Strictosidine synthase: Mechanism of a Pictet-Spengler catalyzing enzyme

Article abstract of DOI:10.1021/ja077190z

The Pictet-Spengler reaction, which yields either a β-carboline or a tetrahydroquinoline product from an aromatic amine and an aldehyde, is widely utilized in plant alkaloid biosynthesis. Here we deconvolute the role that the biosynthetic enzyme strictosidine synthase plays in catalyzing the stereoselective synthesis of a β-carboline product. Notably, the rate-controlling step of the enzyme mechanism, as identified by the appearance of a primary kinetic isotope effect (KIE), is the rearomatization of a positively charged intermediate. The KIE of a nonenzymatic Pictet-Spengler reaction indicates that rearomatization is also rate-controlling in solution, suggesting that the enzyme does not significantly change the mechanism of the reaction. Additionally, the pH dependence of the solution and enzymatic reactions provides evidence for a sequence of acid-base catalysis steps that catalyze the Pictet-Spengler reaction. An additional acid-catalyzed step, most likely protonation of a carbinolamine intermediate, is also significantly rate controlling. We propose that this step is efficiently catalyzed by the enzyme. Structural analysis of a bisubstrate inhibitor bound to the enzyme suggests that the active site is exquisitely tuned to correctly orient the iminium intermediate for productive cyclization to form the diastereoselective product. Furthermore, ab initio calculations suggest the structures of possible productive transition states involved in the mechanism. Importantly, these calculations suggest that a spiroindolenine intermediate, often invoked in the Pictet-Spengler mechanism, does not occur. A detailed mechanism for enzymatic catalysis of the β-carboline product is proposed from these data.

Full text of DOI:10.1021/ja077190z

Published on Web 12/15/2007
Strictosidine Synthase: Mechanism of a Pictet-Spengler
†
Catalyzing Enzyme
†
†
†
‡
Justin J. Maresh, Lesley-Ann Giddings, Anne Friedrich, Elke A. Loris,
§
|
,‡
,⊥
Santosh Panjikar, Bernhardt L. Trout, Joachim St o¨ ckigt,* Baron Peters,* and
,
†
Sarah E. O’Connor*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts
0
2139, College of Pharmaceutical Sciences, Zhejiang UniVersity, Zijingang Campus, 310058
Hangzhou, China, Lehrstuhl f u¨ r Pharmazeutische Biologie, Institut f u¨ r Pharmazie,
Staudingerweg 5, 55099 Mainz, Germany, European Molecular Biology Laboratory, Hamburg
outstation DESY, Notkestrasse 85, 22603 Hamburg, Germany, Department of Chemical
Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and
Department of Chemical Engineering, UniVersity of California, Santa Barbara,
Santa Barbara, California 93106
Received September 17, 2007; E-mail: .
Abstract: The Pictet-Spengler reaction, which yields either a â-carboline or a tetrahydroquinoline product
from an aromatic amine and an aldehyde, is widely utilized in plant alkaloid biosynthesis. Here we
deconvolute the role that the biosynthetic enzyme strictosidine synthase plays in catalyzing the stereose-
lective synthesis of a â-carboline product. Notably, the rate-controlling step of the enzyme mechanism, as
identified by the appearance of a primary kinetic isotope effect (KIE), is the rearomatization of a positively
charged intermediate. The KIE of a nonenzymatic Pictet-Spengler reaction indicates that rearomatization
is also rate-controlling in solution, suggesting that the enzyme does not significantly change the mechanism
of the reaction. Additionally, the pH dependence of the solution and enzymatic reactions provides evidence
for a sequence of acid-base catalysis steps that catalyze the Pictet-Spengler reaction. An additional
acid-catalyzed step, most likely protonation of a carbinolamine intermediate, is also significantly rate
controlling. We propose that this step is efficiently catalyzed by the enzyme. Structural analysis of a
bisubstrate inhibitor bound to the enzyme suggests that the active site is exquisitely tuned to correctly
orient the iminium intermediate for productive cyclization to form the diastereoselective product. Furthermore,
ab initio calculations suggest the structures of possible productive transition states involved in the
mechanism. Importantly, these calculations suggest that a spiroindolenine intermediate, often invoked in
the Pictet-Spengler mechanism, does not occur. A detailed mechanism for enzymatic catalysis of the
â-carboline product is proposed from these data.
1
. Introduction
monoterpene indole alkaloids. These alkaloids are a large and
structurally diverse family of natural products that include
vinblastine, yohimbine, ajmaline, camptothecin, and strychnine
and are found in a wide variety of plant species. Strictosidine
synthase diastereoselectively converts the substrates tryptamine
1a and secologanin 2 to the â-carboline product strictosidine 3
The Pictet-Spengler reaction yields â-carboline and tetrahy-
droquinoline structures essential to the biosynthesis of thousands
of plant natural products. Enzymes that catalyze this reaction
have been isolated from several plant alkaloid biosynthetic
6,7
1
-3
4,5
pathways.
Strictosidine synthase,
one “Pictet-Spen-
(Figure 1). Strictosidine 3 serves as the biosynthetic precursor
glerase,” is the central enzyme in the biosynthesis of the
to all terpene indole alkaloids.
†
Department of Chemistry, Massachusetts Institute of Technology.
Institut f u¨ r Pharmazie and Zhejiang University.
European Molecular Biology Laboratory.
Department of Chemical Engineering, Massachusetts Institute of
The Pictet-Spengler reaction is essentially a two-part reac-
‡
8
,9
§
tion. First, an electron-rich aromatic amine and an aldehyde
condense to form an iminium species (Figure 1, steps 1-3).
Second, an electrophilic aromatic substitution reaction occurs
in which the aryl amine attacks the electrophilic iminium to
yield a positively charged intermediate (Figure 1, step 4) which
is then deprotonated to yield the â-carboline product(s). In
|
Technology.
⊥
University of California, Santa Barbara.
(
(
(
(
(
1) De-Eknamkul, W.; Suttipanta, N.; Kutchan, T. M. Phytochemistry 2000,
5
5, 177-181.
2) Samanani, N.; Liscombe, D. K.; Facchini, P. J. Plant J. 2004, 40, 302-
13.
3) Minami, H.; Dubouzet, E.; Iwasa, K.; Sato, F. J. Biol. Chem. 2007, 282,
3
6
274-6282.
(6) Van der Heijden, R.; Jacobs, D. I.; Snoeijer, W.; Hallard, D.; Verpoorte,
R. Curr. Med. Chem. 2004, 11, 607-628.
(7) O’Connor, S. E.; Maresh, J. Nat. Prod. Rep. 2006, 23, 532-547.
(8) Cox, E. D.; Cook, J. M. Chem. ReV. 1995, 95, 1797-1842.
(9) Pictet, A.; Spengler, T. Ber. Dtsch. Chem. Ges. 1911, 44, 2030-2036.
4) McKnight, T. D.; Roessner, C. A.; Devagupta, R.; Scott, A. I.; Nessler, C.
Nucleic Acids Res. 1990, 18, 4939.
5) Kutchan, T. M.; Hampp, N.; Lottspeich, F.; Beyreuther, K.; Zenk, M. H.
FEBS Lett. 1988, 237, 40-44.
710
9
J. AM. CHEM. SOC. 2008, 130, 710-723
10.1021/ja077190z CCC: $40.75 © 2008 American Chemical Society

Mechanism of Strictosidine Synthase
A R T I C L E S
Figure 1. Pictet-Spengler reaction mechanism. Substrates used in this study are shown.
nonenzymatically catalyzed reactions, two enantiomers are
typically formed, but strictosidine synthase catalyzes the asym-
metric synthesis of the strictosidine 3 diastereomer (asterisk in
Figure 1).
Notably, both carbons 2 and 3 of tryptamine are nucleophilic
(Figure 1). Therefore, after iminium formation, Pictet-Spengler
reactions that utilize indole amine substrates can proceed either
by attack of carbon 2 to directly yield the six-membered ring
intermediate (Figure 1, step 4) or by attack of carbon 3 to yield
a spiroindolenine intermediate (Figure 1, step 4a) that would
then undergo a 1,2-alkyl shift (Figure 1, step 4b) to form the
product. Evidence for both mechanisms in solution exists, and
Figure 2. Structure of strictosidine synthase active site, highlighting
ionizable amino acids surrounding the tryptamine and secologanin. Crystal
structure of enzyme complex with tryptamine (2FPB, 2.8 Å resolution) is
overlaid with structure of secologanin (from 2FPC, 3.0 Å resolution). The
amine of tryptamine is highlighted with a blue arrow, and the aldehyde of
secologanin is highlighted with a red arrow. Surrounding polar residues
are shown in green. Distances from the amine of tryptamine to the
surrounding amino acid side chains are shown. The C2 carbon of tryptamine
is highlighted with a green arrow, and the distance from the carboxyl group
of Glu309 to this carbon is also shown (3.2 Å, green line).
1
0-13
the predominant mechanism is not entirely clear.
mechanism for the enzymatic reaction is not known.
The
Strictosidine synthase (RauVolfia serpentina) has been coc-
rystallized in the presence of both secologanin (PDB code 2FPC)
and tryptamine (PDB code 2FPB),14 and these structures have
allowed the first insights into the substrate binding orientation
and enzymatic mechanism. Strictosidine synthase, originally
isolated from Catharanthus roseus and R. serpentina almost
laboratory match literature values.19,20 The recently reported
strictosidine synthase crystal structure revealed the presence of
only three ionizable residues in the active site, Tyr151, His307,
and Glu309 (R. serpentina numbering), that were located near
the amine of tryptamine 1 and the aldehyde of secologanin 2
15,16
30 years ago,
has been the subject of numerous steady-state
kinetic analyses (for two examples, see refs 17,18). Km values
have been reported for both tryptamine 1 (4 µM) and secolo-
ganin 2 (40 µM), and values previously reported from our
1
4
(
10) Bailey, P. D. J. Chem. Res., Synop. 1987, 202-203.
(Figure 2). His307 appears to be involved in binding to the
glucose moiety of secologanin as evidenced by the crystal
structure and the large increase in secologanin Km after
(
11) Ibaceta-Lizana, J. S.; Jackson, A. H.; Prasitpan, N.; Shannon, P. V. R. J.
Chem. Soc., Perkin Trans. II 1987, 1221-1226.
(
(
(
12) Kowalski, P.; Mokrosz, J. L. Bull. Soc. Chim. Belg. 1997, 106, 147-149.
13) Ungemach, F.; Cook, J. M. Heterocycles 1978, 9, 1089-1119.
14) Ma, X.; Panjikar, S.; Koepke, J.; Loris, E.; Stockigt, J. Plant Cell 2006,
14
mutation. Site-directed mutagenesis of Tyr151 suggests that
1
8, 907-920.
the ionizable hydroxyl group does not play an important role
(
15) Mizukami, H.; Nordlov, H.; Lee, S.-L.; Scott, A. I. Biochemistry 1979,
14
in catalysis. Site-directed mutagenesis experiments do support
1
8, 3760-3763.
(
(
(
16) Treimer, J. F.; Zenk, M. H. Eur. J. Biochem. 1979, 101, 225-233.
17) Pfitzner, U.; Zenk, M. H. Planta Med. 1989, 55, 525-530.
(19) Bernhardt, P.; O’Connor, S. E. Chem. Biol. 2007, 14, 888-897.
(20) McCoy, E.; Galan, M. C.; O’Connor, S. E. Bioorg. Med. Chem. Lett. 2006,
16, 2475-2478.
18) de Waal, A.; Meijer, A. H.; Verpoorte, R. Biochem. J. 1995, 306, 571-
5
80.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 711

