosis of FPPase to CPPase, with
product selectivity gradually switching from FPP to GPP, until replacement of the final α-
helix, whereupon cyclopropanation and branching activity competed with chain elongation.
During the corresponding metamorphosis of CPPase to FPPase, cyclopropanation and
branching activities were lost upon replacement of the first helix in the six-helix bundle.
Mutations of active site residues in CPPase to the corresponding amino acids in FPPase
enhanced chain-elongation activity, while similar mutations in the active site of FPPase
failed to significantly promote formation of significant amounts of irregular monoterpenes.
ACS Paragon Plus Environment

Page 3 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Our results indicate that CPPase, a promiscuous enzyme, is more plastic toward acquiring
new activities; whereas, FPPase is more resistant. Mutations of residues outside of the α
terpene synthase fold are important for acquisition of FPPase activity for synthesis of CPP,
LPP, and MPP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 41
1
2
3
4
INTRODUCTION
5
6
7
8
Isoprenoid compounds are readily distinguished from other natural products by the
distinctive methylbutyl (isoprene) units embedded within their carbon skeletons. With
over 80,000 members,1 they constitute the largest and most structurally diverse group of
compounds synthesized in nature.1 Common metabolites of the isoprenoid biosynthetic
pathway include sterols, carotenoids, ubiquinones, and mono-, sesqui-, and diterpenes.1-3
The carbon skeletons of isoprenoid compounds are constructed from the fundamental C5
isomeric building blocks isopentenyl diphosphate (IPP) and dimethylallyl diphosphate
(DMAPP), which are joined by chain elongation (1’-4), branching (1’-2), cyclopropanation
(c1’-2-3), and cyclobutanation (c1’-2-3-2’) reactions (Figure 1).2 The 1’-4 pattern is the
most common attachment found between isoprene units and is typically called a “regular”
or “head-to-tail” coupling.4 The other three are termed “irregular” or “non-head-to-tail”.5
The 1’-4 pattern results from a reaction that joins C1 in DMAPP, or its higher isoprenologs,
to C4 in IPP. Thus, beginning with DMAPP, a homologous series of linear C10, C15 C20, etc.
allylic diphosphates is formed, where each member can be the substrate for construction of
more complex skeletons, including those with c1’-2-3, 1’-2, and c1’-2-3-2’ substructures.2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 5 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
(1'-4)
(c1'-2-3)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(1'-2)
(c1'-2-3-2')
Figure 1. Four basic patterns for incorporation of isoprene units into metabolites by
joining two isoprenoid diphosphates. Bonds between isoprene units are in green. The
isoprene units are numbered according to the numbering schemes for IPP and DMAPP.
Two structurally distinct families of enzymes catalyze chain elongation. While each
contains enzymes selective for the synthesis of products with different chain lengths, one
family synthesizes allylic isoprenoid diphosphates with E-double bonds, and the other,
allylic diphosphates with Z-double bonds. E-polyprenyl diphosphate synthases typically
produce short-chain products early in the isoprenoid pathway.6 These enzymes belong to
the chain elongation family (PF00348) within the terpenoid synthase superfamily
(CL0613),7 whose members also include protein families for
cyclopropanation/rearrangement (PF00494) and cyclization (PF03936 and PF01397)
(Figure 2). While all proteins in CL0613 contain a characteristic all α-helix α terpene
synthase motif, 7-9 those belonging to PF03936 and PF01397 often have additional β and/or
γ domains.10 Enzymes in the Z-polyprenyl synthase family belong to PF01255 and
synthesize longer-chain products.11,12
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Sequence similarity network (Blast cut-off of e-7) for enzymes in the terpenoid
synthase superfamily (CL0613); chain elongation (PF00348, red),
cyclopropanation/rearrangement (PF00494, gold), cyclization (PF03936, green; PF01397,
turquoise).
Chain elongation is the most common reaction in the isoprenoid pathway. The first
step in E-selective chain elongation couples DMAPP and IPP to produce geranyl
diphosphate (GPP). GPP can then couple with another molecule of IPP to give farnesyl
diphosphate (FPP), and so on.4 The isoprene units are joined by a dissociative electrophilic
alkylation of the nucleophilic double bond in IPP by an electrophilic cation generated from
the allylic diphosphate cosubstrate (DMAPP, GPP, FPP, etc.) (Scheme 1).13, 14 Extension of
the allylic chain by sequential addition of molecules of IPP produces a homologous series of
allylic isoprenoid diphosphates that in turn are substrates for branch point enzymes that
feed into the numerous branches of the terpenoid pathway.15
ACS Paragon Plus Environment

Page 7 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
PPO
H
PPO
H
+
PPi
R
R
OPP
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PPO
PPO
+
H
H+
R
R
Scheme 1. Electrophilic alkylation mechanism for chain elongation. R = CH3, C6H11, C11H19,
etc.
