Synthesis of the Paralytic Shellfish Poisons (+)-Gonyautoxin 2, (+)-Gonyautoxin 3, and (+)-11,11-Dihydroxysaxitoxin

Article abstract of DOI:10.1021/jacs.6b02343

The paralytic shellfish poisons are a collection of guanidine-containing natural products that are biosynthesized by prokaryote and eukaryote marine organisms. These compounds bind and inhibit isoforms of the mammalian voltage-gated Na⁺ ion channel at concentrations ranging from 10^-11 to 10^-5 M. Here, we describe the de novo synthesis of three paralytic shellfish poisons, gonyautoxin 2, gonyautoxin 3, and 11,11-dihydroxysaxitoxin. Key steps include a diastereoselective Pictet-Spengler reaction and an intramolecular amination of an N-guanidyl pyrrole by a sulfonyl guanidine. The IC₅₀s of GTX 2, GTX 3, and 11,11-dhSTX have been measured against rat Na_V1.4, and are found to be 22 nM, 15 nM, and 2.2 ?M, respectively.

Full text of DOI:10.1021/jacs.6b02343

Article
pubs.acs.org/JACS
Synthesis of the Paralytic Shellfish Poisons (+)-Gonyautoxin 2,
(+)-Gonyautoxin 3, and (+)-11,11-Dihydroxysaxitoxin
John V. Mulcahy, James R. Walker, Jeffrey E. Merit, Alan Whitehead, and J. Du Bois*
Department of Chemistry, Stanford University, Stanford, California 94305-0080, United States
S
* Supporting Information
ABSTRACT: The paralytic shellfish poisons are a collection of
guanidine-containing natural products that are biosynthesized
by prokaryote and eukaryote marine organisms. These
compounds bind and inhibit isoforms of the mammalian
voltage-gated Na+ ion channel at concentrations ranging from
10−11 to 10−5 M. Here, we describe the de novo synthesis of
three paralytic shellfish poisons, gonyautoxin 2, gonyautoxin 3,
and 11,11-dihydroxysaxitoxin. Key steps include a diastereoselective Pictet−Spengler reaction and an intramolecular amination of
an N-guanidyl pyrrole by a sulfonyl guanidine. The IC50’s of GTX 2, GTX 3, and 11,11-dhSTX have been measured against rat
NaV1.4, and are found to be 22 nM, 15 nM, and 2.2 μM, respectively.
gonyautoxin 2 (GTX 2) and gonyautoxin 3 (GTX 3) after the
dinoflagellate from which they were isolated. The structures of
GTX 2 and GTX 3 were later revised by Schantz to the
corresponding 11α- and 11β-sulfate esters.10
Over the ensuing 30 years, additional PSPs have been
identified, all having in common a tricyclic, bis-guanidinium
core.2b,5c Electrophysiological and biochemical studies have
revealed that STX and GTX function by binding in the
extracellular pore of voltage-gated Na+ ion channels, making
contacts with the reentrant loops that form the Na+ ion
selectivity filter, and sterically blocking the ion permeation
pathway.11,12
The availability of STX, along with the functionally related
NaV pore-blocker tetrodotoxin (TTX), enabled the earliest
efforts to separate Na+ and K+ currents in intact nerve
fibers.2a,13 Our interest in these natural products follows from
their utility as chemical probes of NaVs and the potential of de
novo synthesis to deliver unique analogue structures to further
explore channel structure and cellular function.14 Despite the
elegant work of Kishi, Jacobi, Nagasawa, Nishikawa, Looper,
and others, access to the PSPs through asymmetric
construction has been exclusive to STX with one excep-
tion.15−20 Because of the difficulty of isolation and highly polar
character of these molecules, chemical modification of the
natural products has proven particularly challenging.2b,21 By
leveraging technologies developed in our lab for guanidine
assembly, we have devised an efficient route to a versatile
intermediate 1, which encompasses the tricyclic perhydro-
pyrrolopurine core of this family of neurotoxins (Figure 2).22
Elaboration of this structure has enabled preparation of three
naturally occurring PSPs, gonyautoxin 2 (GTX 2), gonyautoxin
3 (GTX 3), and 11,11-dihydroxysaxitoxin (11,11-dhSTX), the
INTRODUCTION
■
Voltage-gated Na+ ion channels (NaVs) are responsible for the
rising phase of action potentials, and are essential molecular
components for electrical transmission in neuronal cells.1 The
paralytic shellfish poisons (PSPs) are a family of small molecule
neurotoxins that occlude the outer pore of these channels and
inhibit Na+ ion flux.2 Detailed accounts of human intoxication
by these agents can be found from the 18th century.3
Researchers at the University of California correlated the
occurrence of certain toxic mussels with blooms of Alexandrium
catenella, a species of dinoflagellate algae, following an outbreak
in San Francisco over the summer of 1927.4 Since this
pioneering report, a collection of >50 structurally related small
molecule toxins has been isolated from a variety of
dinoflagellate and cyanobacterial sources (Figure 1).2b,5
Figure 1. Select naturally occurring paralytic shellfish poisons.
