Investigation of the possible pharmacologically active forms of the nicotinic acetylcholine receptor agonist anabaseine

Article abstract of DOI:10.3390/md17110614

Three major forms of the nicotinic agonist toxin anabaseine (cyclic iminium, cyclic imine and the monocationic open-chain ammonium-ketone) co-exist in almost equal concentrations at physiological pH.We asked the question: Which of these forms is pharmacologically active? First, we investigated the pH dependence of anabaseine inhibition of [³H]-methylcarbamylcholine binding at rat brain α4β2 nicotinic acetylcholine receptors (nAChRs). These experiments indicated that one or both monocationic forms interact with the orthosteric binding site for ACh. However, since they occur at equal concentrations near physiological pH, we employed another approach, preparing a stable analog of each form and examining its agonist activities and binding affinities at several vertebrate brain and neuromuscular nAChRs. Only 2-(3-pyridyl)-1,4,5,6-tetrahydropyrimidine monohydrogen chloride (PTHP), the cyclic iminium analog, displayed nAChR potencies and binding affinities similar to anabaseine. The cyclic imine analog 2,3'-bipyridyl and the open-chain ammonium-ketone analog 5-methylamino-1-(3-pyridyl)-1-pentanone (MAPP), displayed ≤1% of the activity predicted if the one form was solely active. The lower potency of weakly basic 2,3'-bipyridyl can be explained by the presence of a small concentration of its monocationic form. Since the open chain ammonium-ketone monocationic form of anabaseine has some structural similarity to the neurotransmitter GABA, we also tested the ability of anabaseine and its 1,2-dehydropyrrolidinyl analog myosmine to activate a mammalian GABA_A receptor, but no activity was detected. We conclude that the monocationic cyclic iminium is the form which avidly binds and activates vertebrate nAChRs.

Full text of DOI:10.3390/md17110614

marine drugs
Article
Investigation of the Possible Pharmacologically
Active Forms of the Nicotinic Acetylcholine Receptor
Agonist Anabaseine
Kristin Andrud 1,†, Hong Xing 1,†, Bjarne Gabrielsen 2, Linda Bloom 3, Vladimir Mahnir 1,
Stephen Lee 4, Benedict T. Green 4, Jon Lindstrom 5 and William Kem 1,
*
1
Department of Pharmacology and Therapeutics, College of Medicine, University of Florida, Gainesville,
FL 32610, USA; Kristin.Andrud@du.edu (K.A.); hong.xing@ufl.edu (H.X.);
wrkem@pharmacology.lufl.edu (V.M.)
Department of Chemistry, University of Florida, Gainesville, FL 32610, USA; bmlmpunta@yahoo.com
Department of Biochemistry and Molecular Biology, College of Medicine, University of Florida, Gainesville,
FL 32610; USA; lbloom@ufl.edu
USDA-ARS Poisonous Plant Research Laboratory, Logan, UT 84341, USA; Stephen.Lee@ars.usda.gov (S.L.);
Ben.Green@ars.usda.gov (B.T.G.)
Department of Neuroscience, University of Pennsylvania, Philadelphia, PA 19104, USA;
jslkk@mail.med.upenn.edu
2
3
4
5
*
Correspondence: wrkem@ufl.edu; Tel.: +1-352-392-0669
†
Denotes first authors.
ꢀꢁꢂꢀꢃꢄꢅꢆꢇ
ꢀꢁꢂꢃꢄꢅꢆ
Received: 23 August 2019; Accepted: 24 October 2019; Published: 29 October 2019
Abstract: Three major forms of the nicotinic agonist toxin anabaseine (cyclic iminium, cyclic imine
and the monocationic open-chain ammonium-ketone) co-exist in almost equal concentrations at
physiological pH. We asked the question: Which of these forms is pharmacologically active? First,
we investigated the pH dependence of anabaseine inhibition of [3H]-methylcarbamylcholine binding
at rat brain α4β2 nicotinic acetylcholine receptors (nAChRs). These experiments indicated that one or
both monocationic forms interact with the orthosteric binding site for ACh. However, since they occur
at equal concentrations near physiological pH, we employed another approach, preparing a stable
analog of each form and examining its agonist activities and binding affinities at several vertebrate
brain and neuromuscular nAChRs. Only 2-(3-pyridyl)-1,4,5,6-tetrahydropyrimidine monohydrogen
chloride (PTHP), the cyclic iminium analog, displayed nAChR potencies and binding affinities similar
to anabaseine. The cyclic imine analog 2,30-bipyridyl and the open-chain ammonium-ketone analog
5-methylamino-1-(3-pyridyl)-1-pentanone (MAPP), displayed 1% of the activity predicted if the one
≤
form was solely active. The lower potency of weakly basic 2,30-bipyridyl can be explained by the
presence of a small concentration of its monocationic form. Since the open chain ammonium-ketone
monocationic form of anabaseine has some structural similarity to the neurotransmitter GABA,
we also tested the ability of anabaseine and its 1,2-dehydropyrrolidinyl analog myosmine to activate
a mammalian GABAA receptor, but no activity was detected. We conclude that the monocationic
cyclic iminium is the form which avidly binds and activates vertebrate nAChRs.
Keywords: acetylcholine; alkaloid; anabaseine; bipyridyl; cholinergic; nicotine; nicotinic acetylcholine
receptor; ring-chain tautomerism; toxin
1. Introduction
Over eighty years ago, the Belgian pharmacologist Z. M. Bacq discovered the presence of a
nicotine-like substance in a marine worm, the Atlantic hoplonemertine Amphiporus lactifloreus [1,2].
Mar. Drugs 2019, 17, 614; doi:10.3390/md17110614
www.mdpi.com/journal/marinedrugs

Mar. Drugs 2019, 17, 614
2 of 12
Small amounts of a weakly basic compound(s), called “amphiporine”, were partially purified using
classical solvent extraction methods, but isolation by crystallization with standard alkaloid precipitating
salts like picric acid was unsuccessful [
properties similar to nicotine, was eventually isolated from the Pacific hoplonemertine Paranemertes
peregina and detected in A. lactifloreus [ ]. A piperideine analog of the tobacco alkaloid anabaseine,
anabaseine also occurs in certain ants [ ]. It is a potent agonist at a variety of nicotinic acetylcholine
receptors, particularly the vertebrate skeletal muscle and 7 neuronal nAChRs that display high
affinities for the snake toxin -bungarotoxin. Anabaseine is also a weak partial agonist at 2 nAChRs
that modulate cognitive and addiction neuronal circuits in the brain [8].
The imine bond of anabaseine, besides increasing its interaction with certain nAChRs, has enabled
this molecule to serve as a useful “lead” compound for designing drug candidates that target
nAChRs [ 11]. 3-(2,4-dimethoxybenzylidene)anabaseine (also called DMXBA and GTS-21), the most
studied anabaseine drug candidate, enhances cognition in humans [12 13], reduces inflammation [14],
inhibits skeletal muscle wasting [15 16] and is neuroprotective in Parkinson’s [17] and Alzheimer’s
3]. Anabaseine, an alkaloid with chemical and pharmacological
4–6
7
α
α
α4β
α
7
9
–
,
,
disease [18] animal models. New anabaseine compounds that are more potent and selective
agonists have been reported [19,20].
α7 nAChR
Most nAChR agonists and competitive antagonists possess a cationic moiety that tightly binds
within an electronegative “aromatic box” of the receptor formed by five aromatic amino acid side
chains [21
experiments where the pH bathing intact skeletal muscle cells is varied [23
approach assumes that altering pH mainly affects ionization of the ligand, rather than that of the nAChR.
Subsequently several laboratories reported myriad effects of pH on several nAChR subtypes [26 27].
,
22]. The dominant role of the monocationic form of nicotine has been demonstrated by
–
25]. Such an experimental
,
Altering extracellular pH was also found to affect muscle chloride permeability and resting membrane
resistance, which indirectly affect response to nicotinic agonists [28,29].
Whereas anabaseine contains a moderately basic (pKa 8.7, [30]) secondary amine group and
mainly exists as a monocation at physiological pH, cyclic anabaseine contains a tertiary amine group
but exists in three different chemical forms at neutral pH (Figure 1): the unionized cyclic imine (I),
cationic cyclic iminium (I+) and monocationic open-chain ammonium ketone (AK+) forms are in
dynamic equilibrium at approximately equal concentrations at pH 7.4 [31,32]. Our initial experiments
on the pH dependence of anabaseine interaction with a brain nAChR, presented below, suggest that
one or both monocationic forms are most active. However, since the two monocationic forms coexist
at equal concentrations regardless of pH, a different experimental approach to assess their separate
activities was required. Thus, we prepared a stable analog for each form and studied its effects on
several vertebrate nAChRs: two fetal-type skeletal muscle (electric fish and human TE671 myosarcoma)
receptors and the two major brain subtypes,
cationic forms of anabaseine and its five-membered ring analog myosmine have some resemblance to
another neurotransmitter, -aminobutyric acid (GABA), we also investigated whether anabaseine and
its myosmine can activate a GABAA receptor.
α4β2 and α7. Since the open chain ammonium ketone
γ

Mar. Drugs 2019, 17, 614
3 of 12
Figure 1. A: Structures of the three major forms of anabaseine that exist in dynamic equilibrium under
physiological conditions; B: Anabaseine analogs synthesized to represent the particular form shown
above. MAPP (also called N-methyl anabaseine) can still occur as a cyclic imine or iminium, but the
equilibrium greatly (≥ 20-fold) favors the ammonium-ketone monocation form.
2. Results
Initially, we measured the pH-dependent binding of [3H]-methylcarbamylcholine ([3H]-MCC)
to rat brain 2 nAChRs. In the absence of anabaseine we found a large (approximately 3-fold)
α4β
enhancement of [3H]-MCC binding as pH increased from 6.0 to 7.8. While interesting in its own
regard, since MCC is permanently ionized, this marked pH dependence of [3H]-MCC nAChR affinity
complicated the analysis of the effect of pH on anabaseine binding, which was indirectly measured
by [3H]-MCC displacement. Anabaseine inhibition of [3H]-MCC binding at each pH in Figure 2 is
expressed with respect to the binding of [3H]-MCC alone at the same pH, so as to compensate for
[3H]-MCC binding pH dependence. If only one or both monocationic forms of anabaseine are active,
one would anticipate that inhibition of [3H]-MCC binding by the fixed concentration of anabaseine
would be greatest at the low pHs and progressively decrease as the pH approaches and exceeds the
pKa of its tertiary amine (iminium) group. However, if only the cyclic imine form binds with high
affinity one would expect the opposite effect, an increase in the inhibition of [3H]-MCC binding as the
pH approaches and exceeds the tertiary amine pKa. Anabaseine displacement of [3H]-MCC binding
did decrease with increased pH, as expected for an active monocationic form, but the decrease was
quantitatively less than would be predicted. Going from 5.8 to 7.8, the unionized cyclic imine (I)
concentration was predicted [31] to increase approximately 20-fold, but the concentration of each
monocationic form (I+ or AK+) would decrease ~4-fold. The pH dependence of anabaseine inhibition
of MCC binding was not as marked as predicted for the monocationic form, but was certainly more
consistent with the hypothesis that one or more monocationic forms of anabaseine is binding to this
brain receptor. One additional complication in the interpretation of the results in Figure 2 is that specific
receptor binding (and its inhibition) is not necessarily proportional to the displacing ligand (anabaseine)
concentration, as it depends on a saturable binding isotherm (For instance, a Michaelis-Menton equation
relating competing ligand concentrations and their respective equilibrium dissociation constants,

Mar. Drugs 2019, 17, 614
4 of 12
which may not have the same pH dependence. The concentration of [3H]-MCC was 10 nM, which
was almost the same as the Kd (11 nM) of this radioligand with rat brain membranes, under our
experimental conditions [5].
Since the two monocationic forms show the same pH dependence and the same concentration
when expressed as % of total anabaseine, another approach was required to assess their activities
separately. We resorted to synthesizing and testing stable analogs that resemble the three anabaseine
forms. PTHP, the tetrahydropyrimidinyl analog of the cyclic iminium form, is predicted to be >99%
ionized at pH 7.4 and does not undergo ring opening within the physiological pH range considered
here, since the basic (pKa >10) amidinium group, unlike the imine group, is not susceptible to hydration
and subsequent ring opening. 2,30-bipyridyl (pKa 4.4) was selected as an analog that possesses the
basic structure of the unionized cyclic imine; at pH 7.4 >99.9% of the molecules of 2,30-bipyridyl
will be unionized. The monocationic ammonium-ketone open-chain form analog selected was
5-methylamino-1-(3-pyridyl)-1-pentanone, MAPP, shown in Figure 1; it mainly (>99%) exists in their
open-chain form in water at physiological pH; its cyclic imine or iminium form, N-methylanabaseine,
is present in very small concentration ( 5%) [31–33]. An unionized open-chain amino-ketone form
≤
which occurs at high pHs [32] can be neglected due to its very small predicted concentrations over the
pH range 5.8 to 7.8 that is of interest here.
Figure 2. Concentrations of the three major forms of anabaseine as a function of pH as determined
by UV spectrophotometry [32]. Left Ordinate: Concentration of each form expressed as % of the
total anabaseine concentration. The red curve is the estimated cyclic imine concentration and the
blue curve is the concentration of the cyclic iminium or the monocationic ammonium-ketone form
of anabaseine (Assuming that KH = 1.0, the concentrations of these ionized forms are equal). Right
ordinate: pH dependence of anabaseine inhibition of [3H]-MCC binding to rat brain
α
4β2 receptors
(Standard error bars included). Anabaseine inhibition (mean
replicate measurements.
±
SEM) at each pH was the average of six
PTHP was the only anabaseine analog that approached anabaseine in potency and binding affinity.
Its potency and affinity for both the frog and human neuromuscular receptors was slightly inferior
to that of anabaseine, but its agonist activity on vertebrate nAChRs had not been reported (Please
see Figures 3 and 4). The other analogs displayed less than 1% of anabaseine’s potency and binding
affinity. The small potency of 2,30-bipyridyl can be interpreted as being entirely due to its monocationic
form, since the pKa of the most basic nitrogen on the 20pyridyl ring is 4.4 [34].

Mar. Drugs 2019, 17, 614
5 of 12
Figure 3. Activation of the human
α4β2 neuronal nAChR expressed in tsA201 cells as measured by
membrane depolarization using the FlexStation assay. Each point generally is the mean of four separate
measurements. All responses were normalized with respect to the response of the cells to 0.5 µM
epibatidine. Potency and efficacy estimates from these curves are found in Table 1.
Figure 4. Activation of the human fetal (TE671 cells) skeletal muscle nAChR by the various stable
anabaseine analogs measured by membrane depolarization using the FlexStation assay. Each point is
generally the mean of four separate measurements. Responses were normalized to the 5 µM epibatidine
response. Potency and efficacy estimates from these curves are found in Table 1.
GABAA receptors are homologous structures that have a common ancestry with nAChRs.
Structure–activity studies on GABAA receptors have shown that both the ammonium and carboxy
groups are not absolutely required for agonist activity and the intermediate structure between these
two ionizable groups can vary considerably as well—in some cases flexibility is minimized by a ring
structure, such as in the agonist muscimol, and in other analogs the hydrocarbon chain is shortened or

Mar. Drugs 2019, 17, 614
6 of 12
lengthened one methylene unit without loss of agonist activity. Since the open-chain cationic form
of anabaseine resembles GABA by having a primary ammonium group that is separated a similar
distance from a carbonyl group, we tested anabaseine and its pyrroline homolog myosmine on a human
GABAA receptor heterologously expressed in a HEK cell line [35]. Figure 5 shows that anabaseine and
myosmine are not GABAA agonists, at least on this particular GABAA receptor subtype.
Table 1. Potencies and affinities characterizing the interactions of anabaseine and its stable analogs
with several neuronal and neuromuscular nAChRs.
Neuronal nAChR
Neuromuscular nAChR
Human
α4β2
EC50
Rat
Rat
Human
TE671
EC50
Torpedo
Compound
α4β
α7
Electric Organ
1
2
3
µM:
Ki
Ki
IC50
Anabaseine
PTHP
2.85 ± 1.0
3.64 ± 1.2
>1000
0.096 ± 0.01
0.38 ± 0.07
>50
1.87 ± 0.10
1.54 ± 0.76
>2300
1.2 ± 0.0
1.7 ± 0.3
660 ± 60
>1000
0.29 ± 0.18
0.47 ± 0.01
343 ± 48
>500
2,30-Bipyridyl
MAPP
>2000
10
>2000
1 Determined by displacement of [3H]-cytisine; 2 Determined by displacement of [125I]
α
-bungarotoxin; 3 Determined
by inhibition of the rate of binding of [125I]-α-bungarotoxin.
3. Discussion
While most nAChR agonists and antagonists binding at ACh orthosteric sites are basic molecules
with at least one ionizable N, there are exceptions, including lophotoxin [36], the neonicotinoids [37 38
,
]
and some recently synthesized pyrimidine agonists [39]. To optimize a molecule to serve as a drug or
selective molecular probe of some receptor it is important to identify the form in which it interacts
optimally with its target. In the case of anabaseine, our identification of the cyclic iminium form
as the most active (and possibly the only) form suggests that structural modifications that enhance
the basicity of its tetrahydropyridyl nitrogen will enhance potency and possibly selectivity for the
intended receptor target. In fact, a major improvement in stability occurs with addition at the 3-position
of the tetrahydropyridyl ring of an electron-conjugated system containing an aromatic ring, as in
3-(4-dimethylaminobenzylidene)-anabaseine [40] or 3-(2,4-dimethoxybenzylidene)-anabaseine, also
called GTS-21 [41]. These anabaseine derivatives do not undergo ring opening (hydrolysis of the imine
bond) under physiological conditions, due to their extended π electron conjugation [42].
Previously, it was assumed (but not experimentally tested) that the cyclic iminium form of
anabaseine is the one which binds avidly to the ACh binding site in nAChRs [8]. This form was
modeled to fit into a muscle type homology model [43]. The crystal structure of the molluscan
acetylcholine binding protein binding anabaseine [44] actually was derived from crystals containing
both the cyclic iminium and the open-chain monocationic form of anabaseine occupying some of the
five identical binding sites. Unlike the cyclic iminium, the ammonium-ketone form ammonium group
did not insert into the “aromatic box.” In the preparation of the AChBP-anabaseine crystals very high
anabaseine concentrations were used so that most of the five sites would be occupied; therefore the
presence of the ammonium-ketone form in the crystal structure does not necessarily indicate that this
form of anabaseine would bind to AChBPs and nAChRs at the low concentrations of anabaseine that
occur in our in vitro functional and binding studies.
The ability of a variety of 2-(aryl)-1,4,5,6-tetrahydropyrmidines to block neuromuscular nAChRs
and produce teratogenic effects in chick embryos was reported long ago [45,46]. Some of these
compounds were reported to be nAChR agonists. The only study of PTHP on neurons was that of
Upshall et al. [47] on leech Retzius cells. Our results also indicate that PTHP is a potent agonist at
neuronal as well as neuromuscular nAChRs. The 2-aryl compounds reported in the Brimblecombe
et al. study did not include PTHP or other compounds having a 2-(3-pyridyl)- substituent, which is
known to be essential for the agonist activity of nicotine and of anabaseine. For PTHP to be equipotent

Mar. Drugs 2019, 17, 614
7 of 12
with anabaseine, based on the I+ form being the only active form, its EC50 would need to be ~1/3
of the EC50 of anabaseine, since the anabaseine (I+) is ~1/3 of its total concentration. The data for
PTHP and anabaseine in Figures 3 and 4 and in Table 1, indicate that PTHP has approximately 30%
of the potency predicted for the anabaseine cyclic iminium form. One possible explanation for this
slightly inferior potency is that the additional N on the tetrahydropyrmidinyl ring must form an H
bond with some other atom within the aromatic box of the nAChR orthosteric binding site and this
may be energetically unfavorable.
Figure 5. Investigation of possible activation of the human GABAA receptor measured by membrane
depolarization using a FlexStation assay. Each point is the mean of three separate measurements.
All responses were normalized with respect to the response of the cells to 1 mM µM GABA. The EC50
value for GABA was 3.0 µM (95% confidence interval = 1.8–5.1 µM).
While 2,30-bipyridyl displayed only a marginal potency at vertebrate receptors that could be
predicted from its pKa, it has been found to be quite paralytic to insects [48] and crustaceans [49 50].
,
At least some arthropod nAChRs are known to not require a cationic ligand for activation, as shown by
many studies with neonicotinoid insecticides such as imidachloprid [37]. These neonicotinoids also
display a limited activity at vertebrate nAChRs [38]. It seems likely that 2,30-bipyridyl affects some
arthropod nAChRs through its unionized form.
One can speculate as to what advantage(s) might accrue to an organism (in this case,
a hoplonemertine or ant) to produce and secrete an imine-bond containing toxin like anabaseine.
The entire body wall integument as well as the contiguous epithelium of the anterior proboscis of
P. peregrina contains very high concentrations of anabaseine [5]. The proboscis everts and wraps
around the prey during its capture allowing extensive contact of this epithelium with its annelid prey.
The mineralized stylet of the proboscis can be observed to produce multiple punctures of the prey
integument, which should further facilitate envenomation. In addition, potential predators will also be
exposed in these ways to anabaseine and related alkaloids. Perhaps a major advantage of possessing
an imine bond, besides an enhanced affinity for the nAChR orthosteric site [8], is its low basicity,
which allows the unionized form to reach a relatively high concentration in the neutral pH range of
the marine environment and thereby enhance passive diffusion across the integument of the prey or
predator. The integument of a marine organism exposed to sea water is likely to have a pH near 8,
well above the pH range in which anabaseine is primarily ionized. Thus, the toxin has a much better
chance of entering the prey/predator than would anabaseine, the more basic secondary amine analog
of anabaseine.
Another consideration is how anabaseine is packaged within the venom gland cells [51]. It seems
most likely that it is accumulated within secretory vesicles, which are generally acidic (pH 5–7).
The ionized forms, especially the very polar open-chain forms, would be much less likely to diffuse
out of the vesicles than would the cyclic imine form, so their dominance under these acidic conditions
should enhance the energy efficiency of anabaseine storage.

Mar. Drugs 2019, 17, 614
8 of 12
Thus, it is possible that each of the coexisting forms of anabaseine have functional significance,
not only for understanding how this potent nAChR agonist interacts with its receptors, but they also
may play different roles in storage and release of the toxin by the nemertine.
We have synthesized other anabaseine compounds whose properties will be reported in the future
(Kem et al., in preparation). The current paper predicts that their relative pharmacological potencies
will be affected by the extent that the cyclic iminium form dominates the equilibrium between these
different forms.
4. Materials and Methods
4.1. Compound Syntheses
2-(3-pyridyl)-1,4,5,6-tetrahydropyrimidinehydrochloride(PTHP):3-pyridylamidinemonohydrochloride
(20 g, 0.127 mol) was dissolved in warmed absolute EtOH. To this solution was slowly added (dropwise
over 5 min) freshly distilled 1,3-diaminopropane (9.42 g, 0.13 mol). This solution was refluxed for 6 h in
a 500 mL round-bottomed flask. Afterwards the entire solution was placed on a steam bath to dryness.
The crude solid was dissolved in approximately 500 mL hot isopropanol; crystals formed as the solvent
◦
was chilled to near 0 C (6.2 g, 21.1% yield, MP 273–275◦). Additional recrystallizations from the
mother liquor raised the yield. Elemental Analysis: Found: % C = 54.71; % H = 6.17; % N = 21.22%.
C9H12N3Cl requires: % C = 54.69; % H = 6.12; % N = 21.26. NMR Spectral data: In CD3OD, referenced
to CD3OD at 3.34 ppm:
δ (ppm): 8.85–9.19 (2H, m, H-2, 6; 8.23–8.55 (1H, dd, H-4); 7.85–7.93 (1H, m,
H-5); 4.80 (2H, broad singlet, CD3OH); 3.64–3.87 (4H, t 4,6-CH2-); 1.97–2.43 (2H, quintet, 3-CH2-).
3-Pyridylamidine hydrochloride (PTHP Precursor): Our procedure was similar to that described
by Schaefer and Peters [52] and Brown and Evans [53]. 3-cyanopyridine (35 g, 0.34 mol) was dissolved
in 300 mL MeOH containing sodium methoxide (1.82 g, 0.034 mol). The flask was fitted with a
condenser equipped with a drying tube and the clear solution stirred at 25 ◦C for 24 h. Ammonium
chloride (19.8 g, 0.37 mol), added according to and the mixture was refluxed and stirred for 24 h.
Not all (~30%) of the ammonium chloride dissolved. After cooling the solution to room temperature
and removing excess NH4Cl by filtration the resulting solution was rotary evaporated and the product
was dissolved in 100 mL hot isopropanol and filtered hot, leaving the NH4Cl behind. After 24 h at 0 ◦C
the crystals were collected by filtration. Two more recrystallizations using isopropanol provided a
total of 30 g product (MP 187–190 ◦C, 57% of theoretical yield). Elemental Analysis: Found, % C =
45.60; % H = 5.17/ % N = 26.61; C6H8N3Cl requires: % C = 45.73; % H = 5.12; % N = 26.66). NMR
Spectral data: In CD3OD, referenced to CD3OD at 3.34 ppm:
δ (ppm): 9.0 (2H, m, H-2, H-5,6 coupling
of 5 H observed); 8.38–8.63 (1H, dd, H-4), H-4,5 coupling of 8 Hz observed, as is H-2,4 coupling of
1.5 Hz); 7.7–8.0 (1H, m, H-5); 5.3 (CD3OH, 4H, broad).
Anabaseine and 5-methylamino-1-(3-pyridyl)-1-pentanone (MAPP) dihydrochlorides (Figure 1)
were synthesized as previously reported [31,
32]. 2,30-Bipyridyl was synthesized by silver acetate
oxidation of anabaseine obtained from Sigma-Aldrich, St. Louis, MO, USA [49].
4.2. Radioligand Binding Assays
[3H]-carbamylcholine ([3H]-MCC) binding assays with rat brain membranes were carried out
as previously reported with a final concentration of 0.5 nM [3H]-MCC [
8
,
54]; 2.8
sulfate (Sigma-Aldrich) was added to inhibit binding to muscarinic receptors. [3H]-Cytisine and
125I]- -bungarotoxin ([125I]-
-BTX) were used in binding experiments with whole rat brain and TE671
cell membranes, respectively, according to Kem et al. [41]. Rat brain membranes (200 g of protein)
on TsA201 cell membranes (100 g protein) expressing human 2 nAChRs were incubated with
0.5 nM [3H]-cytisine in a final volume of 500 L of binding saline for 4 h at 5 ◦C. The experiments on
nAChRs rat brain membranes involved incubation with 0.5 nM [125I]- -Btx for 3 h at 37 ◦C to assure that
µM atropine
[
α
α
µ
µ
α4β
µ
α7
α
equilibrium was reached. Eight different concentrations of the experimental compound were usually
tested in triplicate. Nonspecific binding was measured in the presence of 1 mM (S)-nicotine hydrogen

Mar. Drugs 2019, 17, 614
9 of 12
tartrate (Sigma-Aldrich). Data were fitted using GraphPad Prism software (Version 4 GraphPad
Software, San Diego, CA, USA) by nonlinear regression analyses to a sigmoidal one-site model with
variable slope. Compound affinity for Torpedo electric organ membrane nAChRs was assessed by
α α-Btx binds
inhibition of 0.5 nM [125I]- -BTX binding over a 30 min incubation period at 5 ◦C, since
irreversibly to skeletal muscle nAChRs. The IC50 was also estimated by the same Prism software.
4.3. Cell Culture
TsA201 cells expressing human
α
4
β2 were maintained in media consisting of Dulbecco’s Modified
Eagle medium supplemented with 10% FBS, 100 units/mL penicillin and 100
µg/mL streptomycin,
2 mM l-glutamine, 0.5 mg/mL zeocin and 0.6 mg/mL geneticin [55]. Cells were grown in 75 cm2 culture
flasks, which were housed in a humidified incubator (Fisher Scientific, Atlanta, GA, USA at 37 ◦C in an
atmosphere of 5% CO2. They were grown to around 80–90% confluence after harvesting with 0.25%
trypsin and being split weekly at a subcultivation ratio of between 1:6 and 1:10.
4.4. Nicotinic Receptor FlexStation Functional Assays
Our experimental protocol was based on the initial study of Fitch et al. [56]. Cells were seeded at
a density of roughly 5
with 50 g /ml poly-d-lysine hydrobromide (Sigma-Aldrich, 70–150 kDa) and grown overnight in
100 L culture medium. A proprietary membrane potential dye obtained from Molecular Devices (San
Diego, CA, USA) was prepared by dissolving one bottle of dye into 30 mL of Hanks Saline (pH = 7.4)
containing 20 mM HEPES buffer. The cells were incubated with 100
L of dye for 30 min at 37 ◦C prior
×
104 to 105 cells/well in 96-well flat-bottom black wall culture plates coated
µ
µ
µ
to the robotically controlled concentration-response experiment. Serial dilutions of a compound for
dose-response analysis were prepared in 96-well plates by evaporation of a methanolic stock solution
and then reconstituted in the appropriate volume of Hanks saline. Fluid transfer and readings were
performed by a FlexStation fluorimeter (Molecular Devices). Excitation and emission wavelengths
were set to 530 nm and 565 nm with a cutoff of 550 nm. The first 17 s were used as a basal reading.
At 18 s, a test compound was added to determine the EC50, followed by addition of 25
final concentration) at 160 s to serve as a fluorescence calibrant. Compounds that had no measurable
depolarizing activity on the TsA201 cells were then tested for their ability to inhibit a 5 M control
µL KCl (40 mM
µ
ACh response. Data were fitted and graphed with Prism (GraphPad) to determine the EC50.
4.5. GABA Receptor Flexstation Funcional Assays
WSS1 cells from ATCC (Manassas, VA, USA) which express functional GABAA receptors were
cultured as previously described by Wong et al. [57]. The cells were grown to near confluence and
then transferred to black well, clear bottom 96-well assay plates (Corning Incorporated, Corning, NY,
USA) 12–18 h prior to the assay. Molecular Devices blue dye (Sunnyvale, CA, USA) was reconstituted
as previously described [58]. The WSS1 cells were allowed to come to room temperature, the culture
medium was then aspirated, replaced with the blue dye and equilibrated for 30 min before being placed
on a FlexStation 3 plate reader (Molecular Devices, Sunnyvale, CA, USA). Varying concentrations of
γ
-aminobutyric acid (GABA, Tocris Bioscience, Bristol, United Kingdom), myosmine (Sigma, St. Louis,
MO, USA) and anabaseine [31 32] in blue dye solution were added to the cells by the FlexStation.
Programmed readings and data analysis were performed as previously described [58 59]. The cellular
,
,
responses to GABA were normalized to the maximum dye response generated by 1 mM GABA.
The fifty percent effective concentration for GABA was determined using a sigmoidal dose-response
equation (log(agonist) vs. normalized response—Variable slope) with Prism version 6.03 (GraphPad
Software, Company, San Diego, CA, USA, www.graphpad.com).
Author Contributions: K.A. carried out the FlexStation assays; H.X. and V.M. carried out radioligand binding
assays; B.G. synthesized PTHP; L.B. synthesized compounds (anabaseine, MAPP, DMAPP) and calculated the
predicted anabaseine concentrations shown in Figure 2; B.T.G. and S.L. tested anabaseine and myosmine on

Mar. Drugs 2019, 17, 614
10 of 12
GABA receptor expressing cells; J.L. provided TsA201 cells expressing human
α
4
β
2 nAChRs; W.K. directed the
investigation and wrote the manuscript. All authors read and commented on the manuscript.
Funding: This research was funded by Taiho Pharmaceuticals Company (Tokyo) and a Florida Sea Grant
(R/LR-MB-20) to W.R. Kem.
Acknowledgments: We greatly appreciate the interest and contributions of our late collaborator John Zoltewicz
in the early part of this investigation.
Conflicts of Interest: The authors declare no conflicts of interest.
References
1.
2.
3.
Bacq, Z.M. Les poisons des nemertiens. Bull. Cl. Sci. Acad. R. Belg. 1936, 22, 1072–1079.
Bacq, Z.M. L’Amphiporine et La N mertine Poisons Des Vers N mertiens. Arch. Int. Physiol. 1937, 44, 190–204.
King, H. Amphiporine, an active base from the marine worm Amphiporus lactifloreus. J. Chem. Soc. Lond.
1939, 1365.
é
é
4.
5.
6.
7.
8.
9.
Kem, W.R. A Chemical Investigation of Nemertine Toxins. Ph.D. Thesis, University of Illinois at
Urbana-Champaign, Champaign County, IL, USA, 1969; p. 116.
Kem, W.R. A study of the occurrence of anabaseine in Paranemertes and other nemertines. Toxicon 1971
9, 23–32. [CrossRef]
Kem, W.R.; Abbott, B.C.; Coates, R.M. Isolation and structure of a hoplonemertine toxin. Toxicon 1971
9, 15–22. [CrossRef]
Wheeler, J.W.; Olubajo, O.; Storm, C.B. Anabaseine: Venom alkaloid of Aphaenogaster ants. Science 1981
211, 1051–1052. [CrossRef]
,
,
,
Kem, W.R.; Mahnir, V.M.; Papke, R.L.; Lingle, C.J. Anabaseine is a potent agonist on muscle and neuronal
alpha-bungarotoxin-sensitive nicotinic receptors. J. Pharm. Exp. Ther. 1997, 283, 979–992.
Kem, W.R. The brain alpha7 nicotinic receptor may be an important therapeutic target for the treatment of
Alzheimer’s disease: Studies with DMXBA (GTS-21). Behav. Brain Res. 2000, 113, 169–183. [CrossRef]
10. Slavov, S.H.; Radzvilovits, M.; LeFrancois, S.; Stoyanova-Slavova, I.B.; Soti, F.; Kem, W.R.; Katritzky, A.R.
A computational study of the binding of 3-(arylidene) anabaseines to two major brain nicotinic acetylcholine
receptors and to the acetylcholine binding protein. Eur. J. Med. Chem. 2010, 45, 2433–2446. [CrossRef]
11. Kem, W.; Soti, F.; Wildeboer, K.; LeFrancois, S.; MacDougall, K.; Wei, D.-Q.; Chou, K.-C.; Arias, H.
The nemertine toxin anabaseine and its derivative DMXBA (GTS-21): Chemical and pharmacological
properties. Mar. Drugs 2006, 4, 255–273. [CrossRef]
12. Freedman, R.; Olincy, A.; Buchanan, R.W.; Harris, J.G.; Gold, J.M.; Johnson, L.; Allensworth, D.;
Guzman-Bonilla, A.; Clement, B.; Ball, M.P.; et al. Initial phase 2 trial of a nicotinic agonist in schizophrenia.
Am. J. Psychiatry 2008, 165, 1040–1047. [CrossRef]
13. Kem, W.R.; Olincy, A.; Johnson, L.; Harris, J.; Wagner, B.D.; Buchanan, R.W.; Christians, U.; Freedman, R.
Pharmacokinetic Limitations on Effects of an Alpha7-Nicotinic Receptor Agonist in Schizophrenia: Randomized
Trial with an Extended-Release Formulation. Neuropsychopharmacology 2018, 43, 583–589. [CrossRef]
14. Pavlov, V.A.; Ochani, M.; Yang, L.H.; Gallowitsch-Puerta, M.; Ochani, K.; Lin, X.; Levi, J.; Parrish, W.R.;
Rosas-Ballina, M.; Czura, C.J.; et al. Selective alpha7-nicotinic acetylcholine receptor agonist GTS-21 improves
survival in murine endotoxemia and severe sepsis. Crit. Care Med. 2007, 35, 1139–1144. [CrossRef]
15. Kashiwagi, S.; Khan, M.A.; Yasuhara, S.; Goto, T.; Kem, W.R.; Tompkins, R.G.; Kaneki, M.; Martyn, J.A.
Prevention of burn-induced inflammatory responses and muscle wasting by GTS-21, a specific agonist for
α7 nicotinic acetylcholine receptors. Shock 2017, 47, 61–69. [CrossRef]
16. Schaller, S.J.; Nagashima, M.; Schönfelder, M.; Sasakawa, T.; Schulz, F.; Khan, M.A.S.; Kem, W.R.; Schneider, G.;
Schlegel, J.; Lewald, H.; et al. GTS-21 attenuates loss of body mass, muscle mass, and function in rats having
systemic inflammation with and without disuse atrophy. Pflüg. Arch. Eur. J. Physiol. 2018, 470, 1647–1657. [CrossRef]
17. Suzuki, S.; Kawamata, J.; Matsushita, T.; Matsumura, A.; Hisahara, S.; Takata, K.; Kitamura, Y.;
Kem, W.R.; Shimohama, S. 3-[(2,4-Dimethoxy) benzylidene]-anabaseine dihydrochloride prevents
6-hydroxydopamine-induced Parkinsonism neurodegeneration through α7 nicotinic acetylcholine receptor
stimulation in rats. J. Neurosci. Res. 2012, 91, 462–471. [CrossRef]

Mar. Drugs 2019, 17, 614
11 of 12
18. Takata, K.; Amamiya, T.; Mizoguchi, H.; Kawanishi, S.; Kuroda, E.; Kitamura, R.; Ito, A.; Saito, Y.; Tawa, M.;
Nagasawa, T.; et al. Alpha7 nicotinic acetylcholine receptor-specific agonist DMXBA (GTS-21) attenuates A
β
accumulation through suppression of neuronal -secretase activity and promotion of microglial amyloid-
γ
β
phagocytosis and ameliorates cognitive impairment in a mouse model of Alzheimer’s disease. Neurobiol. Aging
2018, 62, 197–209.
19. Kem, W.R.; Soti, F. Alpha7 Nicotinic Receptor Selective Ligands. U.S. Patent 8,592,458, 14 December 2006.
20. Kem, W.R.; Soti, F.; Xing, H. Bicyclic Tetrahydropyridyl Alpha7 Selective Agonists. U.S. Patent 9,150,558,
6 October 2015.
21. Changeux, J.P. The nicotinic acetylcholine receptor: The founding father of the pentameric ligand-gated ion
channel superfamily. J. Bio. Chem. 2012, 287, 40207–40215. [CrossRef]
22. Brejc, K.; van Dijk, W.J.; Klaassen, R.V.; Schuurman, M.; van Der Oost, J.; Smit, A.B.; Sixma, T.K. Crystal
structure of an ACh-binding protein reveals the ligand-binding domain of nicotinic receptors. Nature 2001
411, 269–276. [CrossRef]
,
23. Barlow, R.B.; Hamilton, J.T. Effects of pH on the activity of nicotine and nicotine monomethiodide on the rat
diaphragm preparation. Br. J. Pharmacol. Chemother. 1962, 18, 543–549. [CrossRef]
24. Hamilton, J.T. The influence of pH on the activity of nicotine at the neuromuscular junction. Can. J.
Biochem. Physiol. 1963, 41, 283–289. [CrossRef] [PubMed]
25. Bartels, E.; Podleski, T.R. Action of nicotine on the electroplax and difference of potency between ionized
and unionized forms. Biochim. Biophys. Acta 1964, 79, 511–520. [CrossRef]
26. Palma, A.; Li, L.; Chen, X.; Pappone, P.; McNamee, M. Effects of pH on acetylcholine receptor function.
J. Membr. Biol. 1991, 120, 67–73. [CrossRef]
27. Abdrakhamanova, G.; Cleemann, L.; Lindstrom, J.; Morad, M. Differential modulation of β2 and β4subunits
of human neuronal nicotinic acetylcholine receptors by acidification. Mol. Pharmacol. 2004, 66, 347–355.
[CrossRef]
28. Hutter, O.F.; Warner, A.E. The pH sensitivity of the chloride conductance of frog skeletal muscle. J. Physiol.
1967, 189, 403–425. [CrossRef]
29. Mallart, A.; Trautmann, A. Ionic properties of the neuromuscular junction of the frog: Effects of denervation
and pH. J. Physiol. 1973, 234, 553–567. [CrossRef]
30. Yamamoto, I.; Kamimura, H.; Yamamoto, R.; Sakai, S.; Goda, M. Studies on nicotinoids as insecticides. Pt. I.
Relation of structure to toxicity. Agric. Biol. Chem. 1962, 26, 709–716.
31. Zoltewicz, J.A.; Bloom, L.B.; Kem, W.R. Quantitative determination of the ring-chain hydrolysis equilibrium
constant for anabaseine and related tobacco alkaloids. J. Org. Chem. 1989, 54, 4462–4468. [CrossRef]
32. Bloom, L.B. Influence of Solvent on the Ring-Chain Hydrolysis Equilibrium of Anabaseine and Syntheses of
Anabaseine and Nicotine Analogs. Ph.D. Thesis, University of Florida, Gainesville, FL, USA, 1990; p. 207.
33. Zoltewicz, J.A.; Bloom, L.B.; Kem, W.R. Hydrolysis of cholinergic anabaseine and N-methylanabaseine:
Influence of cosolvents on the position of the ring-chain equilibrium-compensatory changes. Bioorg. Chem.
1990, 18, 395–412. [CrossRef]
34. Krumholz, P. Structural studies on polynuclear pyridine compounds. J. Am. Chem. Soc. 1951, 73, 3487–3492.
[CrossRef]
35. Green, B.T.; Goulart, C.; Welch, K.D.; Pfister, J.A.; McCollum, I.; Gardner, D.R. The non-competitive blockade
of GABAA receptors by an aqueous extract of water hemlock (Cicuta douglasii) tubers. Toxicon 2015, 108, 11–14.
[CrossRef]
36. Culver, P.; Fenical, W.; Taylor, P. Lophotoxin irreversibly inactivates the nicotinic acetylcholine receptor by
preferential association at one of the two primary agonist sites. J. Biol. Chem. 1984, 259, 3763–3770.
37. Tomizawa, M.; Casida, J.E. Molecular recognition of neonicotinoid insecticides: The determinants of life or
death. Acc. Chem. Res. 2009, 42, 260–269. [CrossRef]
38. Li, P.; Ann, J.; Akk, G. Activation and modulation of human alpha4beta2 nicotinic acetylcholine receptors by
the neonicotinoids clothianidin and imidacloprid. J. Neurosci. Res. 2009, 89, 1295–1301. [CrossRef]
39. Kaczanowska, K.; Camacho Hernandez, G.A.; Bendiks, L.; Kohs, L.; Cornejo-Bravo, J.M.; Harel, M.; Finn, M.G.;
Taylor, P. Substituted 2-aminopyrimidines selective for α7-nicotinic acetylcholine receptor activation and
association with acetylcholine binding proteins. J. Am. Chem. Soc. 2017, 139, 3676–3684. [CrossRef]
40. Kem, W.R. Biochemistry of Nemertine Toxins. In Marine Pharmacognosy: Marine Biotoxins as Probes of Cellular
Function; Martin, D.F., Padilla, G.M., Eds.; Academic Press: New York, NY, USA, 1973; pp. 37–84.

Mar. Drugs 2019, 17, 614
12 of 12
41. Kem, W.R.; Mahnir, V.M.; Prokai, L.; Papke, R.L.; Cao, X.; LeFrancois, S.; Wildeboer, K.;
Prokai-Tatrai, K.; Porter-Papke, J.; Soti, F. Hydroxy metabolites of the Alzheimer’s drug candidate
3-[(2,4-dimethoxy)benzylidene]-anabaseine dihydrochloride (GTS-21): Their molecular properties,
interactions with brain nicotinic receptors, and brain penetration. Mol. Pharmacol. 2004, 6, 56–67. [CrossRef]
42. Zoltewicz, J.A.; Prokai-Tatrai, K.; Bloom, L.B.; Kem, W.R. Long range transmission of polar effects of
cholinergic 3-arylideneanabaseines, Conformations calculated by molecular modelling. Heterocycles 1993
35, 171–179. [CrossRef]
,
43. Mallipeddi, P.L.; Pedersen, S.E.; Briggs, J.M. Interactions of acetylcholine binding site residues contributing
to nicotinc acetylcholine receptor gating: Role of residues Y93, Y190, K145 and D200. J. Mol. Gr. Model. 2013
44, 145–154. [CrossRef]
,
44. Hibbs, R.E.; Sulzenbacher, G.; Shi, J.; Talley, T.T.; Kem, W.R.; Marchot, P.; Taylor, P.; Bourne, Y. Structural
determinants for interaction of partial agonists with acetylcholine binding protein and neuronal α7 nicotinic
acetylcholine receptor. EMBO J. 2009, 28, 3040–3051. [CrossRef]
45. Brimblecombe, R.E.; Hunt, R.R.; Rickard, R.L.; Taylor, J.V. The synthesis and pharmacology of some
1,4,5,6-tetrahydropyrimidines. Br. J. Pharmacol. 1969, 37, 425–435. [CrossRef]
46. Upshall, D.G. Correlation of chick embryo teratogenicity with the nicotinic activity of a series of
tetrahydropyrimidines. Teratology 1972, 5, 287–294. [CrossRef] [PubMed]
47. Woodruff, G.N.; Walker, R.J.; Newton, L.C. Actions of some muscarinic and nicotinic agonists on the Retzius
cells of the leech. Comp. Gen. Pharmacol. 1971, 2, 106–117. [CrossRef]
48. Smith, C.R.; Richardson, C.H.; Shepard, H.H. Neonicotine and certain other derivatives of the dipyridyls as
insecticides. J. Econ. Entomol. 1930, 23, 863–870. [CrossRef]
49. Kem, W.R.; Scott, K.N.; Duncan, J.H. Hoplonemertine worms—A new source of pyridine neurotoxins.
Experientia 1976, 32, 684–686. [CrossRef] [PubMed]
50. Kem, W.R.; Soti, F.; Rittschof, D. Inhibition of barnacle larval settlement and crustacean toxicity of some
hoplonemertine pyridyl alkaloids. Biomol. Eng. 2003, 20, 355–361. [CrossRef]
51. Stricker, S.A.; Cloney, R.A. The ultrastructure of venom-producing cells in Paranemertes peregrina (Nemertea,
Hoplonemertea). J. Morphol. 1983, 177, 89–107. [CrossRef]
52. Schaefer, F.C.; Peters, G.A. Base-catalyzed reaction of nitriles with alcohols. A convenient route to imidates
and amidine salts. J. Org. Chem. 1960, 26, 412. [CrossRef]
53. Brown, D.J.; Evans, R.F. Hydropyrimidines.
Part II. A new general synthesis of substituted
1,4,5,6-tetrahydropyridines. J. Chem. Soc. 1962, 4039. [CrossRef]
54. Boksa, P.; Quirion, R. [3H]-methyl-carbamylcholine, a new radioligand specific for nicotinic acetylcholine
receptors in brain. Eur. J. Pharmacol. 1987, 139, 323–333. [CrossRef]
55. Kuryatov, A.; Luo, J.; Cooper, J.; Lindstrom, J. Nicotine acts as a pharmacological chaperone to upregulate
human alpha4 beta2 acetylcholine receptors. Mol. Pharmacol. 2005, 68, 1839–1851. [CrossRef]
56. Fitch, R.W.; Xiao, Y.; Kellar, K.J.; Daly, J.W. Membrane potential fluorescence: A rapid and highly sensitive
assay for nicotinic receptor channel function. Proc. Natl. Acad. Sci. USA 2003, 100, 4909–4914. [CrossRef]
[PubMed]
57. Wong, G.; Sei, Y.; Skolnick, P. Stable expression of type I gamma-aminobutyric acidA/benzodiazepine
receptors in a transfected cell line. Mol. Pharmacol. 1992, 42, 996–1003. [PubMed]
58. Green, B.T.; Lee, S.T.; Panter, K.E.; Welch, K.D.; Cook, D.; Pfister, J.A.; Kem, W.R. Actions of piperidine alkaloid
teratogens at fetal nicotinic acetylcholine receptors. Neurotoxicol. Teratol. 2010, 32, 383–389. [CrossRef]
[PubMed]
59. Lee, S.T.; Wildeboer, K.; Panter, K.E.; Kem, W.R.; Gardner, D.R.; Molyneux, R.J.; Chang, S.-W.T.; Soti, F.;
Pfister, J.A. Relative toxicities and neuromuscular nicotinic receptor agonistic potencies of anabaseine
enantiomers and anabaseine. Neurotoxicol. Teratol. 2006, 28, 220–228.
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).