The cyclopropanation reaction is also common, but this linkage is subsequently
disassembled in most metabolites and is typically detected indirectly through rearranged
carbon skeletons in downstream metabolites.16 For example, the first pathway specific
reaction in biosynthesis of sterols and carotenoids is a c1’-2-3 cyclopropanation, followed
by a rearrangement to give structures with 1’-1 (head-to-head) linkages between the two
component isoprenoid fragments.17, 18 The mechanism for cyclopropanation is also a
dissociative electrophilic alkylation, where the allylic cation generated by one molecule of
an allylic diphosphate attacks the allylic double bond in another molecule, as illustrated for
the coupling of DMAPP to give chrysanthemyl diphosphate (CPP) in Scheme 2.2 Metabolites
with structures derived from 1’-2 branching, as illustrated by lavandulyl diphosphate
(LPP), are much less common,19 while those with 1’-2-3-2’ cyclobutane structures,
illustrated by maconelliyl diphosphate (MPP), have only been reported for mealy bug
mating pheromones.20,21 The 1’-2 and c1’-2-3-2’ structures are produced from the same
protonated cyclopropane intermediates that give the c1’-2-3 skeletons.2,22
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 41
1
2
3
4
5
6
7
8
9
H
PPO
PPO
DMAPP
H
H
H
CPP
H+
OPP
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H
PPO
+
PPO
H
H
+
PPi
H
H
+
PPO
PPO
H
+
H+
H+
H+
H
H
PPO
PPO
H
MPP
LPP
Scheme 2. An integrated electrophilic alkylation-rearrangement mechanism for formation
of cyclopropane, branched, and cyclobutane skeletons.
Enzymes that catalyze c1’-2-3 cyclopropanation typically couple two molecules of
DMAPP to give chrysanthemyl diphosphate (CPP),22 two molecules of FPP to give
presqualene diphosphate,23,24 or two molecules of geranylgeranyl diphosphate to give
prephytoene diphosphate.25 Presqualene and prephytoene diphosphates are intermediates
in the sterol and carotenoid pathways, respectively. Enzymes that catalyze 1’-2 branching
ACS Paragon Plus Environment

Page 9 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
are found in PF00348 and the structurally unrelated Z-polyprenyl synthase family
PF01255.26 The only wild-type enzyme known to catalyze c1’-2-3-2’ cyclobutanation is
CPPase, where cyclobutanation is a minor promiscuous activity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The enzymes in CL0613 have distinctive aspartate/glutamate-rich sequences that,
along with Mg2+ ions, bind the diphosphate moieties of their substrates and catalyze
carbocation formation.27,28 The structures of proteins in PF00348 and PF00494 suggest
that they diverged from a common chain elongation ancestor early in the evolution of
terpenoid biosynthesis.10 This hypothesis is supported by the recent emergence of a new
cyclopropanation enzyme, chrysanthemyl diphosphate synthase (CPPase) in
chrysanthemums (Chyrsanthemum cinerariifolium)22 and sagebrush (Artemisia tridentata
ssp. Spiciformis)29 from farnesyl diphosphate synthase (FPPase) in the same plants. The
chain elongation and cyclopropanation enzymes are members of PF00348 and share
structural features found in other chain elongation enzymes in that family.6 A. tridentata
contains two FPPases that are highly selective and efficient for chain elongation of DMAPP
to FPP. In contrast, A. tridentata CPPase is an inefficient, promiscuous enzyme that
converts two molecules of DMAPP to CPP (c1’-2-3, cyclopropanation), lavandulyl
diphosphate (LPP, 1’-2, branching), and maconelliyl diphosphate (MPP, c1’-2-3-2’,
cyclobutanation), as well as IPP and DMAPP to GPP (1’-4, chain elongation). These
properties are consistent with those expected for recently evolved enzyme that has gained
a new function but has not yet optimized its ability to catalyze the new reaction.30
E-Polyprenyl chain elongases have five conserved regions within a six-helix bundle
that contains characteristic aspartate-rich DDxx(xx)D diphosphate binding motifs.31,32
Differences between A. tridentata FPPase and CPPase within this bundle are minimal. The
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 41
1
2
3
4
5
6
7
8
9
substantial similarity between FPPase and CPPase prompted us to construct chimeric A.
tridentata proteins where FPPase was transformed into CPPase, and vice versa, by
replacing conserved regions of one enzyme with those from the other.6 The constructs
retained the ability to catalyze chain elongation while acquiring the ability to catalyze
branching, cyclobutanation and cyclopropanation. We now present product and kinetic
studies for chimeras where each helix and loop between the first and last helix in the six-
helix bundle in one enzyme is replaced by the corresponding sequence in the other and for
site-directed mutants where only active-site residues are replaced. Our results illustrate
the importance of residues outside of the active site to influence the regioselectivity of the
reactions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS
Engineered proteins. i. Chimeras. Chimeric enzymes were constructed from A.
tridentata FPPase and CPPase by replacing the amino acids in one protein with those from
the other at 13 different junctions, beginning at the N-terminus using megaprimer whole
plasmid (MEGAWHOP) and structure–based combinatorial protein engineering (SCOPE)
PCR protocols.33,34 The first PCR reaction synthesized a megaprimer that codes for amino
acid sequences from the N-terminus of the chimeric proteins, fused at the appropriate
junction site, to a short segment that codes for amino acids in the C-terminal segment. The
second PCR synthesized an expression plasmid for the chimeric protein (Figure S1).