Structural elucidation, beginning with the elegant synthetic
work disclosed by Rapoport in 1962 and culminating in the X-
ray crystal structures published independently by Schantz and
Clardy, and Rapoport in 1975, first revealed the molecular
architecture of one poison, (+)-saxitoxin (STX).6,7
A
contemporaneous discovery by Shimizu identified two lethal
compounds from the hepatopancreases of contaminated clams,
which were initially assigned as epimeric C11-hydroxylated
forms of STX.8,9 These diastereomeric compounds were named
Received: March 7, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
aziridine intermediate 4, which would open in one of two ways
to afford either the C10- or the C12-substituted product. It is
equally plausible that pyrrole oxidation occurs through a dipole
intermediate in lieu of aziridine 4.25 Mechanistic considerations
notwithstanding, for our purposes either product is suitably
configured for subsequent elaboration of the π-bond to furnish
C11-derived STX targets (e.g., GTX 2/3, 11,11-dhSTX).
The ability to access pyrrole 3 efficiently from commodity
chemicals was an important factor in our strategic planning. By
capitalizing on the nucleophilic reactivity of the pyrrole group,
stereoselective formation of the bicyclic ring structure 6 could
be achieved through a Pictet−Spengler-type reaction (Figure
5).26 Analysis of transition state models for the cyclization
Figure 2. Divergent pathway to C11-derived guanidinium toxins from
intermediate 1.
syntheses of which are described below.23,24 Information gained
from these studies positions us to capitalize on 1 en route to
more complex PSP congeners as well as a varied collection of
modified toxin derivatives.
RETROSYNTHETIC ANALYSIS
■
Our desire to formulate a general synthetic approach to PSPs
led to the identification of 1, which could serve as a common
precursor to both natural and unnatural targets. Having
previously delineated a method to generate 2-iminoimidazoli-
dines through oxidation of aliphatic C−H bonds (Figure 3),
Figure 5. A proposed Pictet−Spengler reaction to access the desired
precursor for Rh-catalyzed aziridination.
event in which allylic strain is minimized across the C6−C5−
N7 bonds suggested that the stereochemistry at C5 would be
established in the requisite configuration. Accordingly, the N-
acyl pyrrole needed to test these ideas could be readily
assembled from commercial serine methyl ester.
Figure 3. Intramolecular amination with 2,2,2-trichloroethoxy-sulfony-
(tces)-protected guanidine substrates.
our analysis focused on the retrosynthetic disconnection of the
five-membered cyclic guanidine.22 In principle, compound 1
could be derived from 2,2,2-trichloroethoxysulfonyl (Tces)-
guanidine 2 by stereospecific amination of the C4−H bond
(Figure 4). Such a transformation, however, could also give an
DIASTEREOSELECTIVE PICTET−SPENGLER
■
CYCLIZATION
Synthesis of the guanidinium toxins commences from serine
ester 8, which is converted selectively to urea 9 using the acid
chloride derived from lithium pyrrolate (Scheme 1).27 Attempts
a
Scheme 1
a
Reagents and conditions: (a) pyrrole-1-carboxylic acid, oxalyl
chloride, cat. DMF, aq NaHCO3, THF, 90%; (b) t-BuPh2SiCl,
imidazole, DMF, 93%; (c) i-Bu2AlH, CH2Cl2, −90 °C.
Figure 4. Strategies to prepare the tricyclic core of the PSPs.
to prepare an analogous guanidine derivative were unsuccessful,
thus prompting a decision to advance the urea compound.
Fortunately, the conversion of urea groups to corresponding
guanidines or guanidinium salts finds ample precedent.15,28,29
Alcohol protection and low temperature ester reduction with
i-Bu2AlH furnishes aldehyde 10; this intermediate is used
immediately in the subsequent Pictet−Spengler cyclization.
Initial experiments to effect imine formation and ring closure
from aldehyde 10 employed allylamine in combination with
isomeric product as a result of competing oxidation of the C6−
H center. Our concerns regarding the control of chemo-
selectivity in this critical reaction prompted us to consider
alternative ideas, from which a pyrrole oxidation strategy
materialized.
Selective amination of 3 offers an attractive method for
assembling the tricyclic core of the PSPs with concomitant
incorporation of an oxygen group on the resulting pyrroline
ring. In principle, Rh-catalyzed oxidation could deliver an
1
Lewis acids. Monitoring of the reaction progress by H NMR
(CD2Cl2) revealed that allylamine addition to 10 occurs rapidly
B
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
and forms a diastereomeric mixture of hemiaminals. Addition of
3.5 equiv of BF3·OEt2 to this solution promotes iminium ion
formation and cyclization to cyclic urea 11. Other strong Lewis
acids, such as Sc(OTf)3, also proved effective in the
transformation. A single product is obtained from this reaction
mixture, which was confirmed by X-ray crystallography to be
the trans-isomer (Figure 6). Assuming the E-imine is formed
with NH4OAc at elevated temperature to yield the desired
guanidinium adduct. Following a similar sequence with urea 14,
however, did not furnish an isolable O-alkyl product. Increasing
reaction temperature, concentration, or time did nothing to
improve the outcome. We surmise on the basis of these
findings that the pyrrole group must strongly reduce the
nucleophilicity of the urea oxygen toward alkylation.
An alternative procedure for O-functionalization of cyclic
ureas appears in work by Overman.28 In this example,
MeOSO2CF3 is used as an alkylating agent in combination
with a sterically hindered base, 2,6-di-t-butyl-4-methylpyridine.