A R T I C L E S
Maresh et al.
the involvement of glutamate residue (Glu309, R. serpentina
numbering) in catalysis (900-fold drop in Vmax for Glu309Ala).
(1H, s), 7.21-7.26 (1H, m), 7.29-7.35 (1H, m), 7.41-7.29 (2H, m),
7.50-7.55 (2H, m), 7.89 (1.33H, dd, J ) 1.4 Hz, 7.3 partial deuteration
of phenyl), 8.01 (1H, dd, 0.72, 8.28). Relative to the commercial starting
14,19
This structural information provides an excellent foundation
for further mechanistic study of the Pictet-Spengler reaction.
Here we report specific roles for acid-base catalysis in
strictosidine synthase using data derived from kinetic isotope
effects (KIE), rate dependence on pH, and comparisons to a
nonenzymatic Pictet-Spengler reaction. The data suggest the
involvement of both an acid-catalyzed step involved in iminium
formation (Figure 1, step 3) and a base-catalyzed step involved
in the final deprotonation step (Figure 1, step 5). We propose
that an active-site glutamate residue previously implicated by
site-directed mutagenesis,14 as well as the protonated tryptamine
substrate, are the mediators of this acid-base chemistry.
Strictosidine synthase does not appear to alter the mechanism
of the reaction, since the KIE data show similar rate-controlling
rearomatization in solution and in the enzymatic reaction.
Additionally, ab initio and crystallographic studies provide
insight into the nature of the enzyme binding site and the
productive transition states involved in the reaction. Notably,
these ab initio calculations suggest that formation of the
spiroindolenine intermediate shown in Figure 1 (step 4a) is not
an important part of the Pictet-Spengler mechanism and that
1
material, the doublet at 6.67 ppm (J ) 3.7 Hz), assigned as 3- H,
collapsed to a sharp singlet, and the doublet at 7.58 ppm (J ) 3.7 Hz)
was absent. Deuteration of the 2-position was measured as >98% by
1
H NMR.
2
2
.2.2. [2- H]-Indole. A solution of NaOH (2.0 M, 40 mL) was added
to a stirring solution of 2-deutero-1-(phenylsulfonyl)indole (3.20 g, 12.4
mmol) in 30 mL of methanol and heated to reflux at 85 °C under
nitrogen overnight until starting material was consumed as monitored
by TLC. Methanol was removed in Vacuo and product was extracted
with ether (5 × 20 mL), washed with brine (3 × 5 mL), dried over
magnesium sulfate, and concentrated in Vacuo to give 1.45 g (98%
1
yield) of a white solid: mp 51-52 °C; H NMR (CDCl
3
, 400 MHz):
δ 6.57 (1H, s), 7.14 (1H, dd, J ) 7.8, 8.1 Hz), 7.22 (1H, dd, J ) 7.8
Hz, 8.1), 7.40 (1H, d, J ) 8.1 Hz), 7.69 (1H, d, J ) 7.8 Hz), 8.51 (1H,
+
NH, br); ESI-MS m/z (rel intens): 119 (MH , 100), 120 (14). Mass
calibration was verified by mixing with natural abundance indole.
1
,2-alkyl shift of this intermediate (Figure 1, step 4b) does not
occur.
2. Materials and Methods
2
.1. General. All reagents except for secologanin were purchased
2
2
.2.3. 2-([2- H]-1H-Indol-3-yl)-2-oxoacetamide. Oxalyl chloride
from Sigma-Aldrich and used without further purification unless
otherwise noted. Secologanin was isolated as previously described.20
HPLC analysis was carried out on a Beckman Coulter System Gold
(
2.15 mL, 25.4 mmol) was added dropwise by syringe to a stirred
solution of 2-deutero-indole (1.0 g, 8.46 mmol) in dry ether under argon
at 0 °C. It was critical that the oxalyl chloride was free of HCl, or else
the deuterium label would be lost. The reaction was allowed to warm
to ambient temperature and was stirred for 1 h. The stirring was then
stopped, the solids were allowed to settle, and excess solution was
removed by syringe and discarded. Ammonia in dioxane (25 mL, 2.0
M) was then added to a cooled solution of the crude product in dry
ether under argon. After stirring overnight at ambient temperature, solid
1
25 HPLC equipped with a model 168 photodiode array detector, Hibar
RT 250-4LiCHrosorb C18 column (Merck), and a home-built column
heater set to maintain 30 °C. Preparative HPLC was performed on a
Beckman Coulter System Gold 127P HPLC equipped with a model
16P single wavelength detector and a Grace Vydac C18 column. Mass
spectroscopy data were obtained on a Waters UPLC-TOF-MS. H and
C 1D NMR spectroscopy were recorded on either a Varian INOVA
00-MHz or Bruker AVANCE 400-MHz or 600-MHz NMR spec-
trometer. Two-dimensional NMR spectroscopy was recorded on a
Bruker AVANCE 600-MHz NMR spectrometer.
2
1
1
13
NaHCO
The resulting solid was loaded dry onto a silica column for purification
by flash chromatography (9:1 f 7:3 CH Cl /CH OH) to yield 1.59 g
93%) of a white solid. H NMR (acetone-d , 400 MHz): δ 7.93 (1H,
), 7.26-7.29 (2H, m), 7.55-7.58 (1H, m), 7.65 (1H, br, NH ),
3
was added, and the solvent was evaporated under vacuum.
5
2
2
3
1
(
6
2
.2. Synthesis of [2- H]-Tryptamine 1b.
br, NH
2
2
8
.35-8.37 (1H, m), 11.25 (1H, br, NH); ESI-MS m/z (rel intens): 190
+
+
(MH , 100), 191 (MH + 1, 9).
2
2
.2.1. [2- H]-1-(Phenylsulfonyl)indole. To a stirring solution of
1
-(phenylsulfonyl)indole (1.03 g, 4.00 mmol, Aldrich) in dry THF (20
2
mL) under dry nitrogen atmosphere, n-butyllithium (1.8 M, 8 mmol,
Aldrich) was added dropwise at -78 °C. The orange mixture was
allowed to warm to 20 °C over 2 h and stirred for 1 h during which
time a precipitate formed. The solution was recooled to -78 °C, and
2.2.4. [2- H]-Tryptamine. Lithium aluminum hydride (10 mL, 1.0
M) was added slowly to a stirring solution of 2-(2-deutero-1H-indol-
3-yl)-2-oxoacetamide (189 mg, 1.00 mmol) in 50 mL of dry 1,4-dioxane
under argon at room temperature over 30 min. At this time, the reaction
mixture was heated to 110 °C and refluxed. After 50 h, wet dioxane
was added cautiously until lithium aluminate precipitated from the
reaction mixture. This mixture was filtered warm through a medium-
porosity fritted glass filter and washed with 100 mL of hot dioxane;
the combined filtrates were dried over anhydrous sodium carbonate,
and solvent was removed in Vacuo to yield an amber oil (crude yield
77%). This oil was purified by preparative RP-HPLC (C18, gradient
D
2
O (1 mL) was added dropwise. The reaction was allowed again to
warm to 20 °C and stirred for 30 min during which time a yellow
precipitate formed. The reaction was quenched with solid K CO and
2
3
stirred with 50 mL of dry ether. After the solution clarified, the reaction
mixture was filtered and concentrated in Vacuo to yield 992 mg (96%)
of crude product as amber oil. Purification by silica gel chromatography
(9:1 hexanes/ethyl acetate) yielded 951 mg (92%) of white crystalline
solid. Alternatively, recrystallization from ether yielded 876 mg (85%)
of water with 0.1% TFA and acetonitrile), evaporated three times from
1
1
3
of white crystals: mp 81-82 °C; H NMR (CDCl , 400 MHz): δ 6.67
H
2
O to afford 165 mg (60%) product as a TFA salt. H NMR (400
7
12 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008
9