Names for the chimeras are based on the enzyme from which the N-terminal
sequence was derived, F for FPPase or C for CPPase, and the junction at which the sequence
changes to the other enzyme. For instance, the “C6” chimera consists of CPPase sequence
ACS Paragon Plus Environment

Page 11 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
from the N-terminus to the sixth junction, with the remaining sequence from FPPase (see
Figure 3). The precise locations of the helices were identified by comparisons of X-ray
structures for FPPase and confirmed by the structure of chimera C3. Chimeras were
constructed from FPPase and CPPase by fusing the proteins at 13 junctions to produce
chimeric fusions at each helix/loop transition from amino acids 53 to 241, including the
junction between amino acids 193 and 194 located on either side of a kink in helix VII. Two
additional FPPase-CPPase chimeras constructed where the six-helix bundle consisting of
sequences from amino acids 53 to 241 from one enzyme (core) were flanked with the
corresponding N- and C-terminal sequences (scaffold) from the other. These core-scaffold
chimeras, with cores from FPPase and CPPase, are named c-F-c and f-C-f, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 41
1
2
3
4
A
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
B
Figure 3. Part A. Sequence alignments for A. tridentata FPPase and CPPase, where helices
in the N-terminal sequence (light gray), six-helix bundle (colored), and C-terminal regions
are highlighted. K193 and T194 lie on either side of a kink between helix VIIa (dark blue)
and helix VIIb (cyan). The thirteen junctions used for morphogenesis are located at the
beginning and end of each helix and at the KT junction at the kink between VIIa and VIIb.
Active site residues in CPPase that are different from those of FPPase are in bold. Part B.
Ribbon structures from the side and the top of a subunit in the A. tridentata FPPase
homodimer.
ACS Paragon Plus Environment

Page 13 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
ii. Site-Directed Mutants. Six amino acids in the active site of FPPase (F92, A161,
T194, F231, D235 and R343) differ from the corresponding residues in CPPase (I92, I161,
G194, Y231, N235, and G343). In FPPase, F92 and A161 form part of the binding pocket for
the hydrocarbon moieties in DMAPP and GPP, the polar amino acids are in contact with the
diphosphate resides or the tightly bound Mg2+ ions, and F231 and D235 form part of the
binding pocket for the hydrocarbon moiety in IPP. Substitution of A161 by I161 shortens
the binding pocket for the hydrocarbon moiety of the allylic substrate in CPPase,
decreasing the preference for binding GPP over DMAPP. Ten site-directed mutants of
FPPase and CPPase were constructed where one, or a combination of, the active site
residues in one enzyme were replaced with those from the other to give FPPase single
mutants A161I, T194G, F231Y, D235N, double mutant F231Y/D235N (2mFPPase), triple
mutant A161I/T194G/D235N (3mFPPase), quintuple mutant A161I/T194G/
F231Y/D235N/ R343G (5mFPPase) and sextuple mutant
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
F92I/A161I/T194G/F231Y/D235N/R343G (6mFPPase); and CPPase single mutant N235D
and triple-mutant I161A/G194T/N235D (3mCPPase). All of the mutants contained an N-
terminal His6 tag to facilitate purification by Ni2+ affinity chromatography.
Product studies. Products were determined for incubations with 0.5 mM IPP and 0.5
mM DMAPP (chain elongation conditions), which gave products from chain elongation,
cyclopropanation, branching, and cyclobutanation reactions, and with 3 mM DMAPP
(irregular conditions), which only gave products from cyclopropanation, branching, and
cyclobutanation reactions. The structures and relative amounts of products were
determined by hydrolysis of the diphosphates with alkaline phosphatase, extraction of the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 41
1
2
3
4
5
6
7
8
9
corresponding alcohols with methyl tert-butyl ether (MTBE), and analysis by GC on HP-5 or
Chiral-HP GC columns and by GC/MS. Structures were established by comparing GC
retention times and mass spectra with authentic samples (Figure 4).6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chain Elongation
PPO
PPO
Neryl Diphosphate
(NPP)
Geranyl Diphosphate
(GPP)
PPO
OH
PPO
Hydroxygeranyl Diphosphate
(hoGPP)
isoGeranyl Diphosphate
(iGPP)
PPO
PPO
(E,E)-Farnesyl Diphosphate
(E,Z-FPP)
(Z,E)-Farnesyl Diphosphate
(Z,E-FPP)
PPO
OH
Hydroxyfarnesyl Diphosphate
(hoFPP)
Cyclopropanation, Branching, Cyclobutanation
H
H
PPO
PPO
H
PPO
H
(1R,3R)-Chrysanthemyl
Diphosphate
(R)-Lavandulyl
Diphosphate
(R,R-LPP)
(R)-Maconelliyl
Diphosphate
(R,R-MPP)
(R,R-CPP)
H
PPO
(1R,3R)-Planococcyl
Diphosphate
(R,R-PPP)
Figure 4. Products from FPPase, CPPase, and FPPase/CPPase chimeras and site-directed
mutants.
ACS Paragon Plus Environment

Page 15 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Under standard chain elongation conditions, FPPase gave E,E-FPP as the major
product, accompanied by a smaller amounts of GPP and the Z,E-FPP (Table 1).14 Since
formation of FPP requires two molecules of IPP and equal concentrations of IPP and
DMAPP were used in the incubations, at the point where IPP was exhausted, only ~55% of
the available DMAPP had been consumed to give an 11:1 ratio of E,E-FPP: GPP. The
preponderance of E,E-FPP reflects substantial preference for utilization of GPP over
DMAPP for chain elongation.35
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Relative percentages of products formed by FPPase, CPPase, chimeras and site-
directed mutants under chain elongation incubation conditions (0.5 mM DMAPP and IPP)
(GPP, geranyl diphosphate; NPP, neryl diphosphate; iGPP, isogeranyl diphosphate; hoGPP,
hydroxygeranyl diphosphate; E,E-FPP, E,E-farnesyl diphosphate; Z,E-FPP, Z,E-farnesyl
diphosphate; hoFPP, hydroxyfarnesyl diphosphate; CPP, chrysanthemyl diphosphate; LPP,
lavandulyl diphosphate).