Only one equivalent of each reagent at a reaction concentration
of 0.1 M is required to generate the O-alkyl product in excellent
yield. Applying these conditions to urea 14, we identified trace
amounts of the desired O-methyl product when the reaction
was performed at room temperature. Elevating the reaction
temperature to 65 °C afforded a modest 33% yield of the O-
methyl product 15. The yield of 15 in this case was diminished
by the competing formation of the N1-methylated compound
(∼1:1 isomeric mixture). Reasoning that a more hindered
electrophile might improve reaction selectivity, EtOSO2CF3
was employed in place of the methyl derivative. This seemingly
small change gave an improved ∼3:1 ratio of the O- and N1-
ethyl products. Additional reaction optimization, which
included increasing the reaction concentration (1.0 M),
lowering the temperature to 37 °C, and altering the base
(2,4,6-tri-t-butylpyrimidine), further boosted product selectivity
such that 79% of the O-ethylated isomer could be isolated as
pure product (∼9:1 ratio O-ethyl/N1-ethyl). Following this
challenge, we were pleased to find that ethyl isourea 15 reacted
smoothly with methanolic ammonia at 70 °C (sealed tube) to
deliver the requisite bis-guanidine 16.
Figure 6. A diastereoselective Pictet−Spengler cyclization.
preferentially, minimization of allylic strain in a transition
structure such as that depicted in Figure 5 rationalizes favored
addition to the re-face of the activated imine. Related
diastereoselective Pictet−Spengler reactions performed on
amino acid-derived starting materials have found general use
in the synthesis of isoquinoline and indole alkaloids.26,30 This
work served as the inspiration for our plan.
Cyclic urea 11 possesses two of the three rings and two of
the three stereogenic centers that form the conserved core of
the PSPs. To install the Tces-guanidine in preparation for our
planned aziridination reaction, 11 is deallylated using Pd-
(PPh3)4 and N,N-dimethylbarbituric acid (Scheme 2).31
a
Scheme 2
INTRAMOLECULAR PYRROLE AMINATION
■
In our initial approach to GTX 2 and 3, the polar guanidinium
salt 16 was N-protected with trichloroacetyl chloride (TCACl).
This reaction proceeded efficiently and furnished 17 as a single
product (Scheme 2). Attempts to functionalize 16 with other
protecting groups (e.g., acetate, tert-butyloxycarbonyl) led to
mixtures of products, presumably due to competing acylation of
the N1 and N16 positions (vide infra). With the TCA-modified
material 17, cyclization to form the five-membered ring
proceeded under the action of catalytic Rh2(esp)2, PhI(OAc)2,
and MgO (Figure 7). The putative aziridine 18 was not
observed in this reaction; instead the resulting tricycle 20 was
obtained as the acetoxylated C10-N,O-acetal, the result of
a
Reagents and conditions: (a) 2 mol % Pd(PPh3)4, 1,3-dimethyl-
barbituric acid, CH2Cl2, then TcesNC(SMe)Cl 13, aq Na2CO3,
96%; (b) EtOTf, 2,4,6-tri-t-butylpyrimidine, CH2Cl2, 79%; (c) NH3,
NH4OAc, MeOH, 70 °C, 85%; (d) Cl3CC(O)Cl, i-Pr2NEt, CH2Cl2,
89%.
Following this operation, the unpurified primary amine is
then treated with thiomethylchloroimidate 13 and aqueous
Na2CO3 to give 14. This protocol represents an extension of
our previous work on Tces-guanidine synthesis, enabling
convenient, one-pot amine deprotection and isothiourea
formation.23
Conversion of the six-membered cyclic urea in 14 to the
corresponding guanidine proved surprisingly difficult given the
available precedent for analogous transformations.28 In their
synthesis of STX, Kishi et al. employed triethyloxonium
tetrafluoroborate to convert a five-membered urea to an O-
ethyl isourea.15 This intermediate was subsequently treated
Figure 7. Intramolecular pyrrole−guanidine oxidative cyclization
furnishes the tricyclic core of the PSPs.
C
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
AcOH incorporation. This product is stable to chromato-
graphic isolation and is generated as a single regio- and
diastereoisomer (62%).
(Figure 8), the same product obtained using more traditional
alkene dihydroxylation methods (e.g., cat. OsO4, NMO).
One plausible mechanism for amination of pyrrole 17 posits
the intermediacy of aziridine 18, which is displaced by AcOH in
a SN2′ addition to give 20 (Figure 7). We have confirmed,
however, through ROESY correlations that the stereochemistry
at C10 is S-configured, a result that appears to discount a
concerted ring-opening reaction. If the ring-strained aziridine
does form, it is likely that rapid opening would ensue to give a
zwitterionic species. It is also possible that this dipolar
intermediate 19 is generated directly in the amination
reaction.25 In either case, assuming that the stereochemical
configuration of the N,O-acetal is kinetically controlled,
directed addition of AcOH by the N9 anion would account
for the observed outcome. Alternatively, si-face addition of
AcOH at C10 may be the result of a stereoelectronic preference
for attack of the iminium ion antiperiplanar to the developing
lone pair on N3 (Stevens’-type model).32 While the
mechanistic underpinnings of this reaction remain opaque,
the formation of 20 through Rh-catalyzed oxidation represents
a crowning success for the application of this technology in
complex synthesis.
Attempts to transpose the acetate group in 20 from C10 to
C12 proved unsuccessful, a result that is perhaps unsurprising
in light of the thermodynamic stability differences between
isomeric dihydropyrroles. As such, a decision was made to
simply reduce N,O-acetal 20. This transformation is best
accomplished using BF3·OEt2 and triethylsilane. The tricyclic
core structure 21 obtained in this way is suitably disposed with
the alkene group at C11−C12 for elaboration to different PSP
targets.
Figure 8. Alkene dihydroxylation prevails over attempts to effect
single-step hydroxyketone formation.
