Substrate specificity and diastereoselectivity of strictosidine glucosidase, a key enzyme in monoterpene indole alkaloid biosynthesis

Article abstract of DOI:10.1016/j.bmcl.2007.11.063

Strictosidine glucosidase (SGD) from Catharanthus roseus catalyzes the deglycosylation of strictosidine, an intermediate from which thousands of monoterpene indole alkaloids are derived. The steady-state kinetics of SGD with a variety of strictosidine analogs revealed the substrate preferences of this enzyme at two key positions of the strictosidine substrate. Additionally, SGD from C. roseus turns over both strictosidine and its stereoisomer vincoside, indicating that although this enzyme prefers the naturally occurring diastereomer, the enzyme is not completely diastereoselective. The implications of the substrate specificity of SGD in metabolic engineering efforts of C. roseus are highlighted.

Full text of DOI:10.1016/j.bmcl.2007.11.063

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 3095–3098
Substrate specificity and diastereoselectivity
of strictosidine glucosidase, a key enzyme in monoterpene
indole alkaloid biosynthesis
Nancy Yerkes, Jia Xin Wu, Elizabeth McCoy, M. Carmen Galan,
Shi Chen and Sarah E. O’Connor*
Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA
Received 15 October 2007; revised 14 November 2007; accepted 15 November 2007
Available online 22 November 2007
Abstract—Strictosidine glucosidase (SGD) from Catharanthus roseus catalyzes the deglycosylation of strictosidine, an intermediate
from which thousands of monoterpene indole alkaloids are derived. The steady-state kinetics of SGD with a variety of strictosidine
analogs revealed the substrate preferences of this enzyme at two key positions of the strictosidine substrate. Additionally, SGD from
C. roseus turns over both strictosidine and its stereoisomer vincoside, indicating that although this enzyme prefers the naturally
occurring diastereomer, the enzyme is not completely diastereoselective. The implications of the substrate specificity of SGD in met-
abolic engineering efforts of C. roseus are highlighted.
Ó 2007 Elsevier Ltd. All rights reserved.
Monoterpene indole alkaloids (MIA) are a large class of
pharmaceutically valuable and structurally complex nat-
ural products.1 Directed biosynthesis studies have
shown that the MIA pathway can produce a variety of
‘unnatural’ alkaloids by utilizing non-natural substrate
analogs.2 This inherent flexibility suggests that MIA bio-
synthesis could provide a robust platform for metabolic
engineering. However, not all substrate analogs are
likely to be incorporated into the pathway with equal
efficiency. If substrate specificity of individual biosyn-
thetic enzymes correlates with rate limiting steps
in vivo, then enzymes having a low catalytic efficiency
for a non-natural substrate could be reengineered to im-
prove turnover of the analog.3 Therefore, evaluation of
enzyme substrate specificity is critical for biosynthetic
engineering efforts. In MIA biosynthesis, the central bio-
synthetic precursor strictosidine 1 is deglycosylated by
strictosidine glucosidase (SGD) to yield a reactive inter-
mediate that rearranges to form the wide variety of MIA
(Scheme 1).4 Here we evaluate the substrate specificity of
SGD from Catharanthus roseus with a variety of strict-
osidine analogs to determine whether SGD could act
as a bottleneck in the production of novel alkaloids
from unnatural strictosidine analogs.
SGD was assayed with strictosidine analogs 2–9 (Table
1).5–8 An HPLC assay was used to monitor both strict-
osidine disappearance and deglycosylated product for-
mation. Since all kinetic data appeared to fit a
sigmoidal rather than a Michaelis–Menten curve, kinetic
constants were obtained from a sigmoidal fit to the data
(Fig. 1).9,10 SGD from C. roseus has been reported to
form aggregates consisting of 4–12 monomers4a;
although a sigmoidal fit has not been previously re-
ported for C. roseus SGD, the oligomeric state of the en-
zyme is compatible with the cooperative mechanism
suggested by the sigmoidal curve.
The catalytic efficiencies (Vmax/K0.5) of indole-substituted
strictosidine derivatives 2–9varied byless than anorder of
magnitude from the naturally occurring strictosidine 1.
Strictosidine analogs with methyl groups in the 9, 10,
11, and 12 positions (compounds 2–5) demonstrated that
steric effects did not disrupt enzyme activity dramatically.
Not surprisingly, replacement of methyl groups with lar-
ger methoxy substituents (compounds 6, 7) resulted in a
small increase in the K0.5 values. Electronic perturbations
did not appear to impact the turnover profoundly; cata-
lytic efficiencies of fluorinated strictosidine analogs (8,
9) did not vary significantly from 1.
Keywords: Alkaloid; Biosynthesis; Strictosidine; Glucosidase.
*
Corresponding author. Tel.: +1 6173240180; e-mail: soc@mit.edu
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.063