hoFPP
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
GPP
NPP iGPP hoGPP E,E-FPP Z,E-FPP
CPP
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
3 ± 1
nd
nd
nd
nd
LPP
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
8 ± 1
nd
nd
nd
nd
FPPase 8 ± 1
<1
<1
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
88 ± 1
87 ± 2
76 ± 1
90 ± 1
81 ± 1
75 ± 3
47 ± 6
38 ± 2
31 ± 2
3 ± 2
4 ± 1
4 ± 1
3 ± 1
2 ± 1
3 ± 1
2 ± 1
<1
C1
C2
9 ± 2
21 ± 1 <1
7 ± 1 <1
16 ± 1 <1
C3
C4
C5
22 ± 3 1 ± 1 nd
51 ± 6 2 ± 1 nd
60 ± 2 2 ± 1 nd
66 ± 2 2 ± 1 nd
92 ± 1 3 ± 1 nd
76 ± 2 2 ± 1 nd
70 ± 4 2 ± 1 nd
85 ± 1 3 ± 1 nd
63 ± 1 2 ± 1 nd
C6
C7
<1
C8
nd
C9
nd
C10
C11
C12
C13
21 ± 2
27 ± 4
10 ± 1
22 ± 1
70 ± 2
85 ± 2
29 ± 1
84 ± 2
nd
nd
nd
nd
A161I 21 ± 2 2 ± 1 nd
7 ± 1
8 ± 1
<1
T194G 7 ± 1
F231Y 71 ± 2 <1
D235N 9 ± 1 <1
<1
nd
nd
nd
7 ± 1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 41
1
2
3
4
5
6
7
8
9
nd
nd
nd
nd
nd
nd
2m
3m
71 ± 2 nd
62 ± 3 nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
29 ± 1
29 ± 1
24 ± 1
4 ± 3
nd
9 ± 1
2 ± 1
<1
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
5m
69 ± 3 5 ± 1 nd
93 ± 3 3 ± 1 nd
6m
c-F-c
12 ± 3 nd
nd
88 ± 3
2 ± 1
nd
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CPPase 47 ± 1 3 ± 1 2 ± 1
nd
35 ± 1 11 ± 1
18 ± 1
11 ± 1
F1
F2
65 ± 1 2 ± 1 <1
93 ± 1 2± 1 <1
4 ± 1
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
<1
nd
4 ± 1
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
<1
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
4 ± 1
nd
F3
F4
F5
F6
81 ± 6 2 ± 1 <1
76 ± 2 2 ± 1 <1
65 ± 2 1 ± 1 nd
72 ± 3 1 ± 1 nd
55 ± 6 1 ± 1 nd
19 ± 6
24 ± 2
33 ± 2
28 ± 3
43 ± 4
81 ± 2
89 ± 2
72 ± 1
60 ± 4
46 ± 4
76 ± 5
nd
<1
<1
1 ± 1
<1
1 ± 1
3 ± 1
3 ± 1
2 ± 1
2 ± 1
<1
F7
F8
14 ± 2 <1
8 ± 2 nd
nd
nd
nd
nd
F9
F10
F11
F12
F13
26 ± 1 <1
38 ± 4 <1
53 ± 4 1 ± 1 nd
21 ± 5 <1 nd
3 ± 1
nd
N235D 63 ± 3 4 ± 1 7 ± 1 22 ± 1
3m
70 ± 2 5 ± 1 6 ± 1 17 ± 1
68 ± 2 2 ± 1 nd nd
2 ± 1
<1
f-C-f
nd
nd
1 ± 0.3 25 ± 1
nd, not detected.
As the N- to C-terminal metamorphorsis of FPPase into CPPase proceeded, FPP was
the major product until the C6 chimera, where helix V was replaced and ~equal amounts of
GPP and FPP were formed (Figure 5A). Small to trace amounts of NPP and Z,E-FPP were
observed for the chimeras that produced substantial amounts of GPP and FPP, respectively.
From the C6 chimera on, the amount of GPP was similar to or greater than FPP, reaching a
maximum of 92% for C9. No irregular products were observed under chain elongation
conditions until the metamorphosis had proceeded to C13, where all helices forming the
active site bundle in FPPase had been replaced, at which point CPP and LPP constituted
ACS Paragon Plus Environment

Page 17 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
~11% (~1:3 CPP:LPP) of the product mixture. Upon completion of the metamorphosis to
CPPase, approximately equal amounts of chain elongation (mostly GPP) and irregular
(~3:1 CPP:LPP) products were seen.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
Figure 5. Products formed under chain elongation conditions during morphogenesis of
FPPase to CPPase and of CPPase to FPPase. GPP, gold; iGPP, light blue; NPP, purple; E,E-
FPP, dark blue; Z,E-FPP, pink; CPP, red; LPP, green. Part A, FPPase to CPPase; Part B,
CPPase to FPPase.
The loss of activity for synthesis of irregular products was dramatic during the
metamorphosis of CPPase to FPPase. Production of CPP and LPP plummeted upon
replacement of helix III in F2 to give an enzyme that predominantly formed GPP. The
remaining metamorphoses produced enzymes that mostly gave increasing amounts of FPP.