The inability to promote single-step ketohydroxylation of 22
compelled us to identify an alternative, stepwise approach for
preparing 23. Dihydroxylation with 2 mol % OsO4 and N-
methylmorpholine-N-oxide (NMO) proceeded efficiently to
afford 24 as a single diastereomer, the result of addition from
the alkene face opposite the 5-membered guanidine (Figure 8).
The product stereochemistry was assigned on the basis of
analysis of ROESY correlations between N9−H and C12−H in
a related compound, and by comparison of the 1H NMR shifts
and coupling constants with corresponding spectral data for
11α- and 11β-hydroxysaxitoxin.34,35 Attempts were made to
perform a two-electron oxidation of this diol, reasoning that
even if the incorrect hydroxy ketone isomer was formed
initially, tautomerization would transform this intermediate to
the desired product 23. All conditions examined for this
oxidation (e.g., Dess−Martin periodinane, TEMPO/PhI-
(OAc)2, DMSO/(COCl)2), however, gave either unreacted
starting material or confounding product mixtures (Figure 9).
Ultimately, we concluded that masking the C11 hydroxyl group
would be necessary to achieve the desired oxidation reaction.
ELABORATION TO GTX 2 AND GTX 3
■
To complete the syntheses of GTX 2 and 3 from alkene 21, we
elected to first replace the silyl ether group at C13 with the
requisite carbamate (Scheme 3). These two steps were
a
Scheme 3
Figure 9. Initial efforts to oxidize diol 24 prove unsuccessful.
Our initial plan for blocking the C11-alcohol in diol 24
aimed to utilize a sulfate-derived protecting group. In a later
transformation, selective cleavage of a sulfate diester would
leave the C11-sulfate anion found in GTX 2 and 3.36 Attempts
to generate 25 were made using 2,2,2-trichloroethoxysulfonyl
chloride and the corresponding imidazolium salt (Table 1).37
Neither electrophile was sufficiently reactive to engage diol 24,
and only starting material was recovered from these experi-
ments. Searching for an alternative solution for selectively
blocking the C11−OH group, a collection of silylating and
acylating conditions was screened against 24 (and a related C13
silyl ether). From these experiments, the combination of
benzoyl cyanide and 4-dimethylaminopyridine was identified.
By employing this protocol, C11-benzoate 26 was obtained as a
single product in 69% yield with the remainder of the material
identified as unreacted diol and bis-acylated side products
(entry 8).
a
Reagents and conditions: (a) BF3·OEt2, Et3SiH, 83%; (b) n-Bu4NF,
THF; (c) Cl3CC(O)NCO, CH2Cl2, then MeOH, 76% (two
steps).
smoothly effected without intermediate purification of the
C13 alcohol. Conversion of carbamate 22 to any of the C11-
sulfated PSPs requires formal ketohydroxylation of the C11−
C12 olefin. Plietker and co-workers have described a Ru-
catalyzed method for alkene ketohydroxylation that employs
RuCl3 as the catalyst in combination with a stoichiometric
persulfate oxidant.33 In prior work from our lab to synthesize
STX, olefin ketohydroxylaton was achieved by employing
catalytic OsCl3 as a substitute for the ruthenium salt.20
Application of either ruthenium or osmium ketohydroxylation
conditions for oxidation of 22, however, gave only diol 24
Subsequent exposure of 26 to Dess−Martin periodinane
(DMP) efficiently transformed the C12 alcohol to the desired
ketone (Scheme 4). Product analysis by 13C NMR revealed that
ketone 27 exists in a fully hydrated form. Similarly, when 27 is
D
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
formed with this compound on Chinese Hamster Ovary Cells
expressing recombinant rat NaV1.4 and confirm that 11β-OH-
Table 1. Reaction Conditions Tested for Selective
Protection of the C11−OH
STX blocks sodium ion flux (IC50 = 9.3
0.7 nM against
rNaV1.4) with 3-fold reduced potency as compared to STX
(IC50 = 2.9 0.1 nM).11b
The synthesis of GTX 3 is completed upon treatment of
purified 11β-OH-STX with SO3·DMF and 2,6-di-t-butyl-4-
methylpyridine in N-methylpyrrolidinone. Other sulfating
reagents such as SO3·pyridine are not effective in this
transformation, affording only unreacted starting material. 2,6-
Di-t-butyl-4-methylpyridine was included as a sterically
hindered base to buffer the reaction mixture. Following our
discovery of these conditions, we became aware of a report by
Laycock in which GTX was desulfated and subsequently
resulfated using H2SO4 and N,N′-dicyclohexylcarbodiimide.40
These conditions are also effective, but in our hands give lower
conversion and isolated product yield (∼30%) than the SO3·
DMF procedure. Synthetic GTX 3 was purified by reversed-
phase HPLC and could be epimerized to a ∼3:1 mixture of
GTX 2/GTX 3 upon prolonged treatment with aqueous
NaOAc (0.3 M).9 Separation of the diastereomeric guanidi-
nium toxins was accomplished by RP-HPLC (Figure 10), and
entry
1
electrophile
base
temp (°C)
% yield
TcesCl
NEt3, DMAP
23
23
a
b
TcesNR2 OTf−
NMI
+
2
3
Et3SiCl
2,6-lutidine
pyridine
23
30−50
c
4
Ac2O
23
c
d
5
Ac2O
TTBP
40
trace
40
c
d
6
(n-PrCO)2O
(PhCO)2O
PhC(O)CN
TTBP
66
c
d
7
TTBP, DMAP
0
40
8
DMAP
−78
69
c
a
b
+
NR2 = N-methylimidazolium. NMI = N-methylimidazole. Experi-
ments performed on an analogous C13-OSitBuPh2 derivative in place
d
of 24. TTBP = 2,4,6-tri-t-butylpyrimidine.
a
Scheme 4
Figure 10. Epimerization of GTX 3 affords GTX 2 as a ∼1:3 mixture
at equilibrium. The two diastereomers are separable by RP-HPLC.