3096
N. Yerkes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3095–3098
deglycosylation was observed when the SGD concentra-
CHO
tryptamine
NH2
tion was increased from approximately 0.6 nM to
50 lM. The ester appears to be positioned more deeply
in the interior of SGD compared to the indole moiety,11
and steric clashes of the pentynyl group with the sur-
rounding enzyme residues may impede effective cataly-
sis. Mutation of SGD may improve turnover of this
substrate.
O-Glc
secologanin
O
N
MeO2C
H
strictosidine
synthase
9
10
11
OH
OH
Notably, the slow turnover rate of 10 parallels the re-
sults of feeding studies in which C. roseus hairy root
cultures were grown in the presence of the precursor
to 10, the pentynyl ester of secologanin. In these feed-
ing studies,3a,12 a significant accumulation of pentynyl
strictosidine 10 was observed, a result consistent with
slow deglycosylation of 10 by SGD (Fig. 3, black
trace). Therefore, the enzyme substrate specificity can
predict bottlenecks in precursor directed biosynthesis.
Despite slow deglycosylation by SGD, 10 is neverthe-
less converted by C. roseus into several downstream
alkaloid products; one alkaloid (serpentine) is shown
in Figure 3 (red trace).12 The production levels of
these alkaloids could be expected to increase further
if a reengineered variant of SGD that more effectively
deglycosylates 10 is incorporated into the MIA
pathway.
O
O
NH
H
N
H
12
O
HO
OH
MeO2C
strictosidine 1
SGD
N
H
NH
H
O
OH
N
H
N
H
MeO2C
MeO2C
OH
examples of MIA
N
H
Strictosidine synthase, the enzyme that forms strictosi-
dine 1 from tryptamine and secologanin (Scheme 1), is
absolutely stereoselective. The vincoside diastereomer
11 is not formed enzymatically in any known MIA
pathway (Fig. 2).1 Although SGD from R. serpentina
does not turn over the vincoside diastereomer,4b we
observed that the C. roseus enzyme does in fact turn
over the ‘non-natural’ vincoside13 as evidenced by an
LC–MS assay and a glucose detection assay.14 Stea-
dy-state kinetic constants were not obtained for this
substrate because vincoside 11 spontaneously formed
the lactam 12 during the course of the assay, thereby
complicating kinetic analysis.15 However, the rate of
deglycosylation of 11 appeared to be qualitatively
slower than that of strictosidine 1. Inspection of the
SGD sequence from R. serpentina, which has 70% ami-
no acid identity with C. roseus glucosidase,4 did not re-
veal any obvious differences in the strictosidine binding
site,11 so the structural basis for the difference in dia-
stereoselectivity between the two enzymes remains to
be determined.
N
H
N
H
H
N
CO2Me
H
MeO2C
OH
Scheme 1. Strictosidine 1 is deglycosylated by strictosidine glucosidase
(SGD) to form a reactive intermediate that is the precursor for
thousands of monoterpene indole alkaloid products.
The recently reported crystal structure of SGD from
Rauwolfia serpentina (PDB Code 2JF6) indicates that
whereas the glucose moiety of 1 is buried within the en-
zyme active site, the indole portion of strictosidine
points toward the surface of the enzyme.11 The results
of these substrate specificity studies suggest that the ac-
tive site of SGD has not evolved to discriminate against
substitutions on the indole ring. Notably, 5 methyl and 6
methyl tryptamine are not turned over by the enzyme
strictosidine synthase (Scheme 1) to form strictosidine
analogs 3 and 4, respectively.5 The specificity of the
early stages of the MIA pathway therefore appears to
be controlled in large part by strictosidine synthase
and not by SGD. Any significant differences observed
in incorporation of indole-substituted tryptamine sub-
strates into alkaloid products are not likely to be due
to the substrate specificity of SGD.
SGD catalyzes deglycosylation of strictosidine analogs
containing indole substituents at relatively high cata-
lytic efficiencies. In contrast, a significant decrease in
catalysis was noted after the methyl ester of strictosi-
dine was replaced with a pentynyl ester (compound
10), a result consistent with accumulation of 10 in feed-
ing studies. Despite the slow turnover of 10 by SGD,
10 is nevertheless converted by C. roseus in vivo into
several downstream alkaloid products.12 We envision
that the levels of these alkaloids will increase if a reen-
gineered variant of SGD that more effectively degly-
cosylates 10 is incorporated into the MIA pathway.
Surprisingly, SGD from C. roseus appears to turnover
vincoside, the diastereomer of strictosidine, indicating
Whereas the modestly sized indole substituents had a
moderate effect on SGD catalysis, replacement of the
methyl ester of strictosidine with the significantly larger
pentynyl ester12 (compound 10, Fig. 2) resulted in a sig-
nificant decrease in turnover by SGD. No formation of
deglycosylated product was observed under the assay
conditions used for strictosidine analogs 2–9, although