Production of FPP increased steadily between F2 and F9, where it reached 90%, and
decreased to 46% at F12 before rising again (Figure 5B).
The influence of the scaffolds surrounding the active-site helical bundles in FPPase
and CPPase was evaluated with core-scaffold chimeras c-F-c and f-C-f. Synthesis of FPP
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 41
1
2
3
4
5
6
7
8
9
was predominant for F13, C1, and the core-scaffold c-F-c chimera. In contrast, more diverse
distributions were seen for C13, F1, and f-C-f. FPP, which constituted 23% of the product
mixture for C13, was detected at 4% for F1 and not detected for the f-C-f chimera.
Furthermore, the CPP:LPP ratios were ~1:3 for C13, ~2:1 for F1, and ~1:25 for f-C-f.
These core-scaffold chimeras highlight the ability of amino acid substitutions outside of the
active site and even outside the α terpenoid fold, to alter product distributions for the
promiscuous CPPase while minimally effecting those of the more robust FPPase.
Single replacements of active site residues in FPPase with the corresponding amino
acids in CPPase did not substantially alter the product distributions, with the exception of
the F231Y mutation, which almost reversed the relative amounts of GPP and FPP. None of
the FPPase mutants with multiple active site replacements gave irregular products in
incubations with IPP and DMAPP. In sharp contrast, a single N235D mutation in CPPase
substantially reduced the production of irregular products by reducing synthesis of CPP
and FPP, and instead gave GPP as the major product, along with 30% of a 1:3 mixture of
iGPP and hoGPP. Similar distributions of C10 products were seen for 3mCPPase, along with
small amounts of both double bond isomers of FPP and hoFPP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Only products with irregular carbon skeletons were formed when the enzymes were
incubated with 3 mM DMAPP (irregular conditions). No products were seen for FPPase, C2,
C9, F2-F7, F11-F13, and c-F-c. LPP (branching) and MPP (cyclobutanation) were the only
products detected for chimeras C1, C3, C4, C8, C10, C11, and F10, as well as 5mFPPase and
6mFPPase (Table 2). The C5-C7 chimeras also produced trace amounts of planococcyl
diphosphate (PPP), a double bond isomer of MPP. CPP was first observed as a minor
product for C12 during the metamorphosis of FPPase to CPPase, and only became
ACS Paragon Plus Environment

Page 19 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
dominant for CPPase. C12, C13, F1, F8, F9, f-C-f, CPPase, the N235D CPPase mutant, and
3mCPPase synthesized compounds with all three of the irregular carbon skeletons.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Relative percentages of products formed by the irregular coupling of two
molecules of DMAPP.
CPP
nd
LPP
MPP
nd
PPP
nd
FPPase
C1
nd
nd
66 ± 8
63 ± 5
48 ± 1
47 ± 2
58 ± 1
57 ± 2
56 ± 4
57 ± 1
94 ± 1
95 ± 1
72 ± 1
63 ± 2
89 ± 2
20 ± 1
35 ± 1
86 ± 1
86 ± 1
83 ± 1
79 ± 1
80 ± 1
83 ± 1
34 ± 8
37 ± 5
52 ± 1
50 ± 1
39 ± 1
40 ± 2
44 ± 4
43 ± 1
6 ± 1
nd
C3
nd
nd
C4
nd
nd
C5
nd
3 ± 1
3 ± 1
3 ± 1
nd
C6
nd
C7
nd
C8
nd
C10
C11
C12
C13
5m
nd
nd
nd
nd
2 ± 0.3
27 ± 0.5
nd
3 ± 1
nd
1 ± 1
nd
37 ± 2
11 ± 1
1 ± 1
nd
6m
nd
nd
CPPase
F1
79 ± 1
64 ± 1
5 ± 1
4 ± 1
nd
nd
1 ± 1
nd
F8
9 ± 1
nd
F9
10 ± 1
17 ± 1
2 ± 1
nd
F10
N235D
3m
nd
19 ± 1
2 ± 1
16 ± 0.7
nd
17 ± 1
1 ± 1
nd
f-C-f
nd
nd, not detected.
Kinetic studies. Apparent Michaelis-Menten constants were determined from single-
point measurements of initial rates for product formation. [β-32P]IPP and [β-32P]DMAPP
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 41
1
2
3
4
5
6
7
8
9
were synthesized from the corresponding monophosphate esters by incubation with [γ-
32P]ATP and isopentenyl phosphate kinase.36 After quenching, assay mixtures were spotted
on silica TLC plates and the developed plates analyzed by phosphorimaging. Radioactivity
only appeared in products from chain elongation for incubations with [β-32P]IPP and
DMAPP. For chain elongation, values for kcat and KM were obtained by fits to the equations
for a sequential bisubstrate binding mechanism.37 The equations for the coupling of two
molecule of DMAPP have squared terms for the substrate in the numerator and
denominator as the result of DMAPP binding at both the donor and acceptor sites.