IC50 values were measured independently for each compound
against heterologously expressed rNaV1.4 (CHO cell). GTX 2
and GTX 3 have measured IC50 values of 22 1.7 and 15 2.1
nM, respectively, in agreement with literature reports of their
relative potency.41 To our knowledge, these studies complete
the first de novo synthesis of any of the sulfated gonyautoxins.
A preparation of GTX 3 by Nagasawa is the only other report
of this kind.17
a
Reagents and conditions: (a) Dess−Martin periodinane, CH2Cl2,
78%; (b) H2, cat. Pd/C, MeOH, CF3CO2H; (c) NH3, MeOH, 83%
(two steps); (d) DMF·SO3, 2,6-di-t-butyl-4-methylpyridine, 71%.
dissolved in d4-methanol, two diastereomeric hemiketal adducts
are generated. In d6-acetone, the 13C NMR shift of the C12
carbon appears at 98.4 ppm, indicating that even in organic
solvents, ketone hydration occurs due to the presence of
adventitious water.
MODIFICATION OF THE SYNTHETIC PLAN AIDS
ANALOGUE SYNTHESIS
■
One goal in developing a preparative synthesis of PSPs was to
make possible access to certain rare and difficult to isolate
natural toxins, such as 11,11-dihydroxySTX.24 A second
objective was to facilitate the preparation of PSP analogues to
enable structure−activity studies against sodium ion channel
subtypes.11,42 Both of these aims motivated our decision to
consider specific alterations to our first-generation synthesis of
GTX that would facilitate our scale-up efforts.
Alkene 22 (see Scheme 3) represents a versatile intermediate
for preparing both natural and unnatural PSPs. The ability to
manipulate this compound, however, using a number of
standard functional group transformations is impeded by
lability of the N16-trichloroacetyl moiety. Accordingly, an
alternative, more robust protecting group was sought.
Having addressed the problem of C12 oxidation, protecting
group cleavage and alcohol sulfation remained to complete the
preparation of GTX 3 from 27. Given the nature of the
blocking groups in 27, we wished to identify a one-pot method
for transforming 27 to C11β-hydroxysaxitoxin (11β-OH-STX).
Initial examination of Zn/AcOH for removing the Tces group,
however, was unsuccessful and gave complex product mixtures.
Clean deprotection of the Tces guanidine can be induced under
hydrogenolytic conditions (H2, cat. Pd/C) in acidic methanol,
conditions that also excise the trichloroacetyl group.36 Filtration
of the reaction mixture followed by treatment with methanolic
ammonia promotes cleavage of the C11-benzoate ester. This
latter reaction occurs quickly and must be carefully monitored;
acidification of the reaction mixture is necessary to minimize
base-promoted decomposition of the product.38 Isolation of the
bis-guanidinium salt by reversed-phase HPLC affords pure 11β-
OH-STX.39 Electrophysiological recordings have been per-
Successful pyrrole oxidation under the action of Rh2(esp)2
necessitates a strongly electron-withdrawing protecting group
on the six-membered guanidine unit. This type of substitution
mitigates a deleterious, uncatalyzed reaction between the
pyrrole unit and PhI(OAc)2. For this reason, a 2,2,2-
E
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
trichloroethoxycarbonyl (Troc) group was selected to replace
the trichloroacetyl in 17. Unlike trichloroacetyl installation,
however, treatment of 16 with TrocCl at 0 °C generated a
mixture of isomeric acyloxy-guanidines 28 and 29, which
favored the latter product (Table 2). Lowering the reaction
contaminated mussels gathered during an intense bloom of
Alexandium tamarense in Eastern Canada.24 These investiga-
tions relied on a new analytical technique, hydrophilic
interaction liquid chromatography and tandem mass spectrom-
etry (HILIC-MS/MS), which made possible the identification
of minuscule quantities of material. One of these isolates, 11,11-
dhSTX, is a most unusual structure bearing three contiguous
carbons at the ketone oxidation level (C4, C11, C12). The lack
of available material has precluded toxicity studies, and no
electrophysiological data measuring the potency of this
compound against NaV subtypes have been reported.
Table 2. Conditions for Selective Protection of N16
Synthesis of 11,11-dhSTX is possible in just two steps from
diol 32 (Figure 11). Initial removal of the guanidinium
entry
electrophile
base
temp (°C)
28/29
% yield
a
1
TrocCl
i-Pr2NEt
i-Pr2NEt
0
−78
23
1:5
1:10
1:1
84
80
80
88
75
2
TrocCl
b
3
TrocNR2+ −OTf
TrocNR′2+ −OTf
TrocNR′2+ −OTf
c
4
23
6:1
c
5
55
>20:1
a
b
+
TrocCl = Cl3CCH2OC(O)Cl. NR2 = N-methylimidazolium.
NR′2 = N-methyl-2-phenylbenzimidazolium.
c
+
temperature to −78 °C marginally increased the ratio in favor
of the N1-Troc product 29. By switching from TrocCl to a
Troc-imidazolium or Troc-benzimidazolium salt, the N16-Troc
compound could be formed preferentially.43 The product ratio
was improved through further reaction optimization, which
included raising the reaction temperature to 55 °C.