N. Yerkes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3095–3098
3097
Table 1. Steady-state kinetic parameters for strictosidine analogs 1–9
Table 1 (continued)
Analogs
K0.5 Vmax Relative
(mM) (mM/min) Vmax/K0.5
Analogs
K0.5
Vmax
Relative
(mM) (mM/min) Vmax/K0.5
O-Glc
O-Glc
NH
NH
N
H
N
O
O
0.22 0.078
0.22 0.018
0.071 0.016
0.092 0.017
0.14 0.016
0.37 0.035
0.43 0.037
0.15 0.038
1
0.12
0.032
0.7
H
H
F
9
1
H
MeO
MeO
O
O
See Ref. 8 for a description of the enzyme assay.
Me
O-Glc
NH
0.2
0.6
0.5
0.3
0.3
0.2
0.7
O 2
N
H
H
MeO
O
Me
O-Glc
NH
N
H
O
H
3
MeO
O
O
O-Glc
NH
N
H
O
H
Me
4
MeO
Figure 1. Representative kinetic data for SGD. The kinetic data for all
substrates tested (substrate 4 shown above) fit a sigmoidal curve. The
equation used was y = A2 + (A1 À A2)/[1 + (x/xo)p]. A2 was used as the
Vmax value, and xo was used as K0.5. All substrates had a Hill
coefficient of 1.8 or greater.
O-Glc
NH
N
H
O
H
5
Me
MeO
O
NH
O-Glc
H
N
H
O
MeO
O-Glc
NH
O
10
N
H
O
O
H
6
O
MeO
NH
H
N
O
O
O-Glc
H
N
N
H
H
O
O
vincoside
lactam 12
O-Glc
NH
MeO2C
vincoside 11
O-Glc
N
H
O
H
MeO
7
Figure 2. Structures of additional analogs 10 and 11 turned over by
SGD. Compound 12 forms spontaneously from 11.
MeO
that the stereoselectivity of the MIA pathway is not
maintained by SGD.
F
O-Glc
NH
N
O
H
H
Acknowledgment
8
MeO
O
We gratefully acknowledge financial support from the
Beckman Foundation, GM074820 and MIT.