Mechanism of Strictosidine Synthase
A R T I C L E S
MHz, D
.13 (t, J ) 7.58 Hz, 1H), 7.05 (t, J ) 7.23 Hz, 1H), 3.16 (t, J ) 6.98
Hz, 1H), 3.00 (t, J ) 6.99 Hz, 1H). ESI-MS m/z (rel intens): 162
2
O): δ 7.53 (d, J ) 9.80 Hz, 1H), 7.39 (d, J ) 8.17 Hz, 1H),
(590 mg, 3.0 mmol) in 5 mL of methanol, 15 mL of water, and 215
7
µL H
condenser at 50 °C for 16 h, was made basic by careful addition of
concentrated NH OH, and was extracted with chloroform. The organic
layer was washed with 1 N NH OH followed by saturated NaCl, dried
over MgSO , and evaporated in Vacuo to yield 600 mg of golden brown
oil. Silica gel chromatography using a gradient of pure CH Cl to 9:1
CH Cl /MeOH afforded 494 mg (82% yield) of yellow crystalline solid.
ESI-MS: m/z 201.14 (expected 201.14). H NMR (400 MHz, CD
with DCl): δ 7.51 (d, J ) 7.83 Hz, 1H, ArH), 7.42 (d, J ) 8.11 Hz,
H, ArH), 7.18 (t, 1H, ArH), 7.10 (t, 1H, ArH), 4.46-4.39 (m, 1H,
2 4
SO (4.0 mmol). The solution was heated and stirred with reflux
+
+
13
(MH , 100), 163 (MH + 1, 3). C NMR (101 MHz, D
2
O): δ 136.9
4
(C9), 127.0 (C4), 122.8 (C7), 120.02 (C5), 118.9 (C6), 112.7 (C8/C3),
4
1
1
09.5 (C3/C8), 40.3 (CH
62.1141(MH ), expected 162.1136. The isotopic enrichment of 2- H
2
), 23.2 (CH
2
N). Exact mass (ESI-MS): m/z
4
+
2
2
2
was measured by ESI-TOF mass spectrometry to be 90.417 ( 0.01%.
2
2
1
3
OD
1
1
3
11
11
CH ), 3.63-3.52 (m, 1H, CH
2
1
2
), 3.29-3.18 (m, 1H, CH
2
), 3.03-
10
14
.85 (m, 2H, CH
.97-1.79 (m, 1H, CH
2
), 2.16 (dqd, J ) 15.05, 7.50, 4.37 Hz, 1H, CH
2
13
),
C
14
15
2
), 1.06 (m, J ) 7.50 Hz, 3H, CH
3
).
3
NMR (101 MHz, CD OD with DCl): δ 136.57, 131.23, 126.48, 122.25,
1
19.57, 118.33, 111.59, 106.85, 54.34, 41.80, 25.61, 19.10, 9.12.
2
.3. Synthesis of Natural Abundance Tryptamine 1a. The identical
procedure was followed for synthesis of natural abundance tryptamine
except H O was substituted for D O following directed lithiation.
Similar yields were obtained. Characterization of final product:
NMR (400 MHz, D O): δ 7.53 (d, J ) 7.93 Hz, 1H), 7.40 (d, J )
.18 Hz, 1H), 7.16 (s, 1H), 7.14 (t, J ) 7.26 Hz, 1H), 7.05 (t, J ) 7.50
2
2
1
H
2
8
13
Hz, 1H), 3.16 (t, J ) 7.00 Hz, 1H), 3.01 (t, J ) 7.00 Hz, 1H).
C
NMR (126 MHz, D O): δ 136.9 (C9), 127.0 (C4), 124.8 (C2), 122.7
2
2
.6. Isolation of Intermediates 8 and 9. The intermediates were
(
C7), 119.96 (C5), 118.8 (C6), 112.6 (C8/C3), 109.6 (C3/C8), 40.3
+
prepared from reactions of 500 mM propanal or acetaldehyde and 1
mM tryptamine in aqueous media acidified to pH 2 with HCl to generate
and 9, respectively. The species 8 was isolated by preparative reverse-
phase HPLC using a gradient of 20-50% aqueous trifluoroacetic acid
0.1%) in acetonitrile. Fractions were chilled on ice and buffered with
(CH
2
), 23.2 (CH
2
N). Exact mass (ESI-TOF): m/z 161.1076 [MH ],
expected 161.1073.
8
(
potassium phosphate (50 mM, pH 5.0) to sufficiently stabilize the
intermediate for evaporation to dryness (by SpeedVac) and character-
ization by NMR (see Supporting Information).
2
.7. Enzymatic Assays. Strictosidine synthase (C. roseus) was
2
.4. Synthesis of Inhibitor 6. A solution of sodium cyanoborohy-
expressed, and activity was measured using an HPLC assay following
previously reported procedures. Strictosidine synthase (80 nM) was
incubated with secologanin and an internal standard in buffer at 30
C. 1-Naphthalene-acetic acid (NAA) was used as an internal standard
at a final concentration of 30 µM. This standard had no effect on the
rate of reaction. Reactions were started by addition of tryptamine (1a
dride (0.086 mmol, in 10 mL methanol) was added to 31.0 mg (0.158
mmol) of tryptamine hydrochloride and 5 activated molecular sieves
in dry methanol (1 mL). Secologanin (50.0 mg, 0.129 mmol) dissolved
in methanol (1 mL) was slowly added to reaction mixture over 4 h.
The reaction was quenched by addition of 400 mg of MP-TsOH resin
20
°
(
Biotage). After 1 h of shaking, the resin was washed with dichlo-
or 1b).
romethane (10 mL) and eluted with ammonia (1 M) in methanol. The
ammonia fraction was collected and evaporated to dryness. The product
was purified by preparative RP-HPLC using a 20 mm × 30 mm reverse-
phase C18 column. Using a 20-50% gradient of acetonitrile: formic
D
(
V/K) values were measured with one substrate varied and the other
. For (V/K)tryp, secologanin
was held at 2.5 mM and tryptamine varied over 0.07-400 µM. For
V/K)sec, tryptamine was held at 1.0 mM, and secologanin concentrations
held at a concentration of at least 50 times K
m
(
acid (0.05%), afforded 4.2 mg (5% yield w/w) of white solid product.
were varied (3.13 µM to 1.6 mM). The greater error in reported V/K
over V values is due to reduced sensitivity of the HPLC assay to changes
in the low micromolar range. The actual concentrations of tryptamine
and secologanin were measured from control samples and calculated
from their area-extinction coefficients. The area-extinction coefficient
for tryptamine was measured directly from standards and for secolo-
ganin from reaction with 2,4-dinitrophenyl hydrazide in pH 5 acetic
acid buffer.
Enzymatic reactions were quenched by addition 2.0 M aqueous
sodium hydroxide at 2 molar equiv over the buffer concentration to
denature the enzyme and prevent any background Pictet-Spengler
condensation. Under these quench conditions the methyl ester of both
secologanin 2 and strictosidine 3 were quantitatively hydrolyzed, and
no side products resulting from this hydrolysis were observed. Initial
rates were obtained from five time points. Quenched aliquots (74 µL)
of the reaction were directly injected onto an analytical HPLC using a
solvent gradient of 22% to 67% acetonitrile in 0.1% aqueous trifluo-
roacetic acid. The absorbance of tryptamine, strictosidine, and NAA
was measured at 228 and 280 nm and exported to Excel spreadsheet
software (Microsoft Corp.) for further analysis.
1
H NMR (500 MHz, CD OD): δ 1.86-1.99 (m, 1 H), 2.63-2.64
3
(
m, 1 H), 2.81-2.85 (m, 1 H), 3.04-3.16 (m, 2 H), 3.16-3.38 (m, 12
H), 3.63-3.69 (m, 4 H), 3.86 (dd, J ) 5.2, 5.1 Hz, 2 H), 3.91 (dd, J
11.8, 1.7 Hz, 1 H), 4.69 (d, J ) 7.9 Hz 1H), 5.28 (d, J ) 10.6 Hz,
H), 5.31 (d, J ) 17.3 Hz, 1 H), 5.54 (d, J ) 6.6 Hz, 1 H), 5.70-5.79
m, 1H), 7.05 (dd, J ) 7.7, 7.2 Hz, 1 H), 7.12 (dd, J ) 7.6, 7.4 Hz, 1
H), 7.19 (bs, 1 H), 7.38 (d, J ) 8.1 Hz, 1 H), 7.55 (bs, 1 H), 7.60 (d,
J ) 7.9 Hz, 1 H), 8.50 (bs, 1 H). C NMR (125 MHz, CD
3.5, 27.9, 31.9, 45.2, 47.5, 52.1, 57.1, 57.4, 62.9, 71.7, 74.7, 78.0,
)
1
(
1
3
3
OD): δ
2
7
1
5
8.6, 97.5, 100.2, 109.9, 110.0, 112.6, 119.0, 120.1, 122.8, 124.3, 128.2,
35.01, 138.4, 154.8, 169.5. Exact mass (ESI-TOF): [M ], calc:
33.2499, found: 533.2495.
+
2
.5. Synthesis of 1-Ethyl-2,3,4,9-tetrahydro-1H-pyrido[3,4-b]-
indole 7. This compound was used as an authentic standard in model
chemical reactions. Propionaldehyde (1.91 g, 33.0 mmol) dissolved in
Strictosidine peak areas were integrated and normalized to the
internal standard. Initial rates were determined from the slope by linear
1
0 mL of water was added to a stirring suspension of tryptamine HCl
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 713