Consequently, we could not obtain Michaelis constants for DMAPP binding at individual
donor and acceptor sites by varying [DMAPP] and instead report K50DMAPP, the
concentration of DMAPP that gives half-maximal velocity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The kinetic constants for chain elongation are presented in Table 3. We were not
able to measure values for chimeras C9 and F2-F7 because of a combination of low
activities and poor levels of expression. The catalytic efficiency for chain elongation of GPP
to FPP (CEGPP, defined by kcatGPP/KMGPP = 2.4 x 106 s-1M-1) by sagebrush FPPase was similar
to CEGPP for the avian enzyme and approximately 100-fold slower than the diffusion
controlled limit for substrate binding (Table 4).35 CEDMAPP for chain elongation of DMAPP to
GPP was ~20-fold lower. This difference favors synthesis of FPP and minimizes
accumulation of GPP during chain elongation from DMAPP to FPP. The values for CEDMAPP
and CEGPP dropped during morphogenesis of FPPase into CPPase and shifted from CEDMAPP
< CEGPP to CEDMAPP > CEGPP, with a concomitant preference for synthesis of GPP. From
FPPase to C3, CEDMAPP decreased 31-fold, while CEGPP decreased 393-fold. From C3 to C10,
CEDMAPP was similar, while CEGPP continued to decline (103-fold, 106-fold relative to
ACS Paragon Plus Environment

Page 21 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
FPPase). Beginning at C11, CEDMAPP dropped substantially, followed by smaller decreases to
a CEDMAPP for CPPase that was 104-fold less than for FPPase. Values for CEGPP cannot be
measured for these enzymes because of a combination of low kcats and high KMDMAPPs.
During morphogenesis of CPPase to FPPase, activity for chain elongation remained low,
except for a spike at F10 from lower values for KMDMAPP and KMGPP (also observed at F1), for
which we have no explanation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Kinetic constants for chain elongation.a
DMAPP
GPP
DMAPP
GPP
IPP
kcat
(s-1)
kcat
(s-1)
KM
KM
KM
(μM)
(μM)
1.1 ± 0.3
3.5 ± 0.9
38 ± 7
36 ± 7
26 ± 4
16 ± 3
25 ± 5
240 ± 50
680 ± 70
4000 ± 800
nd
(μM)
11 ± 3
FPPase
C1
0.96 ± 0.09
1.2 ± 0.1
2.6 ± 0.1
1.5 ± 0.1
9 ± 2
4 ± 1
12 ± 5
C2
0.80 ± 0.07
0.81 ± 0.05
0.22 ± 0.01
0.26 ± 0.01
0.23 ± 0.01
0.42 ±0.03
0.147 ± 0.009
(7.7 ± 0.3) x 10-3
0.010 ± 0.001
nd
74 ± 6
60 ± 20
23 ± 7
C3
0.63 ± 0.06
200 ± 20
330 ± 50
110 ± 20
29 ± 3
C4
1.0 ± 0.1
34 ± 9
C5
0.61 ± 0.04
14 ± 3
C6
0.23 ± 0.02
8 ± 2
C7
0.188 ± 0.008
0.132 ± 0.006
0.64 ± 0.03
140 ± 20
18 ± 2
11 ± 2
C8
14 ± 2
C10
C11
C12
C13
120 ± 20
2000 ± 300
2300 ± 300
350 ± 50
59 ± 10
1300 ± 400
1400 ± 400
1600 ± 400
(7.0 ± 0.6) x 10-2
0.110 ± 0.008
(1.11 ± 0.03) x 10-2
nd
nd
nd
nd
c-F-c
CPPase
F1
(1.2 ± 0.3) x 10-2
(5.1 ± 0.2) x 10-3
(1.25 ± 0.09) x 10-3
(2.1 ± 0.1) x 10-3
(9.1 ± 0.4) x 10-3
(6.7 ± 0.4) x 10-3
(2.7 ± 0.1) x 10-3
(1.77 ± 0.06) x 10-3 3.1 x 10-3 (82 μM)b
(3.7 ± 0.2) x 10-3
(8.7 ± 0.3) x 10-3
nd
290 ± 50
510 ± 90
150 ± 30
nd
nd
80 ± 20
1700 ± 200
18 ± 2
nd
(3.2 ± 0.1) x 10-4
3.2 x 10-3 (247 μM)b 1000 ± 300
8.9 x 10-3 (247 μM)b 1000 ± 200
(7.1 ± 0.4) x 10-3
1.7 x 10-3 (247 μM)b 1100 ± 300
6.1 ± 0.9
nd
F8
1000 ± 200
2000 ± 300
11 ± 2
F9
nd
F10
F11
F12
F13
f-C-f
14 ± 4
7 ± 2
nd
1600 ± 300
1200 ± 100
2800 ± 400
3300 ± 300
440 ± 80
400 ± 70
530 ± 70
nd
3.0 x 10-3 (247 μM)b
9.7 x 10-3 (247 μM)b
nd
nd
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 41
1
2
3
4
5
6
7
8
9
a[DMAPP] and [IPP] were optimized to give maximal values for kcat. Apparent kinetic constants were
determined at a fixed optimal [substrate 1] for varied [substrate 2]. nd- not determined. bValues for KM
were not determined because of low turnover, large differences in optimal concentrations of IPP and GPP and
substrate inhibition at high [IPP]. kcatGPP is an apparent value at optimal [GPP] and 6 mM IPP. Where no
values for KMGPP are listed, kcat is an apparent value at the indicated [GPP].