Identification of conditions to favor 28 was necessary, as the
N1-Troc isomer appears to be quite labile toward hydrolysis
and, more importantly, does not engage effectively in the Rh-
catalyzed pyrrole oxidation. As such, the ability to invert
selectivity in the Troc-protection of 16 was a critical advance.
Dirhodium-catalyzed amination of 28, while lower yielding
than the equivalent transformation of the trichloroacetyl
substrate 17 (41% vs 62%, respectively), is easily performed
on multigram scale to afford large quantities of the desired
tricyclic product (Scheme 5). The same reactions conducted
with 20 and all subsequent trichloroacetyl intermediates can be
performed with the equivalent Troc-derivatives and are
generally more efficient. The availability of 30 has enabled us
to assemble a number of unique PSP analogues, the structures
and activities of which will be disclosed in a future publication.
Figure 11. Elaboration of diol 32 to 11,11-dihydroxysaxitoxin.
protecting groups under hydrogenolytic conditions affords
bis-guanidnium salt 33. In previous work from our lab, we have
employed DMSO and N,N′-dicyclohexylcarbodiimide (DCC),
conditions originally described by Schantz, for C12-alcohol
oxidation on bis-guanidium salts analogous to 33. Using this
combination of reagents, the diol group in 33 is doubly
oxidized to give the C11,12-bis-hemiketal.20,44 Although the
isolated yield of 11,11-dhSTX is low, the reaction itself is
reasonably efficient based on NMR analysis of the unpurified
sample (∼30% conversion). The product is weakly chromo-
phoric (at best) and thus presents a significant challenge for
chromatographic purification on a conventional HPLC instru-
ment. Nevertheless, the analytic data we have obtained on a
pure sample are entirely consistent with those reported in the
literature.24 In addition, we have determined the IC50 of this
compound against rNaV1.4 (CHO cells) to be 2.2 0.2 μM, a
value that is seemingly consistent for a PSP derivative that is
speculated to be a product of detoxification by the host
shellfish. Additional experiments are in progress to measure the
affinity of 11,11-dhSTX against other NaV isoforms and mutant
channels.
PREPARATION OF
(+)-11,11-DIHYDROXYSAXITOXIN
■
A recent report by Quilliam and co-workers describes the
isolation and structure elucidation of three new PSPs from
a
Scheme 5
a
Reagents and conditions: (a) 10 mol % Rh2(esp)2, PhI(OAc)2, MgO, CH2Cl2, 40 °C, 41%; (b) BF3·OEt2, Et3SiH, CH2Cl2, 35 °C, 72%; (c)
nBu4NF, THF; (d) Cl3CC(O)NCO, CH2Cl2, then MeOH, 91% (two steps); (e) OsO4, NMO, THF, 83%; (f) C6H5C(O)CN, DMAP, 61%; (g)
Dess−Martin periodinane, 91%.
F
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(7) (a) Bordner, J.; Thiessen, W. E.; Bates, H. A.; Rapoport, H. J. Am.
Chem. Soc. 1975, 97, 6008−6012. (b) Schantz, E. J.; Ghazarossian, V.
E.; Schnoes, H. K.; Strong, F. M.; Springer, J. P.; Pezzanite, J. O.;
Clardy, J. J. Am. Chem. Soc. 1975, 97, 1238−1239.
CONCLUSION
■
De novo synthesis of natural toxins, including gonyautoxin 2
and 3, and 11,11-dihydroxysaxitoxin has been successfully
achieved through an intermediate that is available in gram
quantities and in just nine steps from L-serine. Highlights of this
route include diastereoselective Pictet−Spengler cyclization and
pyrrole amination reactions to form the tricyclic core structure
common to all PSPs. These chemistries are enabling access to
unique PSP derivatives designed to probe molecular ligand−
receptor interactions with the voltage-gated sodium channel.
(8) Shimizu, Y.; Alam, M.; Oshima, Y.; Fallon, W. E. Biochem.
Biophys. Res. Commun. 1975, 66, 731−737.
(9) Shimizu, Y.; Buckley, L. J.; Alam, M.; Oshima, Y.; Fallon, W. E.;
Kasai, H.; Miura, I.; Gullo, V. P.; Nakanishi, K. J. Am. Chem. Soc. 1976,
98, 5414−5416.
(10) Boyer, G. L.; Schantz, E. J.; Schnoes, H. K. J. Chem. Soc., Chem.
Commun. 1978, 889−890.
(11) (a) Thomas-Tran, R.; Du Bois, J. Proc. Natl. Acad. Sci. U.S.A.
2016, in press. (b) Walker, J. R.; Novick, P. A.; Parsons, W. H.;
McGregor, M.; Zablocki, J.; Pande, V. S.; Du Bois, J. Proc. Natl. Acad.
Sci. U. S. A. 2012, 109, 18102−18107 and references therein..
(12) (a) Choudhary, G.; Shang, L.; Li, X.; Dudley, S. C. Biophys. J.
2002, 83, 912−919. (b) Penzotti, J. L.; Fozzard, H. A.; Lipkind, G. M.;
Dudley, S. C. Biophys. J. 1998, 75, 2647−2657. (c) Lipkind, G. M.;
Fozzard, H. A. Biophys. J. 1994, 66, 1−13.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.6b02343.