A R T I C L E S
Maresh et al.
least-squares fit of five normalized data points for each concentration
Scheme 1
of varied substrate. Kinetic parameters V, V/K, and K
m
were estimated
from the initial rates directly fit to appropriate forms of the Michaelis-
2
1
Menten equation by nonlinear least-squares fit using Origin 7.0
software (OriginLab Corp.).
(
14 µM) in a 50 mM phosphate buffer with 10% glycerol at 30 °C.
To avoid interference from potential trace impurities in 1b, the natural
abundance substrate 1a was synthesized and purified under conditions
identical to those for 1b to ensure that any trace impurities would be
present in both labeled and unlabeled samples. Isotope effects on V
and V/K were determined by dividing the V or V/K values obtained
with 1a from 1b. Although the isotopic substitution of 1b was 90.42%,
reported, KIE values were not corrected. Unlike the case of competitive
KIE measurements, the applied isotopic corrections for the direct
measurements in this work were not found to be significant (usually
within the confidence interval of the uncorrected value).
The pH was adjusted by adding concentrated HCl or NaOH. The
concentration of strictosidine synthase was estimated from the known
extinction coefficient and absorbance at 280 nm.
2
.11. Kinetic Assays of Solution Reaction. Assays contained
distilled propanal (100 mM), tryptamine 1a or 1b (0.3 mM), internal
standard NAA (0.2 mM) and buffer at 30 °C (see Figure S9A,
Supporting Information). Buffer concentration and pH were varied in
acetic acid buffer (pK ′ ) 4.60). Ionic strength was maintained at a
constant ionic strength (I ) 0.5) following the procedure described for
a
enzymatic assay buffers. The assay was initiated by addition of
tryptamine 1a or [2- H]-tryptamine 1b. The initial tryptamine concen-
2
.8. Buffers for Enzyme Kinetics. For assays at pH 7.0, 100 mM
NaH PO , buffer was used. For log V-pH profiles, two different buffer
systems were prepared to control for specific buffer effects on enzyme
activity. To control ionic strength, the “practical pK ” (pK ′) at
experimental conditions was calculated for each buffering agent from
2
2
4
tration was determined from aldehyde-free control samples. Kinetic
data were obtained by HPLC from the peak area at 280 nm and exported
to Excel (Microsoft Corp.) for analysis. The measured area-extinction
coefficient of each species was used to convert corrected areas to
concentrations. The area-extinction coefficients of tryptamine, aldehyde,
and product were measured from authentic samples. To measure the
area-extinction coefficient of the intermediate, this species was isolated
by preparative HPLC to approximately 100 µM. At 10 °C, the
intermediate decomposed cleanly to tryptamine, and the area-extinction
coefficient was determined from tryptamine appearance. The results
of four replicate decomposition reactions were analyzed simultaneously
by nonlinear least-squares fitting in Origin 7.0 (OriginLab Corp.).
These data were fit to various models using the Gepasi computer
program26 to obtain the first-order rate constants. The best fit was
obtained by a system of two coupled reversible rate equations taking
into account the presence of the nonproductive intermediate as in
Scheme 1 (see Supporting Information Figure S9B for example of
quality of fit). For each experiment, the time course data for tryptamine,
intermediate, and product at all aldehyde concentrations were simul-
taneously optimized to the kinetic model by fitting the four parameters
a
a
2
2
the Debye-H u¨ ckel equation as described by Ellis using A ) 0.5161
23
for 30 °C. A set of constant composition buffers with a constant ionic
strength (I ) 0.20) were prepared from a mixed buffer system of acetic
acid (pK
a
′ ) 4.64), Tris (pK
a
′ ) 6.32), and Bis-Tris (pK ′ ) 8.00)
a
2
2
following the procedure of Ellis covering the pH of 3.7 to 8.76 (pH
at 30 °C: 3.70, 4.20, 4.62, 5.06, 5.52, 5.89, 6.89, 6.37, 6.83, 7.29,
7
.81, 8.27, 8.76). Alternatively, overlapping constant ionic strength
buffers (I ) 0.31) were also used: citric acid (pK ) 2.97, pK
.61, pK ) 6.25) covered the pH range from 3.1 to 9.46 (pH at 30
C: 3.10, 3.42, 3.79, 4.10, 4.40, 4.79, 5.12, 5.59, 6.14, 6.67, and 6.93),
potassium phosphate buffers (pK ) 6.72) covered an overlapping
pH range of 6.50 to 8.04 (pH at 30 °C: 6.50, 6.80, 7.19, 7.50, 7.87,
and 8.04), and borate (pK ′ ) 9.04) buffers were used for assays over
a pH range of 8.17 to 9.19 (pH at 30 °C: 8.17, 8.35, 8.70, 8.92, and
.19). Both buffer conditions yielded equivalent results; identical pK
values were calculated from both buffer systems.
.9. pH Rate Profiles. The enzyme was incubated with secologanin
2.5 mM), strictosidine synthase (5 nM), and NAA internal standard
and buffer (from pH 3.1 to 9.7) at 30 °C. Either 1a or 1b was added
0.3 mM) to start the reaction. Samples were quenched with 2.0 M
a
′
1
a 2
′ )
4
a 3
′
°
a 2
′
a
9
a
2
-
14
2
in Scheme 1 (each bound from 10 to 10 ) and the initial tryptamine
concentration of each sample (bound from 0.1 to 0.4 mM). The fastest,
most reliable method was found to be the Hooke and Jeeves pattern-
based direct-search algorithm (50 maximum iterations, 10 conver-
gence tolerance, and F ) 0.2). Order tests with varied tryptamine (10-
(
(
-
8
NaOH sufficient to adjust pH to ∼12 by pH paper as interfering peaks
were observed when the basic buffers were quenched with too much
base. Initial rates from the slope of the linear fit of five data points
were determined as described for enzyme assays above. The data was
1
000 µM) and propanal (5-95 mM) determined that the forward
Pictet-Spengler reaction (k ) was first order in tryptamine and aldehyde,
and the forward side reaction (k ) was first order in tryptamine and
second order in aldehyde. The reverse reactions were both first order.
To measure the parameters kHA and k in eq 7 (Results and
Discussion), the parameter B (total buffer) was defined as B ) [AcOH]
[AcO ]. Substituting this into eq 7 to replace [AcO ] yields eq 2.
P
2
4,25
I
fit to a diprotic model shown in eq 1
Corp.).
using Origin 7.0 (OriginLab
A
Vmax
H
t
t
V )
(1)
-
-
+
K2
1
+
+
K1
H
kobs ) kHA[AcOH] + k (B - [AcOH]) + k
o
(2)
A
t
2
.10. Circular Dichroism. CD measurements were obtained on an
Aviv Instruments, Inc. circular dichroism spectrometer model 202 from
95 to 250 nm. Each wavelength step was 1 nm, and the averaging
Rearrangement of eq 2 yields eq 3, the form used to plot Figure 5B
and directly solve kHA and k from the linear intercepts. Note that
in eq 3 is mathematically identical to fraction of AcOH
HA) plotted in Figure 5B
a
1
[
AcOH])/B
t
time for each wavelength was 6 s. Three CD spectra were obtained
and averaged for each sample at 30 °C. Scans of pure solvent were
used for a baseline and subtracted from the protein measurements. All
samples were degassed. Spectra were obtained of strictosidine synthase
(F
(
kobs - k )/B ) k + (k - k )[AcOH]/Bt
(3)
o
t
A
HA
A
.
(
(
(
21) Northrop, D. B. Anal. Biochem. 1983, 132, 457-461.
22) Ellis, K. J.; Morrison, J. F. Methods Enzymol. 1982, 87, 405-426.
23) Holtzhauer, M. Buffers: Theoretical Considerations. In Basic Methods for
the Biochemical Lab; Springer-Verlag: Heidelberg, Germany, 2006; pp
To solve for k
against total buffer B
buffer (k ) by least-squares fitting. Each sample was corrected by k
and B to find (kobs - k )/B . Samples of the same pH were then used
o
, the rate constants kobs at a given pH were plotted
t
. This plot was then linearly extrapolated to zero
1
91-197.
o
o
(
24) Cleland, W. W. In AdVances in Enzymology and Related Areas of Molecular
Biology; Meister, A., Ed.; John Wiley & Sons: New York, 1977; Vol. 45,
pp 320-348.
t
o
t
to plot Figure 5B. The derived values for 1a are kHA ) 6.2 ( 0.3 ×
(26) Mendes, P. Trends Biochem. Sci. 1997, 22, 361-363.
(25) Cook, P. F.; Cleland, W. W. Enzyme Kinetics and Mechanism; Garland
Science: New York, 2007; pp 325-366.
714 J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008

Mechanism of Strictosidine Synthase
A R T I C L E S
0- , k
4
) 1.5 ( 1.8 × 10-5, k
) (0.38 ( 0.01)[H ] + (1.31 ( 0.05
+
1
A
o
Table 1. Data Collection and Refinement Statistics of Strictosidine
Synthase-Inhibitor Complex
-
5
-1 -1
-4
×
10
10-5, k
M
s
) and 1b are kHA ) 2.3 ( 0.3 × 10 , k
A
) 2.0 ( 1.4
+
-6
-1 -1
×
o
) (0.26 ( 0.02)[H ] + (6.7 ( 0.6 × 10
M
s ).
wavelength (Å)
space group
0.8148
R3
2.12. Ab Initio Calculations. Geometry optimizations and vibrational
frequencies were computed using the 6-31G* basis set and the
PW1PW91 density functional (M. J. Frisch, Gaussian Inc., Wallingford,
CT (2003). See Supporting Information for full citation.) which has
recently been found to provide a better description of organic cyclization
Unit Cell (Å)
a ) b ) 150.0
c ) 121.7
119018
20319
1.0
total reflections
unique reflections
mosaicity
27
reactions than B3LYP. Transition states were found using the growing
string method and a memory augmented eigenvector following
resolution (Å)
completeness (%)
I/σ(I)
20-3.00
100 (100)
2
8
a
algorithm. Single-point solvent-corrected energies were computed
using Tomasi’s polarized continuum model (PCM) with the 6-311+G*
basis set and the PW1PW91 functional. All calculations and reported
18.1 (3.6)
10.1 (52.9)
20-3.00
18.7/24.7
48.6
Rmerge (%)b
resolution (Å)
2
9
values were determined by the Gaussian ‘03 program.
Rcryst/Rfree (%)c
2
.13. Inhibition Kinetics of Inhibitor 6. To assess whether
average B (Å) for protein
average B (Å) for inhibitor
average B (Å) for water
compound 6 acted as an inhibitor for strictosidine synthase, IC50 was
determined by HPLC with tryptamine (20 µM), secologanin (80 µM),
internal standard NAA (60 µM), phosphate buffer (100 mM, pH 7.0),
C. roseus strictosidine synthase (10 nM), and inhibitor at 30 °C. The
reaction was quenched with 2.0 M sodium hydroxide (to 200 mM final)
at four time points. Tryptamine disappearance and strictosidine ap-
pearance were monitored at 280 nm for assays with inhibitor,
respectively. Peak areas were normalized to the internal standard as in
other experiments and initial rates determined as described for enzyme
assays above. The rates were fit to a sigmoidal logistic curve with Origin
52.4
30.5
Number of Atoms
non-hydrogen
water
4896
14
rms Deviations
bond (Å)
angles (deg)
0.019
1.94
a
The values in parentheses correspond to the last resolution shell. b Rmerge
∑hkl∑i|Ii (hkl) - 〈I(hkl)〉|/∑hkl∑i〈Ii(hkl)〉, where 〈I(hkl)〉 is the average
intensity over symmetry-equivalent reflections. c R
- |Fc(hkl)||/∑hkl|Fo(hkl)|, where Fo and Fc are observed and calculated
)
7
.0 (OriginLab Corp.) with a Hill slope factor of 1. An IC50-value of
(R ) ) ∑ ||F (hkl)|
cryst
free
hkl
o
3
( 0.5 nM was obtained.
structure factors, respectively.
2
.14. Cocrystallization of Strictosidine Synthase with Inhibitor
6
. The strictosidine synthase gene (excluding the signal peptide) from
program REFMAC5.36 The final R factor is 18.7% for all data in the
resolution range 20-3.0 Å (Rfree ) 24.7%). Of the non-glycine residues,
R. serpentina was expressed in Escherichia coli and purified as
3
0
previously described. The enzyme was concentrated to 4-5 mg/mL
in 10 mM Tris-HCl, pH 8.0, before crystallization. Crystals of
strictosidine synthase in complex with inhibitor were obtained by
cocrystallization in the presence of 0.1 mM inhibitor in the enzyme
solution using the hanging drop vapor diffusion method. The reservoir
solution contained 0.8 M potassium sodium tartrate tetrahydrate and
3
7
8
5.6% fall in the most favored regions of a Ramachandran plot
38
generated by PROCHECK. The refinement statistics are summarized
in Table 1.
3. Results and Discussion
0
.1 M HEPES, pH 7.3.
3.1. Enzymatic Kinetic Isotope Effect: Rate-Controlling
Step. To probe the specific role that strictosidine synthase plays
in catalysis, the KIE of the enzymatic reaction (C. roseus, 89%
homology to the enzyme from R. serpentina) was measured with
The crystals of strictosidine synthase-inhibitor complex were
incubated for 20 s in the reservoir solution containing 23% glycerol
prior to flash cooling at 100 K in a liquid nitrogen stream. X-ray data
were collected using synchrotron radiation at the X11 beamline of
EMBL-Hamburg. The complete data were collected to 3.0 Å resolution.
ThedatawereprocessedusingDENZOandscaledusingSCALEPACK.3231
The data collection statistics are shown in the Table 1. The scaled data
was nonisomorphous with respect to the native data. The structure was
2
[2- H]-tryptamine 1b to determine which of the reaction steps
2
could be rate controlling. Tryptamine 1a and [2- H]-tryptamine
1
b (Figure 1) were prepared following parallel synthetic
procedures, and the rates of reaction with these substrates were
3
9
compared using an HPLC assay (Figure 3). At saturating
solved using molecular replacement method using the Auto-Rickshaw
software pipeline.32 The high-resolution structure of strictosidine
concentrations of both substrates 1a/b and 2 and optimal pH
D
7
.0, a primary isotope effect of 2.67 ( 0.13 for V was observed
synthase (PDB code 1FP8) was used as a search model. Within the
33
at 30 °C, indicating that the loss of the proton at the indole
2-position under Vmax conditions is rate controlling (Figure 1,
pipeline, the program MOLREP was used for molecular replacement,
and the rigid body positional as well as B-factor refinement was
3
4
40,41
D
performed using the program CNS. The residual electron density for
inhibitor was clearly visible and was built using the graphics program
step 5).
The primary KIE of V reveals that under Vmax
conditions, the chemical steps of the enzymatic reaction are
largely rate limiting.
35
COOT. Subsequent refinement procedures were carried out using the
The ring closure and deprotonation of the Pictet-Spengler
reaction (Figure 1, steps 4 and 5) is essentially an electrophilic
(
27) Matsuda, S. P. T.; Wilson, W. K.; Xiong, Q. Org. Biomol. Chem. 2006, 4,
30-543.
28) Peters, B.; Heyden, A.; Bell, A. T.; Chakraborty, A. J. Chem. Phys. 2004,
20, 7877-7886.
5
(
1
(36) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr., Sect. D
1997, 53, 240-255.
(
29) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027-2094.
30) Ma, X. Y.; Koepke, J.; Fritzsch, G.; Diem, R.; Kutchan, T. M.; Michel,
H.; St o¨ ckigt, J. Biochem. Biophys. Acta 2004, 1702, 121-124.
31) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307-326.
32) Panjikar, S.; Parthasarathy, V.; Lamzin, V. S.; Weiss, M. S.; Tucker, P. A.
Acta Crystallogr., Sect. D 2005, 61, 449-457.
(
(37) Ramakrishnan, C.; Ramachandran, G. N. Biophys. J. 1965, 5, 909-933.
(38) Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. J.
Appl. Crystallogr. 1993, 26, 283-291.
(
(
(39) Parkin, D. W. Methods for the Determination of Competitive and
Noncompetitive Kinetic Isotope Effects. In Enzyme Mechanism from Isotope
Effects; Cook, P. F., Ed.; CRC Press: Boca Raton, 1991; pp 269-290.
(40) Northrop, D. B. Biochemistry 1981, 20, 4056-4061.
(
(
33) Vagin, A.; Teplyakov, A. J. Appl. Crystallogr. 1997, 30, 1022-1025.
34) Br u¨ nger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros, P.;
Grosse-Kunstleve, R. W.; Jiang, J. S.; Kuszewski, J.; Nilges, M.; Pannu,
N. S.; Read, R. J.; L. M., R.; Simonson, T.; Warren, G. L. Acta Crystallogr.,
Sect. D 1998, 54, 905-921.
(41) For enzyme reactions, isotope effects on the limiting macroscopic rate
D
D
m a
constants Vmax and Vmax/K are shortened to V and (V/K ) respectively
(the subscript “a” indicates the varied substrate). For chemical reactions
D
(
35) Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D 2004, 60, 2126-2132.
H
k /k
D
is simply k.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 715