GPP
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The catalytic efficiencies for synthesis of CPP and LPP from two molecules of
DMAPP (kcatDMAPP/K50DMAPP) and for synthesis of GPP from IPP and DMAPP
(kcatDMAPP/KMDMAPP) by CPPase were ~10,000-fold lower than for synthesis of GPP by
FPPase (Tables 4 and 5). During morphogenesis of CPPase to FPPase, the rate for synthesis
of the irregular products dropped 10-fold for the F1 chimera and 100- to 1000-fold for the
remaining chimeras in the series. Irregular activity was not detected for FPPase. For the
FPPase to CPPase series, a low level of irregular activity was first detected for C1 and
remained low through C12, where kcatDMAPP was ~20-50-fold was lower than that for wt
CPPase. For these chimeras, catalytic efficiencies for the irregular couplings were
substantially less than for chain elongation and the major irregular products from
branching (LPP) and cyclobutanation (MPP) were typically not detected in competition
with chain elongation. At C13, kcatDMAPP increased to a level similar to CPPase. However,
LPP was the major irregular product and the catalytic efficiency for chain elongation was
~6-fold greater than that for irregular coupling. In comparison, CPP was the dominant
irregular product for CPPase, and the catalytic efficiencies for irregular coupling and chain
elongation were similar.
ACS Paragon Plus Environment

Page 23 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
Table 4. Catalytic efficiencies for chain elongation.
DMAPP
GPP
kcatDMAPP /KM
kcatGPP/KM
9
(s-1 M-1)
(s-1 M-1)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FPPase
C1
1.0 x 105
3.0 x 105
1.1 × 104
3.2 × 103
3.0 × 103
5.5 × 103
7.9 × 103
1.34 × 103
7.3 × 103
5.3 × 103
35
2.4 x 106
4.3 x 105
2.1 × 104
6.1 × 103
1.0 × 104
1.4 × 104
1.7 × 104
6.1 × 102
C2
C3
C4
C5
C6
C7
C8
11
C10
C11
C12
C13
f-C-f
CPPase
F1
2.5
-
48
-
32
-
41
-
10
-
8.3
52
F8
2.1
-
F9
9.1
4.8 × 102
-
F10
F11
F12
F13
c-F-c
1 × 103
2.5
-
-
-
-
4
9.6
16
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 41
1
2
3
4
5
6
7
8
Table 5. Kinetic constants for formation of irregular products.
DMAPP
DMAPP
K50
DMAPP
kcat
(s-1)
kcat/K50
(s-1 M-1)
9
(μM)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C1
C3
3.7 × 10-4a
6.9 × 10-4a
5.0 × 10-4a
1.1 × 10-3a
6.6 × 10-4a
9.5 × 10-4a
4.0 × 10-4a
9.6 × 10-4a
8.6 × 10-4a
5.9 × 10-4a
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
-
-
-
-
-
-
-
-
-
-
C4
C5
C6
C7
C8
C10
C11
C12
C13
f_C_f
(1.1 ± 0.05) × 10-2 2100 ± 200
1.8 × 10-3a
nd
5.2
-
CPPase (2.0 ± 0.08) × 10-2 4100 ± 400
4.8
F1
F8
1.9 × 10-3a
3.2 × 10-5a
3.0 × 10-4a
6.3 × 10-4a
nd
nd
nd
nd
-
-
-
-
F9
F10
nd, not determined. aMeasured at 6 mM DMAPP.
Structural studies. Models for A. tridentata FPPase·IPP·DMAPP, C13·IPP·DMAPP,
C13·IPP·DMAPP, CPPase·IPP·DMAPP, and CPPase·DMAPP·DMAPP were constructed from
X-ray structures of E. coli FPPase liganded with IPP and an unreactive thiolo analogue of
DMAPP38, 39 in a “closed” active conformation (Brookhaven PDB: 1RQI; 2.4 Å) and apo C13
(Brookhaven PDB: 4KK2; 2.2 Å). C13, which is the first chimera during the morphogenesis
of FPPase to CPPase where formation of irregular products becomes competitive with
chain elongation and the only protein in the Artemisia group whose crystal structure has
ACS Paragon Plus Environment

Page 25 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
been reported, was used as the base structure for the modelling, as described in Supporting
Information.
7
8
9
Six active-site amino acids F92, A161, T194, F231, D235, and R343 in FPPase are
replaced by I92, I161, G194, Y231, N235, and G343, respectively, in CPPase. The locations
of substrates and amino acids at positions 193, 194, 231, 235 and 343 in the A. tridentata
enzymes are shown in Figure 6 with IPP in the acceptor region of A. tridentata FPPase (part
A) and IPP (part B) or DMAPP (part C) in the acceptor region of A. tridentata CPPase.
DMAPP is located in the donor regions of the active sites. In each case, the structures are
oriented so carbon-2 in the substrate in the acceptor site is located directly behind carbon-
3. This orientation highlights the differences among the conformations of the hydrocarbon
moieties of isopentenyl and dimethylallyl bound in the acceptor site. For IPP, the carbon
atoms attached to carbon-3 are restricted to a common plane and carbon-1 located above
the plane. In A. tridentata FPPase·IPP·DMAPP, the dihedral angle between carbon-1 and the
methyl group (Θ) is 75°, which is similar to the dihedral angle seen in E. coli
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FPPase·IPP·DMAPP.29 For A. tridentata CPPase·IPP·DMAPP, Θ = 31°. In contrast to IPP, all
of the carbon atoms in the dimethylallyl moiety are restricted to a common plane (Θ = 0°).
Corresponding active site structures for C13 are essentially identical to those for CPPase.
In FPPase, F231 is located perpendicular to the plane of C2, C3, C4, and the methyl
group in IPP with its edge ~4 Å above the C3-C4 double bond in position to stabilize
developing positive charge in the IPP moiety during the electrophilic alkylation reaction.
One of the terminal R343 guanidinium nitrogens is located ~3.4 Å almost directly above
the center of the F231 aromatic ring. The other terminal nitrogen is located 3.9 Å from one
of the non-bonding P1 oxygen atoms in IPP. R343 seems to have two roles – to help
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 41
1
2
3
4
5
6
7
8
9
immobilize the diphosphate group in IPP and to lock the aromatic ring of F231 over the
double bond in IPP. D235 is part of the DDXXD motif that binds the diphosphate unit in
DMAPP. The carboxylate group in D235 is coordinated with one of the active site
magnesium atoms that bind DMAPP diphosphate and the two carboxylate oxygens are
about ~3.5 Å from the IPP methyl group. The side chain hydroxyl group in T194 is
hydrogen bonded to the amide carbonyl of G228 and the T194 methyl group is 4.8 Å from
the methylene carbon at C4 in IPP. The amino group in K193 helps bind the DMAPP
diphosphate groups in FPPase·IPP·DMAPP and CPPase·IPP·DMAPP. The non-conserved
amino acids at positions 92 and 161 in FPPase and CPPase are located in a hydrophobic
pocket that interacts with the terminal isoprene unit in GPP and do not make contact with
DMAPP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In CPPase, the plane of the aromatic ring in Y231 makes a ~50° angle with the plane
of the C1-C4 carbons in IPP and its edge is 3.6 Å above the IPP methyl group. The
diphosphate moiety in IPP is anchored to Y231 by a hydrogen bond between a non-
bonding P1 oxygen and the hydroxyl group in the tyrosine side chain. The amide group in
N235 is twisted relative to the position of the D235 carboxylate in FPPase and is located 3.1
Å from the IPP methyl group. The backbone chain of G194 is displaced toward IPP in
CPPase relative to the T194 backbone, and is located 4.8 Å from the IPP C4 methylene
carbon. Water molecules occupy similar locations in the active sites of FPPase and CPPase
where they coordinate with the three magnesium atoms that bind to the diphosphate in
DMAPP and form a shell around the diphosphate in IPP.
ACS Paragon Plus Environment

Page 27 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Active site structures for A. tridentata FPPase and CPPase. Part A:
FPPase·IPP·DMAPP - isopentenyl moiety (yellow), dimethylallyl moiety (raspberry), Mg2+
(olive); Part B: CPPase·IPP·DMAPP - isopentenyl moiety (pale green), dimethylallyl moiety
(salmon), Mg2+ (orange); Part C: CPPase·DMAPP·DMAPP – dimethylallyl moiety
(isopentenyl region, forest green), dimethylallyl moiety (firebrick red), Mg2+ (light orange).
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 41
1
2
3
4
5
6
7
8
9
Figure 7 (part A) shows active site overlays for A. tridentata FPPase·IPP·DMAPP and
A. tridentata CPPase·IPP·DMAPP and are arranged to superimpose the carbon-2-carbon-3
bonds of the substrates in the acceptor sites. The only substantial conformational changes
seen between the active site amino acids in FPPase and CPPase relative to the substrates
are at positions 194, 231, and 235. The side chains in F231 and Y231 are positioned to
stabilize the tertiary carbocationic intermediate formed during the alkylation step (see
Scheme 1). However, the Y231 aromatic ring in CPPase is displaced upward toward
phosphorus-1 in IPP relative to the F231 phenyl group in FPPase. The IPP methyl group is
located above the D235 carboxylate equidistant from both carboxylate oxygens. In contrast,
the amide group in N235 side chain amide of CPPase is displaced upward to within
hydrogen bonding distance of the bond-bridging oxygens of P2 in DMAPP. These changes
alter the topology of the binding pocket occupied for the methyl group in IPP and move C1
in IPP toward the plane of the four other carbon atoms in IPP. The T194 to G194 mutation
allows the binding pocket to accommodate methylene group as the dihedral angle in IPP
decreases from 75° in FPPase to 31° in CPPase. As shown in Figure 7 (part B), CPPase can
accommodate either IPP or DMAPP in the acceptor region of the active site by small
changes in the conformation of the diphosphate moiety and the tilt of the aromatic ring in
Y231.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 29 of 41
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Overlays of active sites of A. tridentata FPPase and CPPase. Part A:
FPPase·IPP·DMAPP – IPP (yellow), DMAPP (raspberry), Mg2+ (olive), and amino acids
(blue); CPPase·IPP·DMAPP – IPP (pale green), DMAPP (salmon), Mg2+ (orange), and amino
acids (sky blue). Part B: CPPase·IPP·DMAPP – IPP (pale green), DMAPP (salmon), Mg2+
(orange), and amino acids (sky blue); CPPase·DMAPP·DMAPP – DMAPP (acceptor, dark
green; donor, firebrick red), Mg2+ (light orange), and amino acids (light blue).
DISCUSSION
The α terpenoid synthase fold, found in all members of CL0613, most likely evolved
from a gene duplication-fusion event of a primordial 4-helix bundle protein.10 Presumably
an ancestral fusion protein, with hallmark DDXXD/DDXXD motifs on mirror image helices,
evolved the ability to bind an allylic diphosphate and catalyze the electrophilic alkylation
ACS Paragon Plus Environment