3098
N. Yerkes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3095–3098
6. Strictosidine glucosidase (SGD) was expressed in Esche-
richia coli as an N-terminal maltose-binding protein
(MBP) or C-terminal 6 His-tag fusion using a codon
optimized synthetic gene previously described in Ref. 5.
The MBP fusion was used for the determination of all
kinetic constants.
7. Synthesis of strictosidine analogs has been previously
described in Ref. 5. Briefly, strictosidine analogs 1, 2, 5, 8,
and 9 were synthesized enzymatically by incubating
strictosidine synthase with secologanin and the corre-
sponding tryptamine analog followed by purification via
preparative HPLC. Strictosidine analogs 3, 4, 6, and 7
were synthesized as diastereomeric mixtures chemically by
incubating secologanin and the corresponding tryptamine
analog in pH 2, 100 mM maleic acid. The strictosidine
diastereomers were purified by preparative HPLC. Exact
masses and representative NMR data of the purified
products are reported in Supplementary Material.
8. Quantitative SGD kinetic assays were performed at the
reported optimal pH of the enzyme (pH 6, citrate
phosphate buffer, see Ref. 4). Assays were conducted at
37 °C in the presence of 0.61 nM SGD. Nine or more
different substrate concentrations were used for each
analog, and six time points were measured for each
substrate concentration. An HPLC assay was used to
monitor strictosidine consumption at 238 nm with naph-
thalene acetic acid as an internal standard. The concen-
tration of all strictosidine analogs was determined with a
standard curve. A logistic curve using origin was used to
fit the data, and R2 values ranged from 0.979 to 0.999.
9. Hammes, G. G.; Wu, C.-W. Annu. Rev. Biophys. Bioeng.
1974, 3, 1.
Figure 3. LC–MS traces of extracts from C. roseus root culture
incubated with a pentynylated secologanin substrate analog. More
pentynyl strictosidine analog 10 (m/z 583, black trace) accumulates
relative to the final alkaloid analog product pentynyl serpentine (m/z
401, red trace). In contrast, less natural strictosidine 1 (m/z 531, green
trace) is observed relative to natural serpentine alkaloid (m/z 349, blue
trace). See Ref. 12 for detailed structural characterization of the
alkaloid products.
Supplementary data
Supplementary data contain exact mass data for sub-
strates and products, representative NMR data, and as-
say data for substrate 11. Supplementary data
associated with this article can be found, in the online
version, at doi:10.1016/j.bmcl.2007.11.063.
10. A sigmoidal curve was observed for SGD expressed as
both an N-terminal maltose-binding protein fusion and
as a C-terminal His-tag fusion indicating that a specific
affinity tag does not alter the kinetic parameters
significantly.
11. Barleben, L.; Panjikar, S.; Ruppert, M.; Koepke, J.;
Sto¨ckigt, J. Plant Cell 2007, 19, 2886.
References and notes
12. See Galan, M. C.; McCoy, E.; O’Connor, S. E. Chem.
Commun. 2007, 3249, Briefly, a pentynyl secologanin
derivative (500 lM) was incubated with a C. roseus hairy
root culture. After two weeks, alkaloids were extracted
from the cell cultures and analyzed by MS and NMR to
demonstrate that the pentynyl ester had been incorporated
into the MIA biosynthetic pathway.
1. (a) van der Heijden, R.; Jacobs, D. I.; Snoeijer, W.;
Hallard, D.; Verpoorte, R. Curr. Med. Chem. 2004, 11,
607; (b) O’Connor, S. E.; Maresh, J. J. Nat. Prod. Rep.
2006, 23, 532.
2. McCoy, E.; O’Connor, S. E. J. Am. Chem. Soc. 2006, 128,
14276.
3. Recent examples of enzyme engineering in the MIA pathway
are: (a) Chen, S.; Galan, M. C.; Coltharp, C.; O’Connor, S. E.
Chem. Biol. 2006, 13, 1137; (b) Bernhardt, P.; McCoy, E.;
O’Connor, S. E. Chem. Biol. 2007, 14, 888; (c) Loris, E. A.;
13. Vincoside and deuterated vincoside were synthesized by
incubating tryptamine or b,b-D tryptamine with secolog-
2
anin in pH 2, 100 mM maleic acid for 12 h at 37 °C
followed by purification via preparative HPLC. MS and
NMR data for this compound are shown in Supplemen-
tary Material.
Panjikar, S.; Ruppert, M.; Barleben, L.; Unger, M.; Schubel,
¨
H.; Sto¨ckigt, J. Chem. Biol. 2007, 14, 979.
14. To further validate the activity of SGD, a glucose
detection reagent was used to validate that glucose was
produced in the presence of SGD. Zhou, M.; Diwu, Z.;
Panchuk-Voloshina, N.; Haugland, R. Anal. Biochem.
1997, 253, 162.
4. (a) Geerlings, A.; Iban˜ez, M. M.-L.; Memelink, J.; van der
Heijden, R.; Verpoorte, R. J. Biol. Chem. 2000, 275, 3051;
(b) Gerasimenko, I.; Sheludko, Y.; Ma, X.; Sto¨ckigt, J.
Eur. J. Biochem. 2002, 269, 2204.
5. McCoy, E.; Galan, M. C.; O’Connor, S. E. Bioorg. Med.
Chem. Lett. 2006, 16, 2475.
´
´
´
´ ´
15. Patthy-Lukats, A.; Kocsis, A.; Szabo, L.; Podanyi, B. J.
Nat. Prod. 1999, 62, 1492.