■
S
Complete experimental procedures and characterization
(PDF)
(13) Narahashi, T.; Moore, J. W.; Scott, W. R. J. Gen. Physiol. 1964,
47, 965−974.
AUTHOR INFORMATION
Corresponding Author
*jdubois@stanford.edu
Notes
The authors declare the following competing financial
interest(s): John Mulcahy and Justin Du Bois are cofounders
and own equity shares in SiteOne Therapeutics, a pharma-
ceutical startup company aimed at developing sodium channel
selective inhibitors as anti-nociceptive agents.
(14) (a) Parsons, W. H.; Du Bois, J. J. Am. Chem. Soc. 2013, 135,
10582−10585. (b) Ondrus, A. E.; Lee, H.-L. D.; Iwanaga, S.; Parsons,
W. H.; Andresen, B. M.; Moerner, W. E.; Du Bois, J. Chem. Biol. 2012,
19, 902−912. (c) Andresen, B. M.; Du Bois, J. J. Am. Chem. Soc. 2009,
131, 12524−12525.
(15) Tanino, H.; Nakata, T.; Kaneko, T.; Kishi, Y. J. Am. Chem. Soc.
1977, 99, 2818−2819.
(16) (a) Jacobi, P. A.; Martinelli, M. J.; Slovenko, P. J. Am. Chem. Soc.
1984, 106, 5594−5598. (b) Jacobi, P. A.; Brownstein, A.; Martinelli,
M.; Grozinger, K. J. Am. Chem. Soc. 1981, 103, 239−241.
(17) (a) Iwamoto, O.; Shinohara, R.; Nagasawa, K. Chem. - Asian J.
2009, 4, 277−285. (b) Iwamoto, O.; Koshino, H.; Hashizume, D.;
Nagasawa, K. Angew. Chem., Int. Ed. 2007, 46, 8625−8628.
(18) Sawayama, Y.; Nishikawa, T. Angew. Chem., Int. Ed. 2011, 50,
7176−7178.
(19) Bhonde, V. R.; Looper, R. E. J. Am. Chem. Soc. 2011, 133,
20172−20174.
(20) (a) Fleming, J. J.; McReynolds, M. D.; Du Bois, J. J. Am. Chem.
Soc. 2007, 129, 9964−9975. (b) Fleming, J. J.; Du Bois, J. J. Am. Chem.
Soc. 2006, 128, 3926−3927.
(21) Robillot, C.; Kineavy, D.; Burnell, J.; Llewellyn, J. E. Toxicon
2009, 53, 460−465.
(22) (a) Kim, M.; Mulcahy, J. V.; Espino, C. G.; Du Bois, J. Org. Lett.
2006, 8, 1073−1076. (b) Burkhard, J. A.; Du Bois, J., manuscript in
preparation.
■
ACKNOWLEDGMENTS
■
Brian Andresen and Rhiannon Thomas-Tran are graciously
acknowledged for assistance with electrophysiological record-
ings. This work has been supported by a grant from the NIH
(R01 NS045684) and generous gifts from Pfizer, Amgen,
Boehringer-Ingelheim, and GlaxoSmithKline. J.V.M. and A.W.
were the recipients of fellowships from the National Science
Foundation and American Cancer Society, respectively.
REFERENCES
■
(1) (a) Voltage Gated Sodium Channels, Handbook of Experimental
Pharmacology 221; Reuben, P. C., Ed.; Springer-Verlag: Berlin, 2014.
(b) Dib-Hajj, S.; Priestly, T. In Ion Channels: From Structure to
Function, 2nd ed.; Kew, J. N. C., Davies, C. H., Eds.; Oxford University
Press: Oxford, 2010; pp 131−169. (c) Hille, B. Ion Channels of
Excitable Membranes, 3rd ed.; Sinauer Associates: Sunderland, MA,
2001.
(2) (a) Kao, C. Y., Levinson, S. R., Eds. Tetrodotoxin, Saxitoxin, and
the Molecular Biology of the Sodium Channel. Annals of the New York
Academy of Sciences; New York Academy of Sciences: New York, 1986;
Vol. 479. (b) Thottumkara, A. P.; Parsons, W. H.; Du Bois, J. Angew.
Chem., Int. Ed. 2014, 53, 5760−5784. (c) Moczydlowski, E. G. Toxicon
2013, 63, 165−183.
(23) Mulcahy, J. V.; Du Bois, J. J. Am. Chem. Soc. 2008, 130, 12630−
12631.
(24) Dell’Aversano, C.; Walter, J. A.; Burton, I. W.; Stirling, D. J.;
Fattorusso, E.; Quilliam, M. A. J. Nat. Prod. 2008, 71, 1518−1523.
(25) (a) Padwa, A.; Flick, A. C.; Leverett, C. A.; Stengel, T. J. Org.
Chem. 2004, 69, 6377−6386. (b) Padwa, A.; Stengel, T. Org. Lett.
2002, 4, 2137−2139. Also see: Shibuta, T.; Sato, S.; Shibuya, M.;
Kanoh, N.; Taniguchi, T.; Monde, K.; Iwabuchi, Y. Heterocycles 2014,
89, 631−639.
(26) For a general review on Pictet−Spengler reactions, see:
(a) Stockigt, J.; Antonchick, A. P.; Wu, F.; Waldmann, H. Angew.
̈
(3) (a) Quayle, D. B. Paralytic Shellfish Poisoning in British Columbia;
Fisheries Research Board of Canada: Canada, 1969; Vol. 168.
(b) Vancouver, G. A Voyage of Discovery to the North Pacific Ocean,
and Round The World; London, 1798; Vol. 2.