A R T I C L E S
Maresh et al.
44
concentrations, and the acidity of the leaving proton. Assuming
that the steps prior to the rate-limiting deprotonation step are
reversible in the enzyme active site, a primary KIE should be
expected if the rates of the reverse reaction (i.e., k-3 in Figure
1
) are fast relative to the rate of deprotonation (k4 in Figure 1).
.2. Enzymatic KIE under V/K Conditions. A significant,
3
but smaller, primary KIE was observed at limiting concentrations
D
of either tryptamine 1a/b or secologanin 2 ( (V/Ktryp) ) 1.87
D
(
0.37 and (V/Ksec) ) 1.45 ( 0.28) at pH ) 7.0 and 30 °C.
At conditions of low substrate, where V/K is the pseudo-second-
order rate constant, the active site is not saturated. For a substrate
that reacts to give products faster than it dissociates from the
enzyme-tryptamine-secologanin ternary complex (a “sticky”
substrate”), the isotope effect on V/K is reduced by an external
4
9,50
D
forward commitment to catalysis.
Since (V/Ktryp) is slightly
D
greater than (V/Ksec), the tryptamine substrate is likely released
at a rate faster than the rate for release of secologanin from the
enzyme-tryptamine-secologanin ternary complex.
Equation 4 describes the relationship between the Michaelis
constant (Km) and the dissociation constant KD. From the V/K
50
D
data, we estimate KD for secologanin and tryptamine as KD‚tryp
)
(
2 µM (Km‚tryp ) 4 ( 1 µM) and KD‚s ) 11 µM (Km‚s ) 39
4 µM), respectively.
D
V - 1
)
K /K
(4)
m
D
D
(V/K) - 1
The magnitude of D(V/K) isotope effects also provides
Figure 3. (A) Representative HPLC trace for enzyme catalyzed reaction
of tryptamine 1a with secologanin 2 to yield strictosidine 3. Under quench
conditions, the methyl ester of secologanin and strictosidine have been
hydrolyzed. (B) Representative kinetic data for 1a and 1b (2.5 mM
secologanin 2, 80 nM strictosidine synthase, 30 °C) obtained using this
assay.
information about binding order for multisubstrate enzymes. For
D
an ordered mechanism, (V/K) approaches unity if the varied
5
1
substrate binds first. At saturating concentrations of either
tryptamine 1 or secologanin 2, primary KIE values greater than
D
D
1 were measured ( (V/Ktryp) ) 1.87 ( 0.37 and (V/Ksec) )
.45 ( 0.28), suggesting that the substrates bind in random
1
aromatic substitution reaction. Since the aromaticity is broken
during the cyclization step (Figure 1, step 4), it might be
order. This is consistent with crystallographic evidence which
demonstrates that both tryptamine 1a and secologanin 2
substrates can bind to strictosidine synthase independently of
42
reasonable to assume that cyclization would be rate controlling
and that no primary isotope effect would be observed with
14
one another.
2
[
2- H]-tryptamine 1b. While it is true that some nonenzymatic
3
.2.1. Acid-Base Catalysis in Strictosidine Synthase. The
electrophilic aromatic substitution reactions do exhibit small KIE
values, there are in fact many examples that exhibit primary
KIE values for the replaced proton at the site of electrophilic
aromatic substitution.4 These previously reported primary KIE
values indicate that the final deprotonation step (Figure 1, step
pH dependence of kinetic parameters (V and V/K) can result
from the titration of groups involved in catalysis/substrate
binding or maintenance of enzyme structural conformation.
Enzyme conformational changes can also impact the kinetic
parameters. However, the structure of strictosidine synthase
without substrate compared to the enzyme structure bound to
either tryptamine, secologanin, or a bisubstrate inhibitor (see
3,44
5
) is very often rate limiting in nonenzymatic electrophilic
aromatic substitution reactions.
The emergence of a primary KIE for nonenzymatic electro-
philic aromatic substitution reactions appears to be sensitive to
the rate of the reverse reaction. In other words, as the rate of
the reverse reaction k-3 increases relative to deprotonation k4
14
below) are very similar. This suggests, although does not
prove, that no major conformational enzyme rearrangements
occur upon substrate binding. We therefore make the assumption
that since the enzyme-substrate(s) complex dominates under
Vmax conditions, the pH dependence of this parameter of this
enzymatic reaction is only sensitive to catalysis, not binding or
(
Figure 1), the deprotonation becomes rate controlling, and the
45-48
observed KIE increases.
Factors contributing to the emer-
gence of primary KIEs include sterics, base and reactant
24,25
conformation, when the chemical reaction is rate limiting.
Substrate concentrations were 100 fold above the Km for
tryptamine and 50 fold above the Km value for secologanin
substrate as measured at pH 6.8. In this analysis, we make the
(
42) Jackson, A. H.; Smith, A. E. Tetrahedron 1968, 24, 403-413.
43) Zollinger, H. In AdVances in Physical Organic Chemistry; Gold, V., Ed.;
Academic Press: London, 1964; Vol. 2, pp 163-196.
(
(
44) Berliner, E. In Progress in Physical Organic Chemistry; Cohen, S. G.,
Streitweiser, A., Taft, R. W., Eds.; Interscience: New York, 1964; Vol. 2,
pp 253-321.
(49) Northrop, D. B. J. Chem. Edu. 1998, 75, 1153-1157.
(50) Klinman, J. P.; Matthews, R. G. J. Am. Chem. Soc. 1985, 107, 1058-
1060.
(
45) Jackson, A. H.; Lynch, P. P. J. Chem. Soc., Perkin Trans. II 1987, 1483-
1
488.
(46) Zollinger, H. HelV. Chim. Acta 1955, 38, 1623-1631.
(47) Zollinger, H. HelV. Chim. Acta 1955, 38, 1617-1622.
(48) Zollinger, H. HelV. Chim. Acta 1955, 38, 1597-1616.
(51) Karsten, W. E.; Cook, P. F. In Isotope Effects in Chemistry and Biology;
Kohen, A., Limbach, H.-H., Eds.; Taylor & Francis: Boca Raton, 2006;
pp 793-809.
716 J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008

Mechanism of Strictosidine Synthase
A R T I C L E S
important functional group, is less straightforward. The active
site of the enzyme is largely hydrophobic, with only the residues
Tyr151, His 307, and Glu309 located near the binding pocket.14
Mutation of the tyrosine residue (expected pKa in solution is
-
1
1
0) to phenylalanine results in only a small drop in kcat (78 min
-
1
for wild type to 58 min for Tyr151Phe), suggesting that this
residue does not play a critical role in acid-base catalysis.14
Mutation of His307 (typical pKa within an enzyme active site
2
4
is 5.5-7) reduces the kcat value by approximately 40 fold,
significantly less than the 900-fold decrease observed with the
1
4
Glu309Ala mutant. This histidine residue appears to be over
Å away from the amino group of the bound tryptamine, and
4
over 5 Å from the aldehyde oxygen of secologanin; thus, it is
concluded from analysis of the crystal structure that this residue
is not optimally positioned for a direct role in catalysis.14 No
ordered water molecules appear in the enzyme structure,
suggesting that these ionizable residues do not play a role in
polarizing water molecules for catalysis. It is possible that the
pKa value may be represented by the tryptamine substrate,
suggesting that the protonation state of tryptamine plays a role
in catalysis. Although the pKa of tryptamine is 10.2 in aqueous
Figure 4. pH profile of enzymatic activity under saturating (Vmax) substrate
conditions. Tryptamine 1a is represented by filled symbols; [2- H]-
tryptamine 1b is represented by open symbols. Circles, squares, triangles
2
represent citrate, phosphate, and borate buffer conditions, respectively (see
D
Materials and Methods). For a plot of the V depenedence on pH see
Supporting Information.
5
2
media, the pKa of a primary amine may be observed at 7.5-
0.5 in the low dielectric environment of an enzyme binding
assumption that the Km values for the enzyme do not change
significantly over the pH range measured and that the ternary
complex is still accessible to proton transfer with the solution.
Although the pH dependence of strictosidine synthase has
previously been measured,1 the pKa values corresponding to
these changes in enzyme activity have not been reported. We
therefore measured the rate of strictosidine synthase in a series
of buffers at fixed ionic strength under saturating substrate
concentrations. Loss of activity at both acidic and basic pH
values was observed (Figure 4). Fit of these data to a diprotic
model suggested pKa values of two ionizable species at 4.70
and 8.28.
1
24
pocket. However, since the rate-controlling step occurs well
after alkylation of the primary amine, it is not clear that an effect
on this moiety would be observed in the pH dependence of V.
It is also possible that the pKa value of 8.3 is representative of
the iminium intermediate; under basic conditions, the iminium
could be deprotonated, thereby preventing electrophilic attack
by the indole (step 4, Figure 1). In short, the pH profile data
strongly suggest that an acidic residue is important for catalysis,
and these data are supported by site-directed mutagenesis studies
implicating Glu309 in the mechanism. The meaning of the basic
pKa value is less clear, since no obvious basic amino acid
appears to be involved in the mechanism.
6,18
At Vmax conditions, it is generally assumed that rate is limited
D
by chemical catalysis and product release. A primary V effect
3.2.2. Isotope Effects in the Nonenzymatic Pictet-Spengler
indicates that the chemistry of the reaction is significantly rate-
controlling in comparison to product release. To verify that the
chemistry (not product release) remained rate limiting over the
pH range, we also assayed 1b to verify the presence of the
primary KIE over the pH range (Figure 4). DV decreased
somewhat at acidic pH, and this effect may be due to
contributions from either rate-limiting product release or from
another mechanism. Although the existence of this effect
hampers exact assignment of the lower pKa value, the significant
primary isotope effect does indicate that the residue titrated at
low pH is a catalytic species having a pKa value of ap-
proximately 4.6. Circular dichroism (CD) spectroscopy sug-
gested that the protein secondary structure was unchanged at
basic and acidic pH values (data not shown), although CD
spectroscopy does not provide a detailed picture of the protein
structure.
Reaction. The Pictet-Spengler reaction occurs nonenzymati-
cally in acidic media and is, in fact, an important reaction in
8
organic synthesis. Surprisingly, to the best of our knowledge,
isotope effects have not been reported for the Pictet-Spengler
reaction under nonenzymatic conditions. The closest analogous
system is the report of primary isotope effects between 2 and 3
2
for coupling of [2- H]-3-methyl-indole with electrophilic p-
4
5
nitrobenzenediazonium ion at 30 °C. Therefore, to provide a
D
context for our observed enzymatic V of 2.67, we studied a
model Pictet-Spengler reaction in solution.
Acetic acid (AcOH), with a pKa value similar to that of the
glutamate residue implicated in enzymatic catalysis, was used
to buffer the reaction under mild, aqueous reaction conditions.
We found the ester of secologanin is hydrolytically labile on
the slower time scale of solution kinetics, and thus the model
aldehyde propanal 4 (Figure 1) was used in these experiments.
As assayed by HPLC, the equilibrium constant was less than
Although other effects, such as structural perturbations, may
contribute to the reduced activity of strictosidine synthase at
extreme pH values, Glu309, an amino acid with a pKa of
approximately 4.5, has been demonstrated to be critical for
1
, and excess aldehyde or tryptamine was required to drive the
reaction to completion. Kinetic analysis was complicated by
the presence of an intermediate that was eventually consumed
by the end of the reaction. Isolation and structural analysis
suggested that this intermediate was a carbinolamine resulting
activity as evidenced by site-directed mutagenesis (900-fold drop
in Vmax for Glu309Ala),1
4,19
indicating that this residue is
important in the catalytic mechanism. The assignment of the
basic pKa value, assuming that the loss of enzyme activity
represents a change in ionization state of a catalytically
(
52) Perrin, D. D. Dissociation Constants of Organic Bases in Aqueous Solution
(and Supplement, 1972); Butterworths: London, 1965.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 717

A R T I C L E S
Maresh et al.
from condensation of aldehyde with the indole nitrogen (Sup-
porting Information). When isolated by preparative HPLC, the
intermediate decomposed cleanly to tryptamine and aldehyde.
Therefore, it was simply treated as a reversible, nonproductive
intermediate for estimation of rate constants (see Materials and
Methods). No other intermediates were detectable by HPLC,
LC-MS, or NMR. Attempts to trap intermediates by reductive
amination led to complex mixtures consisting primarily of over-
alkylated product (data not shown).
The kinetic behavior of the reaction of tryptamine 1a/b and
propanal 4 in solution was surveyed in a variety of pH values,
buffers, and substrate concentrations using an HPLC assay (see
Figure S11, Supporting Information). As expected, the Pictet-
Spengler reaction was first order in both reactants. The apparent
KIE for the reaction under these varied conditions ranged from
2
.0 to 2.6, indicating that deprotonation (Figure 1, step 5) is at
least a partially rate-controlling step in both the solution and
enzymatic reactions and the expression of the KIE is sensitive
to the reaction conditions.
3
.2.3. The Role of General Acid-Base Catalysis in the
Pictet-Spengler Reaction. Electrophilic aromatic substitution
reactions that exhibit primary isotope effects are base cata-
lyzed.44 However, the Pictet-Spengler reaction is generally
carried out under acidic conditions. To understand the role of
acid-base catalysis in the mechanism, we examined the
dependence of the primary KIE on the solution reaction
conditions in more detail. The apparent rate constant kobs (eq
5) was measured for several pH values between 4 and 5 in acetic
acid buffers of constant ionic strength for reaction of propanal
2
4
with tryptamine 1a and [2- H]-tryptamine 1b (Figure S10,
Supporting Information). These data revealed that the overall
rate of the reaction is increased by the presence of acid.
Therefore, although the primary rate-controlling step [deproto-
nation (Figure 1, step 5)] is base-catalyzed, the overall Pictet-
Spengler reaction is apparently acid catalyzed.
Figure 5. (A) Apparent second-order rate constants plotted against
concentration of acetic acid CH3CO2H [AcOH] for tryptamine 1a and 1b.
(
B) Plotted rate constants are normalized for total buffer concentration (Bt).
By extrapolation, the KIE is 1 for kA (the limiting rate constant for AcO-
in the absence of acid) and 2.65 for kHA (the limiting rate constant for AcOH
in the absence of general base).
observed rate ) kobs[tryptamine 1][propanal 4] (5)
To determine kHA and kA in eq 7, the rate constants in Figure
5
A were corrected for ko and plotted against the fraction of
We further noted that as the pH was kept constant, the
reaction rate increased with concentration of the acetate buffer
-
protonated acetic acid buffer, a continuum from AcO to AcOH
(see Materials and Methods), graphically revealing the effects
(AcOH), suggesting that the reaction is also catalyzed by buffer.
of the acidic and basic species on the KIE. It was found that
the KIE extrapolated to 1 as [AcOH] approached zero (Figure
5
the KIE reached a maximum value of 2.65 ( 0.10. By
considering the expression of the KIE as a mechanistic probe,
we hypothesize that in the absence of general acid, the acid-
catalyzed step becomes rate-controlling. Therefore, the depro-
tonation step is no longer rate limiting, and the primary KIE is
no longer observed.
We combined these two factors, pH and buffer concentration,
by considering the reaction rate as a function of protonated
buffer [AcOH] (eq 6). [AcOH] was calculated from the literature
-
B). Conversely, as [AcOH] increased and [AcO ] decreased,
5
2
pKa value adjusted for ionic strength, the measured pH, and
known buffer concentration (Materials and Methods). When kobs
was plotted against [AcOH], a linear dependence in rate was
observed (Figure 5A). Such a linear dependence was not
-
observed for [AcO ] concentration. In this aqueous model
reaction, the rate does not appear to be linearly dependent on
pH but on protonated buffer, AcOH. Therefore the Pictet-
Spengler reaction under aqueous acidic conditions obeys general
acid catalysis.
-
kobs ) kHA[AcOH] + k [AcO ] + k
(7)
A
o
Two possible steps may account for the apparent acid
catalysis. Acid could be required to protonate the iminium
kobs ) kHA[AcOH] + k′
(6)
5
3
(Figure 1, KA(imine)). The protonated iminium is a stronger
electrophile than the imine and is therefore a better substrate
for cyclization (Figure 1, step 4). Another candidate for the acid-
catalyzed step is protonation of the carbinolamine involved in
Since isotope effects reveal that a base-catalyzed step,
deprotonation, is rate limiting, we introduced general base
catalysis by acetate ion AcO to eq 6 in the form of eq 7. In
-
this equation, ko represents non-buffer-mediated catalysis (ca-
talysis by solvent H2O and H ) for each sample in Figure 5A.
(
53) (a) Hupe, D. J. New Compr. Biochem. 1984, 6, 271-301. (b) Jencks, W.
+
P. Catalysis in Chemistry and Enzymology; McGraw-Hill: New York, 1969.
718 J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008

Mechanism of Strictosidine Synthase
A R T I C L E S
the formation of the iminium species (Figure 1, step 3).53 To
distinguish between these possibilities, N-ω-methyltryptamine
that the newly deprotonated Glu309 (like acetate in the solution
reaction) could also catalyze the rate-limiting deprotonation step
(Figure 1, step 5; Figure 2).
5
(Figure 1) was reacted with propanal 4 in the same acetic
acid buffers used for buffer dependence studies with tryptamine
Strictosidine synthase catalyzes iminium formation, which
can be rate limiting in solution. The enzyme acts as a scaffold
to increase the local concentrations of the tryptamine, aldehyde,
and acid catalyst species. To qualitatively provide a sense of
how important this effect is, we compared the rate laws for the
solution (nonenzymatic) reaction with acetic acid catalyst and
2
1
a and [2- H]-tryptamine 1b. The iminium resulting from
reaction of this secondary amine with aldehyde is charged and
electrophilic, independent of acid. Therefore, if protonation of
the neutral imine species limited available electrophile (Figure
1
1
, KA(imine)), then the acid catalysis observed with tryptamine
a and 1b should be abolished with 5. However, the reaction
-4
-2 -1
propanalsubstrate4(rate)(6.25×10 M s )[AcOH][aldehyde][tryptamine])
with 5 displays the same dependence on general acid as
tryptamine 1a and 1b (Figure S12, Supporting Information),
leaving protonation of the carbinolamine as the most likely rate-
controlling step.
with a previously reported rate law for the enzymatic reaction
-1
19
at Vmax (rate ) (1.67 s )[enzyme]). Setting the concentrations
of the catalysts to an arbitrary concentration of 1 M, these rate
-
4
-1 -1
laws become (6.25 × 10
M
s )[aldehyde][tryptamine] and
-
1
3
.2.4. Comparison of Solution and Enzymatic Reactions.
(1.67 M s ), respectively. Setting these equal to each other
3
2
Primary KIEs between 2 and 3 have often been reported for
yields [aldehyde][tryptamine] ) 2.7 × 10 M . Assuming that
the local concentrations of each reactant are equal, this analysis
gives an equivalent concentration of 52 M for each reactant in
the enzyme active site to achieve a rate equal to that of the
acid-catalyzed solution reaction. This qualitative analysis sug-
gests that increasing the local concentration of substrates and
catalytic acid group to promote rapid iminium formation is a
major mechanism of rate acceleration in strictosidine synthase.
The similar KIE values imply that the relative rates of the two
rate-controlling steps are similar in solution and enzyme-
catalyzed reactions. Therefore, to a first approximation, stric-
tosidine synthase does not appear to catalyze electrophilic
aromatic substitution any more than would be expected by an
increased local concentration of reactive species.
43
electrophilic aromatic substitution reactions; theoretical ex-
44
planations for these relatively small values have been reviewed.
However, in the case of the Pictet-Spengler reaction, the
presence of a second rate-controlling step additionally limits
full expression of the intrinsic primary KIE. The steps of
iminium formation and â-carboline formation have opposite
catalytic requirements; under conditions when the acid-catalyzed
step is faster, the base-catalyzed step becomes more rate limiting.
Given the presence of two competing steps, our extrapolated
maximal KIE of 2.7 in the nonenzymatic reaction describes the
conditions at which the final deprotonation is most rate
controlling and the protonation of carbinolamine step is
relatively fast.
D
3.2.5. Cocrystallization of Strictosidine Synthase with a
Bisubstrate Inhibitor. A key difference between the enzyme-
catalyzed and solution reaction is that strictosidine synthase not
only increases the local concentration of substrates but also acts
as a scaffold to appropriately orient the imnium intermediate
as it undergoes cyclization to produce a diastereoselective
product (Figure 1, step 4). While the KIE data, the pH
dependence of the rate, and the model solution reaction
described above can be used to make mechanistic predictions
about steps 3 and 5 (Figure 1), these data do not provide any
insight into step 4, which is when this key stereocenter is set.
Step 4 of the Pictet-Spengler reaction using indole amine
substrates also has an additional complication. These reactions
have been proposed to proceed through a spiroindolenine
intermediate resulting from attack from the 3-position (Figure
The observed enzymatic KIE V of 2.7 at the optimal pH of
the enzymatic reaction (pH 7) is a close match to the maximal
KIE observed in the nonenzymatic reaction. These similar
magnitudes suggest that the enzyme primarily catalyzes the first,
acid-catalyzed rate-limiting step of iminium formation. Steric
differences between propanal 4 and secologanin 2 do not appear
to significantly influence the KIE since a variety of aldehydes
gave similar KIE values for solution reactions (data not shown).
Nonchemical steps often contribute to reduced expression of
KIEs in enzymes.50 Given our current understanding of the
solution Pictet-Spengler reaction, the expression of a significant
KIE for deuterium substitution at the 2-position of tryptamine
indicates that the enzyme efficiently catalyzes carbinolamine
protonation.
Therefore, we propose that tryptamine 1 enters the active site
with the primary amine within hydrogen-bonding distance to
Glu309, as suggested by the crystal structure (2FPB) (Figure
1
3
1
, steps 4a and 4b), although direct attack by the 2-position
(
Figure 1, step 4) also appears to contribute to the mecha-
1
2,54,55
nism.
It is not clear which mechanism would be favored
2
). Tryptamine 1 is most likely protonated under the conditions
for the enzyme. For all of these reasons, we wished to further
examine step 4 in detail.
of the assay (pH 7) and, after entering the active site, transfers
its proton to deprotonated Glu309. This proton transfer is
consistent with the orientation of tryptamine relative to Glu309
in the crystal structure, and the importance of the tryptamine
protonation state to the mechanism is consistent with the loss
of enzyme activity at basic pH. After transferring a proton to
Glu309, tryptamine would then be nucleophilic enough to attack
the aldehyde substrate to rapidly generate the carbinolamine
species (Figure 1, step 2). The protonated Glu309 could then
act as the general acid that is required to protonate the
carbinolamine and catalyze the formation of the iminium species
through loss of water (Figure 1, step 3). Given that there are no
other ionizable residues in the vicinity (Figure 2), we propose
Structural analysis could provide insights into the model of
binding of the imine as it is oriented relative to carbons 2 and
3
during the cyclization step. However, the imine intermediate
is not sufficiently stable for cocrystallization with strictosidine
synthase. Therefore, we synthesized the reductive amination
product of tryptamine 1a and secologanin 2, compound 6, with
a flexible amine linkage between the indole and secologanin
moieties which could approximate the conformation of the
(
54) Casnati, G.; Dossena, A.; Pochini, A. Tetrahedron Lett. 1972, 5277-5280.
55) Kowalski, P.; Bojarski, A. J.; Mokrosz, J. L. Tetrahedron 1995, 51, 2737-
2742.
(
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 719

A R T I C L E S
Maresh et al.
Figure 6. (A) Structure of inhibitor 6. (B) Overlay of 6 (in green) with tryptamine 1a (in blue, from 2FPB) and secologanin 2 (in orange, from 2FPC). The
aldehyde of secologanin is highlighted with an orange arrow, and the amine of tryptamine is highlighted with a blue arrow. The indole moiety of 6 (from
2
VAQ) does not overlay well with 1a. Surrounding amino acids from 2FPB, 2FPC, and inhibitor structure (3.0 Å resolution) are shown for comparison
(tryptamine complex, 2FPB, cyan; secologanin complex, 2FPC, blue; inhibitor complex, 2VAQ, magenta). (C) Detail of the structure of inhibitor 6 as it
crystallized in the strictosidine synthase active site. The nucleophilic carbons 2 and 3 (highlighted by arrows) are distant from the electrophilic carbon of the
iminium species (highlighted by arrow). The methyl ester of inhibitor 6 is omitted for clarity. The structures of tryptamine 1a (in blue, from 2FPB) and
secologanin 2 (in orange, from 2FPC) are shown for comparison.
iminium. We demonstrated that 6 is a potent inhibitor of
strictosidine synthase with an IC50 value of 3 ( 0.5 nM (Figure
mechanism (Figure 7). Free energies were computed relative
to the trans form of the iminium species, the lowest-energy
ground-state structure according to the calculations. To facilitate
computational analysis, instead of secologanin 2, the calculations
were performed with a simplified structure (Figure 7; R ) Me
in Figure 1). For both 2-attack (Figure 1, step 4) and 3-attack
(Figure 1, step 4a), eight different transition states were found
which link either trans- or cis-iminium species with either five-
or six-membered rings via attack with the R-group oriented
either “over” or “off” of the initial ring during the attack (Figure
7). All transition-state species had a single imaginary frequency,
and the vibrational contribution to the free energy was included
at 300 K. Figure 7 shows the free energies of the stable species
and the lowest-energy transition states required to connect the
network of stable species. Transition states below a free energy
on the left axis are accessible from the trans-iminium species
on the corresponding time scales given on the right-axis.
Transition states of higher energy than the five shown provide
redundant connections in the network, and thus they have been
omitted for clarity (see Supporting Information for the other
transition-state structures).
The calculations indicate that interconversion from the
iminium (both cis and trans forms) directly to a six-membered
ring by 2-attack is several orders of magnitude faster than
interconversion from cis- or trans-iminium to five-membered
ring spiroindolenines. This suggests that the spiroindolenine
intermediate is more rarely visited in the mechanism. Further-
more, transition-state searches for the 1,2-shift between the six-
membered ring and the spiroindolenine searches repeatedly
resulted in transition states leading back to the iminium species.
For configurations that lie between the spiroindolenine and the
6
A).
Inhibitor 6 readily cocrystallized with strictosidine synthase
R. serpentina), and structural data of the complex were obtained
(
at 3 Å resolution. However, the electron density suggested that
the indole of 6 was bound in a different conformation compared
to the conformation when tryptamine alone was bound to
strictosidine synthase (2FPB) (Figure 6B). In contrast, the
secologanin moiety of the inhibitor overlaid well with the
previously reported enzyme-secologanin complex (Figure 6B).
The conformation of the inhibitor that was suggested by the
electron density was not representative of a productive cycliza-
tion state; neither carbon 2 (4.78 Å) or 3 (3.86 Å) of the inhibitor
was positioned close to the atom that corresponded to the
electrophilic iminium carbon (Figure 6C). These crystallography
data suggest that the tryptamine binding pocket can allow
binding modes that are not conducive to Pictet-Spengler
2
cyclization. We suspect that the rigidity of the sp hybrized
iminium moiety is critical for fostering a productive binding
orientation that can lead to cyclization, and that the more flexible
3
sp hybridized amine of the inhibitor can adopt conformations
inaccessible to the iminium intermediate.
3.2.6. Theoretical Calculations. Since structural analyses did
not provide insight into the orientation of the iminium involved
in the cyclization step, we turned to theoretical models of the
transition states on the pathway to step 4 (Figure 1). Ab initio
calculations were used to model the transition states of both
the spiroindolenine (Figure 1, step 4a) and the six-membered
ring (Figure 1, step 4) intermediates, and the free energies of
these transition states were compared to predict the favored
720 J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008

Mechanism of Strictosidine Synthase
A R T I C L E S
Figure 7. Free energy ∆F minima (represented by structures A, B, E, G, I, K) and transition states (represented by structures C, D, F, H, J in brackets)
plotted to show the connectivity of the free energy landscape. The timescales on the right are from transition-state theory and reflect the time scale for
passage through each column from the trans-iminium state. There is also a pathway from spiroindolenine A to six-membered ring K through a higher-energy
transition state 12.43 kcal/mol above the trans-iminium species.
six-membered ring, the calculations suggest that electron density
easily shifts from the migrating carbon to the nitrogen, resulting
in increased sp character at the migrating carbon center. Our
results suggest, but do not prove, the 1,2-shift does not occur
and that the spiroindolenine only rearranges to reform the
imnium species. However, the availability of a lower-energy
pathway from the iminium to a six-membered ring relative to
the iminium-spiroindolenine pathways suggests that, if a 1,2-
shift does exist, it would not significantly contribute to the
mechanism.
4. Summary
Strictosidine synthase stereoselectively constructs the central
alkaloid biosynthetic intermediate strictosidine 3. The Pictet-
Spengler reaction also occurs in solution, although without the
2
presence of a chiral catalyst the reaction is not stereoselec-
58,59
tive.
Several strategies were undertaken to probe the
mechanism of this enzymatic Pictet-Spengler reaction. First,
a primary KIE value was measured for both the enzymatic and
solution Pictet-Spengler reactions. The magnitude of the
enzymatic KIE was similar to the KIE measured for the solution
reaction catalyzed by acetic acid in the absence of solvent water
catalysis (kHA, Figure 5B). Therefore, deprotonation to form the
final product (Figure 1, step 5) appears to be rate controlling in
both solution and in the enzyme active site. Although we initially
expected cyclization to be rate controlling (Figure 1, step 4),
previously reported studies of electrophilic aromatic substitution
suggest that, as the rate of the reverse reaction increases relative
to the rate of deprotonation, the deprotonation becomes rate
These results are consistent with the empirical rule that
6-endo-ring closures are favored over 5-endo-trig cyclizations.
Moreover, the stereospecificity of the Pictet-Spengler reaction
can be rationalized by considering a direct attack mechanism
from the 2-position.56 Finally, the ab initio results confirm
previous conclusions of Kowalski et al. Their semiempirical
calculations did find a transition state for the 1,2-shift, but it
was of significantly higher energy than the transition state for
4
3,44
a direct 2-attack.1
2,55
controlling, and a large KIE is observed.
Notably, direct evidence for formation of
Interestingly, a primary KIE has been observed for the Pictet-
Spengler reaction catalyzed by norcoclaurine synthase, an
enzyme that utilizes a dopamine substrate instead of tryptamine
the spiroindolenine intermediate by isotopic scrambling has been
observed with several indole substrates.1
0,11,13,57
There is only
an 8 kcal/mol barrier to form the spiroindolenine intermediate,
6
0
1
a. This suggests that deprotonation is also rate controlling
so that if deprotonation (Figure 1, step 5) is slower than kT/h
in enzymatic Pictet-Spengler reactions that utilize non-indole
substrates. The authors speculated that norcoclaurine synthase
changes the rate-controlling step of the reaction, and that
rearomatization (i.e., Figure 1, deprotonation step 4) is not rate
controlling for nonenzymatic catalyzed reactions. However, our
nonenzymatic reactions show that, at least with indole amine
substrates, deprotonation is, in fact, rate controlling in the
solution Pictet-Spengler reaction, provided that a general acid
is available to ensure that iminium formation is rapid.
(-8/kT)
e
, accessing the spiroindolenine during the course of the
reaction could reasonably be expected to occur. Interestingly,
the free energy landscape shows that the spiroindolenine
(structure G, Figure 7) provides a relatively low-energy pathway
for interconversion of cis- and trans-iminium isomers. The
calculations however, suggest that the 1,2-shift from the
spiroindolenine to form the six-membered ring (Figure 1, step
4
b) does not occur; thus, access to the spiroindolenine is an
unproductive step in the Pictet-Spengler mechanism. Therefore,
while step 4a in Figure 1 is accessiblesthough less energetically
favorable than step 4sstep 4b does not appear to be accessible.
While the rate-controlling step of deprotonation is base-
(58) Seayad, J.; Seayad, A. M.; List, B. J. Am. Chem. Soc. 2006, 128, 1086-
1
087.
(
59) Taylor, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 2004, 126, 10558-10559.
(
56) Li, D.; Czerwinski, K.; Cook, J. M. Tetrahedron Lett. 1991, 32, 175-178.
57) Jackson, A. H.; Smith, P. Chem. Commun. 1967, 128, 264-266.
(60) Luk, L. Y. P.; Bunn, S.; Liscombe, D. K.; Facchini, P. J.; Tanner, M. E.
(
Biochemistry 2007, 46, 10153-10161.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 721

A R T I C L E S
Maresh et al.
Figure 8. Proposed mechanism of strictosidine synthase involving Glu309 and protonated tryptamine 1a.
catalyzed, the dependence of the solution reaction on pH and
buffer concentration suggests that the overall Pictet-Spengler
reaction is general acid-catalyzed. Based on the observed pKa
values calculated from the pH dependence curve of enzymatic
rate (Figure 4), the known residues surrounding the tryptamine
thereby supplying the proton for acid-catalysis (Figure 8). No
acid-base catalysis is required for the cyclization step during
which the stereocenter is formed (Figure 8), and it is likely that
the enzyme holds the iminium intermediate in the appropriate
conformation to achieve the correct diastereomer after cycliza-
tion. The nonproductive conformation of a flexible bisubstrate
inhibitor 6 bound in the enzyme crystal structure suggests that
1
a and secologanin 2 binding sites, and previously reported site-
14
directed mutagenesis studies, Glu309 appears to be the only
residue available for acid-base catalysis. Therefore, Glu309
likely plays the role of general acid catalyst in strictosidine
synthase. A number of possible ionizable groups, including the
tryptamine substrate or iminium intermediate, could account for
the observed basic pH dependence.
2
the enzyme also relies on the relatively rigid sp hybridized
iminium to ensure this intermediate is in a correct conformation
for cyclization (Figure 1, step 4). Ab initio calculations were
used to provide insight into this step of the mechanism. Energy
levels of theoretically predicted transition states suggest that
the transition state leading to the spiroindolenine intermediate
is higher in energy than the transition state resulting from direct
Although little mechanistic data is available for enzymatic
Pictet-Spengler catalysis, enzymatic iminium formationsthe
first part of the Pictet-Spengler reactionshas been well studied
in enzymes such as aldolases, acetoacetonate decarboxylase,
pyridoxal- and pyruvate-containing decarboxylases, and dehy-
dratases.53 Imine formation is a two-step process involving
addition of an amine to form a carbinolamine followed by
dehydration to the imine (Figure 1, steps 2-3). In these
enzymes, it is proposed that a catalytic residue, often a glutamic
2-attack. Moreover, since ab initio calculations suggest that the
1
,2-alkyl rearrangement of the spiroindolenine does not occur,
if the spiroindolenine is formed, it is most likely an unproductive
intermediate (Figure 8). It is difficult to predict which face of
attack from the indole is most favored by the constraints of the
enzyme active site. To facilitate the ab initio calculations, the
structures of the calculated transition states lacked the secolo-
ganin moiety; the absence of this group complicated attempts
to accurately dock the calculated structures into the active site.
Additionally, the structure of the enzyme-inhibitor 6 complex
illustrates that the indole moiety can bind in the active site in
more than one orientation, which further complicated docking
efforts. We speculate, although cannot prove, that strictosidine
synthase would be likely to favor one of the lower-energy
transition states. For example, if formation of the trans-iminium
isomer is catalyzed by the enzyme, then structure D (Figure 7)
would be a likely transition state. Additional bisubstrate
inhibitors with different linkages between the secologanin and
tryptamine moieties may crystallize in a “productive” conforma-
tion that could be compared and overlaid with the calculated
structures.
61,62
63
acid
or an ordered water molecule protonates the carbino-
lamine to catalyze dehydration to the iminium species. An
ordered active-site water molecule does not appear to be present
in the crystal structure of strictosidine synthase.
With the current evidence, we propose that Glu309 is
responsible for protonation of the carbinolamine and that the
resulting iminium formation is the key acid-catalyzed step of
the Pictet-Spengler reaction (Figure 8). Protonated tryptamine
can enter the active site and transfer its proton to Glu309,
(
(
(
61) Lorentzen, E.; Siebers, B.; Hensel, R.; Pohl, E. Biochemistry 2005, 44,
222-4229.
62) St-Jean, M.; Lafrance-Vanasse, J.; Liotard, B.; Sygusch, J. J. Biol. Chem.
4
2
005, 280, 27262-27270.
63) Fullerton, S. W. B.; Griffiths, J. S.; Merkel, A. B.; Cheriyan, M.; Wymer,
N. J.; Hutchins, M. J.; Fierke, C. A.; Toone, E. J.; Naismith, J. H. Bioorg.
Med. Chem. 2006, 14, 3002-3010.
722 J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008

Mechanism of Strictosidine Synthase
A R T I C L E S
It is possible that Glu309 is also responsible for the final
deprotonation step, since there do not appear to be any other
appropriate residues to perform this role (Figure 8). Glu309 is
near carbon 2 of tryptamine 1a as measured from the crystal
structure of the enzyme-tryptamine complex 2FPB (see Figure
substrates, along with a general acid-base catalyst, in high local
concentration to achieve efficient catalysis.
Data deposition: The atomic coordinates and structure factors
have been deposited in the Protein Data Bank, www.pdb.org
(PDB ID code 2VAQ [PDB], R2VAQSF (SRUCTURE FAC-
TORS)).
2, green line, 3.2 Å). Also, the careful kinetic analysis in Figure
4
B indicates that a carboxylate ion is sufficient to catalyze
Acknowledgment. Financial support was provided from MIT
to S.E.O. L.A.G. is supported by Ford Foundation and NSF
predoctoral fellowships. A.F. was supported by Deutscher
Akademischer Austausch Dienst (German Academic Exchange
Service). Deutsche Forschungsgemeinschaft (Bonn, Bad-Godes-
berg, Germany) to J.S. and S.P. Research was also supported
by the European Community (Research Infrastructure Action
under the FP6 “Structuring the European Research Area
Programme” contract No. RII3/CT/2004/5060008). The Bio-
physical Instrumentation facility (NSF-0070319 and GM68762),
the Chemistry Instrumentation Facility (NIH 1S10RR015926),
and the MIT/Harvard Center for Magnetic Resonance (NIH EB-
002026) at MIT are gratefully acknowledged. We gratefully
acknowledge Joseph Gregg for assistance in the preparation of
inhibitor 6.
deprotonation (kA).
Although the Pictet-Spengler reaction can occur spontane-
ously in solution, an enzymatic catalyst is always utilized in
alkaloid biosynthesis. This is partly due to the fact that very
little reaction would occur on a reasonable time scale at the
low (submillimolar) concentrations of secologanin and tryptamine
substrates in ViVo. Additionally, strictosidine synthase also
ensures that a single diastereomer is produced for subsequent
biosynthetic steps; notably, the R diastereomer of strictosidine
(vincoside), formed in approximately 60% yield in the solution
reaction (Figure S9A, Supporting Information) is not incorpo-
6
4
rated into downstream monoterpene indole alkaloids. The
mechanistic studies reported here strongly suggest that stricto-
sidine synthase does not significantly alter the mechanism of
the reaction as it occurs in solution under conditions of rapid
iminium formation. Notably, the Pictet-Spengler cyclization
is an example of an electrophilic aromatic substitution reaction
where the final deprotonation step is rate controlling, provided
that formation of the iminium is fast. The strictosidine synthase
catalyst has evolved to arrange the amine and aldehyde
Supporting Information Available: Figures S1-S14 and
Tables S1-S3: characterization of the transient intermediate
observed in the nonenzymatic reaction, representative solution
kinetics data, the buffer and pH dependence of reaction with
D
substrate 5, dependence of V on pH, calculated transition-state
structures with tabulated energies of all structures; full citation
of Gaussian reference. This material is available free of charge
via the Internet at http://pubs.acs.org.
(64) Heckendorf, A. H.; Hutchinson, C. R. Tetrahedron Lett. 1977, 48, 4153-
4
154.
JA077190Z
J. AM. CHEM. SOC.
9
VOL. 130, NO. 2, 2008 723