Chem., Int. Ed. 2011, 50, 8538−8564. (b) Cox, E. D.; Cook, J. M.
Chem. Rev. 1995, 95, 1797−1842.
(27) Boger, D. L.; Patel, M. J. Org. Chem. 1987, 52, 2319−2323.
(28) (a) Coffey, D. S.; McDonald, A. I.; Overman, L. E.; Rabinowitz,
M. H.; Renhowe, P. A. J. Am. Chem. Soc. 2000, 122, 4893−4903.
(b) Overman, L. E.; Rabinowitz, M. H.; Renhowe, P. A. J. Am. Chem.
Soc. 1995, 117, 2657−2658.
(4) (a) Covell, W. P.; Whedon, W. F. Arch. Pathol. 1937, 24, 411−
418. (b) Sommer, H.; Meyer, K. F. Arch. Pathol. 1937, 24, 560−598.
(5) (a) Cusick, K. D.; Sayler, G. S. Mar. Drugs 2013, 11, 991−1018.
(b) Wiese, M.; D’Agostino, P. M.; Mihali, T. K.; Moffitt, M. C.; Neilan,
B. A. Mar. Drugs 2010, 8, 2185−2211. (c) Llewellyn, L. E. Nat. Prod.
Rep. 2006, 23, 200−222.
(29) For a general discussion of guanidine synthesis, see: (a) Zhang,
W.-X.; Xu, L.; Xi, Z. Chem. Commun. 2015, 51, 254−265. (b) Maki, T.;
Tsuritani, T.; Yasukata, T. Org. Lett. 2014, 16, 1868−1871 and
references therein..
(6) Schuett, W.; Rapoport, H. J. Am. Chem. Soc. 1962, 84, 2266−
2267.
G
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(30) (a) Myers, A. G.; Kung, D. W. J. Am. Chem. Soc. 1999, 121,
10828−10829. (b) Cox, E. D.; Hamaker, L. K.; Li, J.; Yu, P.;
Czerwinski, K. M.; Deng, L.; Bennett, D. W.; Cook, J. M. J. Org. Chem.
1997, 62, 44−61. (c) Waldmann, H.; Schmidt, G.; Jansen, M.; Geb, J.
Tetrahedron Lett. 1993, 34, 5867−5870.
(31) Garo-Helion, F.; Merzouk, A.; Guibe,
6109−6113.
́
F. J. Org. Chem. 1993, 58,
(32) (a) Stevens, R. V. In Strategies and Tactics in Organic Synthesis;
Lindberg, T., Ed.; Academic Press: Orlando, FL, 1984; Vol. 1, pp 275−
300. (b) Stevens, R. V. Acc. Chem. Res. 1984, 17, 289−296. Also see:
Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.; Woerpel, K.
A. J. Am. Chem. Soc. 2005, 127, 10879−10884.
(33) (a) Plietker, B. Tetrahedron: Asymmetry 2005, 16, 3453−3459.
(b) Plietker, B. J. Org. Chem. 2003, 68, 7123−7125.
(34) A correlation between N9−H and C12−H is observed in the
ROESY spectrum of diol 33.
(35) Wichmann, C. F.; Boyer, G. L.; Divan, C. L.; Schantz, E. J.;
Schnoes, H. K. Tetrahedron Lett. 1981, 22, 1941−1944.
(36) Liu, Y.; Lien, I. F. F.; Ruttgaizer, S.; Dove, P.; Taylor, S. D. Org.
Lett. 2004, 6, 209−212.
(37) (a) Ingram, L. J.; Desoky, A.; Ali, A. M.; Taylor, S. D. J. Org.
Chem. 2009, 74, 6479−6485. (b) Ingram, L. J.; Taylor, S. D. Angew.
Chem., Int. Ed. 2006, 45, 3503−3506.
(38) Our observation that the fully deprotected form of 27
decomposes under basic conditions is consistent with other accounts
of the stability of PSPs, see: Wong, J. L.; Brown, M. S.; Matsumoto, K.;
Oesterlin, R.; Rapoport, H. J. Am. Chem. Soc. 1971, 93, 4633−4644.
(39) Shibue, M.; Mant, C. T.; Hodges, R. S. J. Chromatogr. A 2005,
1080, 68−75.
(40) Laycock, M. V.; Kralovec, J.; Richards, R. Nat. Toxins 1997, 5,
36−42.
(41) (a) Alonso, E.; Alfonso, A.; Vietyes, M. R.; Botana, L. M. Arch.
Toxicol. 2016, 90, 479−488. (b) Frace, A. M.; Hall, S.; Brodwick, M.
S.; Eaton, D. C. Am. J. Physiol. 1986, 251, C159−C166.
(42) (a) Dib-Hajj, S. D.; Cummins, T. R.; Black, J. A.; Waxman, S. G.
Annu. Rev. Neurosci. 2010, 33, 325−347. (b) England, S.; Groot, M. J.
Br. J. Pharmacol. 2009, 158, 1413−1425. (c) George, A. L., Jr. J. Clin.
Invest. 2005, 115, 1990−1999.
(43) Watkins, B. E.; Kiely, J. S.; Rapoport, H. J. Am. Chem. Soc. 1982,
104, 5702−5708.
(44) Koehn, F. E.; Ghazarossian, V. E.; Schantz, E. J.; Schnoes, H. K.;
Strong, F. M. Bioorg. Chem. 1981, 10, 412−428.
H
DOI: 10.1021/jacs.6b02343
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX