Superagonist, Full Agonist, Partial Agonist, and Antagonist Actions of Arylguanidines at 5-Hydroxytryptamine-3 (5-HT3) Subunit A Receptors

Article abstract of DOI:10.1021/acschemneuro.6b00196

Introduction of minor variations to the substitution pattern of arylguanidine 5-hydroxytryptamine-3 (5-HT₃) receptor ligands resulted in a broad spectrum of functionally-active ligands from antagonist to superagonist. For example, (i) introduction of an additional Cl-substituent(s) to our lead full agonist N-(3-chlorophenyl)guanidine (mCPG, 2; efficacy % = 106) yielded superagonists 7-9 (efficacy % = 186, 139, and 129, respectively), (ii) a positional isomer of 2, p-Cl analog 11, displayed partial agonist actions (efficacy % = 12), and (iii) replacing the halogen atom at the meta or para position with an electron donating OCH₃ group or a stronger electron withdrawing (i.e., CF₃) group resulted in antagonists 13-16. We posit based on combined mutagenesis, crystallographic, and computational analyses that for the 5-HT₃ receptor, the arylguanidines that are better able to simultaneously engage the primary and complementary subunits, thus keeping them in close proximity, have greater agonist character while those that are deficient in this ability are antagonists.

Full text of DOI:10.1021/acschemneuro.6b00196

Subscriber access provided by UNIV OF OREGON
Article
Super-agonist, full agonist, partial agonist and antagonist
actions of arylguanidines at h5-HT3A receptors
Katie Alix, Shailesh Khatri, Philip D. Mosier, Samantha Casterlow, Dong Yan,
Heather L Nyce, Michael M. White, Marvin K. Schulte, and Malgorzata Dukat
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.6b00196 • Publication Date (Web): 17 Aug 2016
Downloaded from http://pubs.acs.org on August 20, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Chemical Neuroscience is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Super-agonist, full agonist, partial agonist and antagonist actions of
arylguanidines at h5-HT3A receptors
Katie Alix,† Shailesh Khatri,‡ Philip D. Mosier,† Samantha Casterlow,† Dong Yan,§ Heather L.
Nyce,§ Michael M. White,§ Marvin K. Schulte,‡ and Małgorzata Dukat*,†
†Department of Medicinal Chemistry, School of Pharmacy, Virginia Commonwealth University,
Richmond, VA 23298, USA
‡Department of Pharmaceutical Sciences, Philadelphia College of Pharmacy, University of Sciences,
Philadelphia, PA 19104, USA
§Department of Biochemistry and Molecular Biology Drexel University College of Medicine,
Philadelphia, PA 19102, USA
*Corresponding author. Tel.: 8048285256; Fax: 8048287625.
Eꢀmail address: mdukat@vcu.edu (M. Dukat)
Address: Department of Medicinal Chemistry, School of Pharmacy, Virginia Commonwealth University, PO Box
980540, Richmond, VA 23298ꢀ540, USA
ACS Chemical Neuroscience_ Revised _August 16, 2016
1
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 36
1
2
3
4
Abstract
5
6
7
8
9
Introduction of minor variations to the substitution pattern of arylguanidine 5ꢀHT3 ligands
resulted in a broad spectrum of functionallyꢀactive ligands from antagonist to superꢀagonist. For
example: i) introduction of an additional Clꢀsubstituent(s) to our lead full agonist Nꢀ(3ꢀ
chlorophenylguanidine (mCPG) (2; efficacy % = 106) yielded superꢀagonists 7–9 (efficacy % =
186, 139, and 129, respectively); ii) a positional isomer of 2, pꢀCl analog 11, displayed partial
agonist actions (efficacy % = 12); iii) replacing the halogen atom at the meta or para position
with an electron donating OCH3 group or a stronger electron withdrawing (i.e., CF3) group
resulted in antagonists 13–16. We posit based on combined mutagenesis, crystallographic, and
computational analyses that for the 5ꢀHT3 receptor, the arylguanidines that are better able to
simultaneously engage the primary and complementary subunits, thus keeping them in close
proximity, have greater agonist character while those that are deficient in this ability are
antagonists.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords:
Arylguanidines, 5ꢀHT3A receptors, functional activities, binding affinities, 3Dꢀgraphic models,
siteꢀdirected mutagenesis, SAR
2
ACS Paragon Plus Environment

Page 3 of 36
ACS Chemical Neuroscience
1
2
3
4
INTRODUCTION
5
6
7
8
The homomeric 5ꢀHT3A receptor is a member of the Cysꢀloop pentameric ligandꢀgated ion channel
(LGIC) superfamily. To date, five distinct 5ꢀHT3 receptor subunits, AꢀE, are known. Antagonists of 5ꢀ
HT3 receptors are currently used in the treatment of chemotherapyꢀ and radiationꢀassociated nausea and
vomiting. Partial agonists are being investigated for potential treatment of irritable bowel syndrome
(IBS), whereas a therapeutic role for 5ꢀHT3 receptor agonists has not yet been established.1
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nꢀ(3ꢀChlorophenyl)biguanide (mCPBG; 1) was one of the first selective 5ꢀHT3 receptor agonists
demonstrating utility for characterizing this receptor population. Nearly 20 years ago we showed that the
structure of 1 could be truncated to Nꢀ(3ꢀchlorophenyl)guanidine (mCPG; 2), and that mCPG (2) and a
naphthyl counterpart, Nꢀ(2ꢀnaphthyl)guanidine (3), possessed 5ꢀHT3 receptor agonist character.2 For
example, both 2 and 3 were active in a von Bezold–Jarisch assay and rabbit bladder assay, and their
actions were blocked using a 5ꢀHT3 receptor antagonist.2 mCPG (2) also served as a training drug in drug
discrimination studies using rats. The mCPG stimulus, generalized to compounds 1 and 3 as well as
several other 5ꢀHT3 receptor agonists, and was blocked by pretreatment of the animals with 5ꢀHT3
receptor antagonists.3
Additional arylguanidines were prepared in order to optimize affinity and formulate structure–affinity
relationships. Affinities of the analogs, depending upon the aryl substitution pattern and other structural
3
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 4 of 36
1
2
3
4
5
modifications, were found to span a >10,000ꢀfold range.2,4,5 These arylguanidines were also employed in
QSAR studies4,5 to better understand how they might interact at 5ꢀHT3 receptors.
6
7
8
9
Although a distinct possibility, there is no reason to suspect that structurally–related agonists and
antagonists bind differently at a receptor. Given the premise that structurally–related agonists might,
perhaps, have a greater likelihood of binding in a manner more similar to one another than to that of
antagonists, and vice versa, it was important to determine the functional activity of 2ꢀlike agents. That is,
subsequent structure–activity studies could focus on agonist analogs and antagonist analogs individually;
this could provide useful information with regard to what structural feature(s) contribute to which action.
Accordingly, we examined the functional activity of sixteen (2–17; Tables 1 and 2) arylguanidine analogs
at 5ꢀHT3A receptors expressed in Xenopus oocytes using a two–electrode voltage clamp assay.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Here we show that by making minor variations to the substitution pattern of a small, unconstrained
arylguanidine core scaffold, ligands with a broad spectrum of functional activity from antagonist to superꢀ
agonist are obtained. We postulate, based on combined mutagenesis, crystallographic, and computational
analyses, that for the 5ꢀHT3 receptor, the arylguanidines that are better able to simultaneously engage the
primary and complementary subunits, thus keeping them in close proximity, have greater agonist
character while those that are deficient in this ability are antagonists.
RESULTS and DISCUSSION
Ligand Synthesis. mCPBG (1) and arylguanidines 2, 3, 7–9 and 11–17 (Tables 1 and 2) were
resynthesized as previously reported by us.2,4 The nitrate salt of arylguanidine 4 (Table 1) was prepared
according to a literature procedure.6 Arylguanidine 5 (Table 1; Scheme 1) was prepared using our
standard protocol for the synthesis of arylguanidines as depicted in Scheme 1. The free base of 6 (Table
1) was synthesized following the literature procedure7 and then converted to its nitrate salt. In short, the
corresponding commercially available anilines (for 4–6) were converted to their HCl salts and heated at
reflux in a condensation reaction with cyanamide in ethanol. The hydrochloride salts of the obtained
4
ACS Paragon Plus Environment

Page 5 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
arylguanidines were converted to the nitrate salts (4–6) with ammonium nitrate. Nꢀ(3ꢀChloroꢀ4ꢀ
hydroxyphenyl)guanidine hydrochloride (10) (Table 1) was synthesized by the literature method outlined
in Scheme 1. The nitro intermediate, 19, was obtained by sonication of 2ꢀchlorophenol (18) with nitric
acid (9% aqueous).8 The nitro group was then reduced using sodium hydrosulfite and aqueous sodium
hydroxide.9 An ethanolic solution of the HCl salt of aniline 20 was heated at reflux with cyanamide (50%
aqueous) to give 10.10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1 and 2
Effects of arylguanidines on 5-HT3 receptor currents. Both the endogenous ligand 5ꢀHT, and
mCPBG (1), produced concentrationꢀdependent inward currents in Xenopus oocytes expressing h5ꢀHT3A
receptors (EC50 = 2.34 and 0.96 ꢀM, respectively) (Table 1). The examined arylguanidines can be divided
into two major groups: agonists and antagonists. Compounds 2–11, similar to 5ꢀHT and mCPBG (1),
produced concentrationꢀdependent inward currents (Table 1; Figure S1). That is, all metaꢀsubstituted
heavy halogen (Cl, Br, I; 2, 5 and 6, respectively) arylguanidine derivatives mimicked the action of 5ꢀHT
and mCPBG (1) with comparable efficacy (Table 1), whereas introduction of additional chloroꢀgroups as
seen in arylguanidines 7–9 resulted in substantially moreꢀefficacious agonists, i.e. “superꢀagonists” (Table
1). A “superꢀagonist” is a ligand that elicits a maximal effect greater than that of the endogenous ligand.
This phenomena was reported for one of the oldest and most studied LGICRs, nAChRs11,12 as well as for
GPCRs.13 Table 3 shows the relative efficacy for each of the compounds 7−10 compared to the
endogenous agonist 5ꢀHT. Using unpaired t tests, 7−9 were all shown to produce statistically significant
increases in efficacy compared to 5ꢀHT (p<0.02 for all). In contrast, introduction of a paraꢀhydroxyl
group to the parent mCPG (i.e., 10) resulted in a weaker agonist with no significant difference in efficacy
from the control, 5ꢀHT (Tables 1 and 3).
Table 3
5
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 6 of 36
1
2
3
4
5
6
7
8
9
The arylguanidine 4 bearing a light halogen (fluoro group) in the metaꢀposition yielded not only a less
potent, but also a less efficacious analog compared to the parent mCPG (2); compound 4 behaved as a
partial agonist. Similarly, a positional isomer of mCPG (2), analog 11, as well as a bioisostere of 7,
naphthylguanidine 3, produced partial agonist actions (Table 1). Representative concentration–response
curves for superꢀagonists 7–9 and agonist 10 in comparison to the endogenous ligand 5ꢀHT are depicted
in Figure 1. Interestingly guanidines 12–17 failed to produce current when examined alone (data not
shown), however they produced concentration–dependent inhibition of 2.5 ꢀM 5ꢀHTꢀelicited current, thus
behaving as antagonists (Table 2; Figure S2). Representative traces for currents induced by full agonist 2,
superꢀagonist 9, partial agonist 11, and antagonist 15 in comparison to responses obtained for a control
concentration of 5ꢀHT are shown in Figure 2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1 and 2
Site-directed mutagenesis. The orthosteric binding site (OBS) of 5ꢀHT3 receptors (and other Cysꢀloop
receptors), the presumed ligand binding site for the compounds studied here (because they bind at 5ꢀHT3
receptors in a competitive manner), contains a large proportion of aromatic amino acid residues, several
of which have been implicated as playing a role in ligand–receptor interactions.14ꢀ24 The effect of mutating
selected amino acid residues in the OBS on ligand affinity are discussed in the modeling section below.
Molecular Modeling. Many small–molecule and peptide ligands are able to bind to the 5ꢀHT3A OBS,
located between two adjacent subunits in the extracellular domain (ECD), and either trigger or prevent the
firing of the channel.25 A widelyꢀaccepted hypothesis regarding the activation of LGICs is that the degree
of closure of the C loop, located on the principal subunit, is directly correlated with the degree of agonism
associated with a particular ligand.26 However, the efficacy of a particular orthosteric ligand appears not
only to be correlated with variation in the closure of loop C as has been previously suggested, but also
with its related ability to form a strong interꢀsubunit bridge between the primary and complementary
subunits of the OBS.27 As the OBS in the 5ꢀHT3 receptor (and other Cysꢀloop LGICs) features an
6
ACS Paragon Plus Environment

Page 7 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
“aromatic box”, aromatic and/or piꢀcation interactions are of central importance. The 5ꢀHT3 receptor also
features a glutamate residue (E236) on the C loop that contributes to the OBS and is able to engage in
strong interactions with cationic ligands. In addition, halogen bonds have also been shown to play a major
role in the formation of the bridge, for example, 1ꢀ(pyridinꢀ3ꢀyl)ꢀ1,4ꢀdiazepine agonists acting at the
nicotinic α4β2 receptor.28 Closure of the binding site upon activation has also been observed in a recent
crystal structure of a glutamateꢀgated chloride channel, wherein the βꢀstrands of the complementary
subunit move closer to the primary subunit in activated states of the receptor.29 A recentlyꢀreported
molecular dynamics (MD) simulation of the activation of the 5ꢀHT3A receptor proposed that changes in
the OBS include an outward motion of the C loop by ~3 Å when bound with 5ꢀHT.30
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To gain insight into the mechanism by which the arylguanidines modulate the activity of the 5ꢀHT3
receptor, we generated a population of mouse 5ꢀHT3 (m5ꢀHT3) models with varying degrees of Cꢀloop
openness in the OBS of the extracellular ligand binding domain (LBD) to model the inherent flexibility of
this loop. The models were based on the recentlyꢀdescribed Xꢀray crystal structure of the m5ꢀHT3
receptor31 (PDB ID: 4PIR; see Experimental Methods and Supporting Information). Interꢀ and intraꢀ
subunit hydrogen bond and ionic interactions between amino acid residues on the periphery of the OBS
are involved in the stabilization of the OBS. Such interactions include the side chain hydroxyl group of
Y143 on the D loop with T186 and D189, for D229 and R92, for E129 and T179, and for E236 and K238
(green dashed lines in Figure 3A).
Figure 3
Docking studies using these models were subsequently conducted with the arylguanidine ligands
reported here to identify potential binding modes of the agonists and antagonists. Using clustering
methods applied to the docking results, we identified binding modes for all arylguanidines corresponding
to a common degree of C loop openness (15 Å; Figure 3A), reflecting the similar size of the ligands. Our
analysis of the automated docking results for the arylguanidines indicated that both agonists and
7
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 8 of 36
1
2
3
4
5
6
7
8
9
antagonists adopt a common binding mode within the “aromatic box”32 (defined by W183, F226 and
Y234 on the principal subunit and W90 and Y153 on the complementary subunit) in which the phenyl
ring and its substituent are sandwiched between W183 on the principal subunit and W90 on the D loop of
the complementary subunit, and the guanidinium group is ionꢀpaired with the side chain carboxylate
group of E236 and hydrogenꢀbonded with the side chain of T181 on the C loop of the primary subunit
(Figure 3A). This pose is very similar to that of 5ꢀHT after the initial 30ꢀns equilibration in the recently
reported MD simulation of the gating mechanism of 5ꢀHT3.30 However, subtle but important differences
exist between the docked poses of the agonists and antagonists. Most notably, the phenyl ring of the
agonists is able to effectively interact with the aromatic side chain of Y234 on the C loop, maintaining
van der Waals contact in an edgeꢀface manner, strengthening the bridge between principal and
complementary subunits. This proposed agonist property is also consistent with the aforementioned MD
receptor gating study wherein the interaction between the corresponding aryl portion of 5ꢀHT and Y207
(cognate residue of Y234) was purported to be important in the activated form of the receptor.30 In
contrast, the phenyl ring of the antagonists is seated lower in the crevice formed by W183 and W90 and is
not able to maintain the close van der Waals distance needed for effective engagement of Y234, and a
correspondingly weaker interaction with the C loop.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The interaction modes of the ligands correspond to their associated functional properties (Figures 3
and S8–9). A classical principle of medicinal chemistry and drug design states that compounds that are
able to effectively fill a pocket in a target macromolecule will increase its affinity for the target. The
increase in affinity associated with simply filling such a pocket for even a small properly positioned
methyl groupꢀsized substituent can be two orders of magnitude.33 For the superꢀagonists 7–9, the presence
of multiple chloro groups allows them to fill the OBS in a stereoelectronically favorable manner,
enhancing their interaction via hydrophobic interactions with W183 and W90 as well as potential
hydrogen bonds with R92 (Figure 3B). The agonist 2 lacks the additional chloro groups of 7–9, but its mꢀ
Cl group is optimally sandwiched between W183 and W90 (Figure 3C), consistent with reduced affinity,
8
ACS Paragon Plus Environment

Page 9 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
potency and efficacy. Other analogs with halogens that are larger and more polarizable than chlorine in
the metaꢀposition show progressive increases in affinity and potency. The mꢀF analog 4 shows
dramatically reduced affinity, potency and efficacy likely due to the unfavorable placement of the fluoro
group either between the tryptophan rings or proximal to the R92 guanidinium group, as fluorine is
neither polarizable nor has appreciable hydrogenꢀbonding capability. Analog 11 with a pꢀCl group, is a
weak partial agonist and, due to the location of the paraꢀsubstituent, may allow the chloro group to Hꢀ
bond with R92, preventing its phenyl ring from forming optimal interactions with W183/W90 and R92
(Figure 3D; cf. 12 below).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the antagonists, the phenyl ring is located deeper in the groove formed by W90 and W183, either
due to lack of an aromatic substituent (12) or one that is too large. Arylguanidine 15 for example, cannot
place its mꢀCF3 group between W183 and W90 due to steric bulk (and electrostatic repulsion), and must
direct it toward R92 (Figure 3E) where it also cannot engage in productive interactions due to lack of
polar or aromatic complementarity, providing an explanation for its low affinity. The methoxy group of
13 cannot form a strong hydrogen bond with R92 due to the bulk of its methyl group, which is directed
toward R92. Analog 16 was found to have higher affinity and potency than other antagonists, likely due
in part to the small favorable interaction of its pꢀCF3 group with the D loop backbone corresponding to
residues W90–R92. Aryl CF3 groups have been recognized as being able to interact favorably with the piꢀ
system of amides with a binding free energy of about 1 kcal/mol.34ꢀ36 The presence of a mꢀCF3 and a pꢀCl
group (17) is predicted to make the ligand too large to fit completely into the aromatic box, and an
alternative binding mode was predicted for it in which the guanidinium group is piꢀstacked with W183,
the phenyl ring is sandwiched between W90 and R92, and the trifluoromethyl group is directed toward
the D loop backbone (Figure S9), and may reflect the compound’s enhanced affinity and potency. In
general, this binding mode also cannot be ruled out as a contributing factor for the other compounds in the
series, although it is less able to explain the SAR.
9
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 10 of 36
1
2
3
4
5
6
7
8
9
Interestingly, strong correlations were observed for the arylguanidines between their pKi values and
either their pEC50 values (compounds with agonist activity: r = 0.91) or their pIC50 values (antagonists: r
= 0.92). In addition, for the compounds with agonist activity, a modest correlation was observed between
the pKi values and their percent efficacy (r = 0.73) (Figure S10). Moreover, the putative binding modes
are also consistent with mutagenesis data (Tables 4–6). The W90F mutation reduces the affinity of all
arylguanidines by at least 10ꢀfold as a result of the removal of a significant aromatic portion of the
aromatic box; this mutation may also contract the size of the box, making it more difficult for the ligands
to bind. The R92A mutation shows more modest reductions in affinity for the agonists, a likely result of
the loss of Hꢀbonding character provided by the arginine guanidinium group. The N128A mutant
essentially retains affinity for most ligands (±2ꢀfold). N128 interacts directly with the agonist
arylguanidines by donating one of its sideꢀchain primary amide group hydrogens to the piꢀsystem of the
ligand guanidinium group. Mutation of the C loop residue F226 to tyrosine uniformly reduced the affinity
of all compounds by at least ~6ꢀfold. Although the models indicate that the additional tyrosine hydroxyl
group does not directly interact with the bound ligands, its presence may stabilize a different, less
complementary conformation of the OBS with respect to the arylguanidines. E236 is proposed to form a
salt bridge with the ligand guanidinium group. Removal of its anionic charge (E236N) uniformly and
significantly reduced the affinity of all tested ligands. Removing the bulk of the E236 side chain (E236A)
also substantially reduced affinity for all ligands except the partial agonist 11 (paraꢀchloro group) and the
antagonist 13 (metaꢀmethoxy group). Although our initial models do not provide an obvious explanation
for this particular observation, additional modeling/simulation (e.g. molecular dynamics) may yield
further insight. Finally, the effect of the hydroxyl groups of tyrosines 143 and 153 was examined. Both
the Y143F and Y153F mutations resulted in modest decreases for mCPBG (1) and 2, with a significantly
larger effect found for Y143F. As mentioned above, the hydroxyl group of Y143 interacts directly with
two other principal subunit residues, and loss of these interactions most probably destabilize the OBS.
The effect of these mutations on the binding of 5ꢀHT was much larger; however 5ꢀHT has been proposed
to bind in a different manner than the arylguanidines.37
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10
ACS Paragon Plus Environment

Page 11 of 36
ACS Chemical Neuroscience
1
2
3
4
Table 4ꢀ6
5
6
7
8
9
CONCLUSIONS
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Previous studies have attempted to explain the SAR of arylguanidine binding at 5ꢀHT3 receptors in
the absence of available information on their functional activity. Here, we conducted for the first time,
studies with 16 arylguanidines and compared the results with those of the standard agonists 5ꢀHT and
mCPBG (1). What was found was that their functional activity spanned a broad spectrum from superꢀ
agonist, to full agonist, to partial agonist, to antagonist at 5ꢀHT3 receptors. As such, they represent a
congeneric series of agents that not only displayed different functional activities, but did so with minimal
structural alterations. This was a nearly ideal situation whereby siteꢀdirected mutagenesis combined with
molecular modeling/docking studies could be conducted relative to the different functional actions, and
with a minimum of structural perturbation (on the part of the ligands). This facilitated the elucidation of a
proposed common binding mode that incorporated ligand functionality and led to additional insights into
the ligandꢀbased modulation of the receptor. The results presented here confirm the utility of
phenylguanidine as an extremely versatile scaffold in the design of 5ꢀHT3 ligands with a ‘tunable’ level of
agonist or antagonist activity. Thus, phenylguanidine derivatives will continue to serve as useful tools in
the study of 5ꢀHT3 receptor structure and function. In addition, they may also lead to novel therapeutic
agents, most notably for cases in which 5ꢀHT3 antagonism or partial agonism has been shown to have
relevance.
METHODS
Chemistry. Melting points were taken in glass capillary tubes on a Thomas–Hoover melting point
apparatus and are uncorrected. 1H NMR spectra were recorded with a Varian EMꢀ400 spectrometer with
tetramethylsilane (TMS) as an internal standard. Peak positions are given in parts per million (δ). Infrared
11
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 12 of 36
1
2
3
4
5
6
7
8
9
spectra were obtained on a Nicolet Avatar 360 FTꢀIR spectrophotometer. Microanalyses were performed
by Atlantic Microlab Inc. (Norcross, GA) for the indicated elements and results are within 0.4% of
calculated values. Chromatographic separations were performed on silica gel columns (Silica Gel 62, 60–
200 mesh, SigmaꢀAldrich). Flash chromatography was performed on a CombiFlash Companion/TS
(Teledyne Isco Inc., Lincoln, NE) apparatus. Reactions were monitored by thinꢀlayer chromatography
(TLC) on silica gel GHLF plates (250 µm, 2.5 × 10 cm; Analtech Inc., Newark, DE).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3ꢀFluorophenylguanidine Nitrate (4). A mixture of 3ꢀfluoroaniline hydrochloride (1.00 g, 6.78 mmol) and
cyanamide (0.38 g, 9.04 mmol) was heated at reflux in abs EtOH (5 mL) for 6 h. The solvent was
removed under reduced pressure to give a paleꢀyellow oil. The oil was dissolved in H2O (2 mL) and
NH4NO3 (1.00 g, 12.5 mmol) was added in excess. The solvent was removed under reduced pressure and
the residue was recrystallized four times from H2O. The lightꢀbrown crystals were collected by filtration
and washed with cold Et2O (3 × 5 mL) to give 4 (0.81 g, 55%): mp 145–146 °C (H2O). The product was
subsequently recrystallized from EtOH, mp 146–147 °C (Lit6 165 °C EtOH); IR (KBr, cmꢀ1): 3342
(amine), 3330 (amine), 3195 (cyclic alkene), 1669 (cyclic alkene), 1597 (alkene), 1493 (amine salt), 1369
(nitrate), 1143 (fluorine); 1H NMR (DMSOꢀd6) δ 7.21–7.04 (m, 3H, ArH), 7.53–7.45 (m, 5H, ArH, NH2,
ex with D2O), 9.72 (s, 1H, NH, ex with D2O). 13C NMR (100 MHz, DMSOꢀd6): δ 162.9 (quat., 1JCF =
244.1 Hz), 156.1 (quat.), 137.7 (quat., d, 3JCF = 10.4 Hz), 131.8 (CH, d, 3JCF = 9.5 Hz), 120.7 (CH, d, 4JCF
= 2.9 Hz), 113.6 (CH, d, 2JCF = 20.9 Hz), 112.0 (CH, d, 2JCF = 24.2 Hz). Anal. calcd for (C7H8FN3·HNO3)
C, H, N. A sample of 4 was free based and recrystallized from benzene, mp 93–94 °C (Lit6 96 °C
benzene).
3ꢀBromophenylguanidine Nitrate (5). 3ꢀBromoaniline hydrochloride (4.0 g, 19.2 mmol) and cyanamide
(2.1 g, 50.0 mmol) were added to abs EtOH (20 mL). The reaction mixture was heated at reflux for 24 h
and cooled to room temperature. Distilled H2O (5.0 ml) was added to the solution followed by addition of
saturated solution of NH4NO3 (10.0 mL). A white precipitate promptly formed and was recrystallized
from distilled H2O. The product was further purified by flash chromatography (CH2Cl2:MeOH ꢀ 9:1) to
12
ACS Paragon Plus Environment

Page 13 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
afford white crystals (0.72 g, 14%): mp 160–162 ºC; ¹H NMR (300 MHz, DMSOꢀd6, δ): 7.26 (ddd, J =
7.80 Hz, 1H, ArH), 7.41 (t, J = 7.80 Hz, 1H, ArH), 7.47–7.52 (m, 6H, ArH and NH2, D2O exchangeable).
13C NMR (100 MHz, DMSOꢀd6): δ 156.1 (quat.), 137.7 (quat.), 132.0 (CH), 129.6 (quat.), 127.5 (CH),
123.9 (CH), 122.5 (CH). Anal. calcd for (C7H8BrN3·HNO3) C, H, N.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3ꢀIodophenylguanidine Nitrate (6). 3ꢀIodoaniline hydrochloride (4.4 g, 17.22 mmol) and NH2CN (2.1 g,
50.0 mmol) were added to abs EtOH (20 mL). The reaction mixture was heated at reflux for 24 h and
cooled to room temperature. Distilled H2O (5.0 ml) was added to the solution followed by addition of
saturated solution of NH4NO3 (10.0 mL). A white precipitate promptly formed and was recrystallized
from distilled H2O. The product was further purified by flash chromatography (CH2Cl2:MeOH ꢀ 9:1) to
afford white crystals (0.40 g, 7%): mp 166–168 ºC; ¹H NMR (300 MHz, DMSOꢀd6, δ): 7.25 (t, J = 8.40
Hz, 1H, ArH), 7.27–7.30 (m, 1H, ArH), 7.49 (br s, 4H, NH2, D2O exchangeable), 7.62–7.63 (m, 1H,
ArH), 7.65–7.69 (m, 1H, ArH). 13C NMR (100 MHz, DMSOꢀd6): δ 156.1 (quat.), 137.5 (quat.), 135.5
(CH), 133.2 (CH), 132.0 (CH), 124.4 (CH), 95.7 (quat.). Anal. calcd for (C7H8IN3·HNO3) C, H, N.
Molecular modeling. The 5ꢀHT3 ligands in this study were sketched using SYBYLꢀX 2.1 (Certara, St.
Louis, MO) and energyꢀminimized using the Tripos Force Field (TFF) with Gasteiger–Hückel charges
and a distanceꢀdependent dielectric constant of 4.0 D/Å to an energy gradient cutꢀoff of 0.05
kcal(mol×Å)ꢀ1.
Models representing the ligand binding domains (LBDs) of two adjacent subunits of the
homopentameric mouse 5ꢀHT3A receptor were generated using the A (primary) and E (complementary)
subunits from the m5ꢀHT3A receptor crystal structure (PDB ID: 4PIR). The orthosteric ligand binding
site, consisting of loops A–G, was modeled at the interface of subunits A and E. Residues preceding T35
and following P247 were not included in the models. A comparison of the 5ꢀHT3A LBDs between human
13
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 14 of 36
1
2
3
4
5
6
and mouse reveals that no significant differences in amino acid composition exist in the OBS, where the
present compounds are predicted to bind (Figures S3–S4).
7
8
9
It has been noted that loop C in antagonistꢀbound crystal structures adopts a more ‘open’
conformation than is observed for agonistꢀbound structures.38 Thus, a population of 17 models of the LBD
were generated with varying degrees of C loop openness following a procedure described by Sander et
al.39 (Figure S5). In this procedure, centroids representing the tip of the primary subunit C loop (defining
atoms: S227.CA, I228.CA, D229.CA, I230.CA) and the complementary subunit “back wall” (defining
atoms: Y91.CA, Y91.C, Y91.O, K154.CA) were created. Initial crystal structure openness = 17.4 Å; after
minimization = 15.6 Å.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
GOLD Suite v.5.2 (Cambridge Crystallographic Data Centre)40 was used to dock the arylguanidines
into each member of the population to generate candidate binding modes. For the arylguanidine
derivatives, four rotamers were defined about the guanidineꢀaryl bond representing four discrete local
energy minima (+synclinal, 60° to 80°; +anticlinal, 100° to 120°; –anticlinal, –120° to –100°; –synclinal,
–80° to –60°). These rotamers were identified using two different energy minimization routines and
observed crystal structures (Figure S6) and subsequently incorporated into the gold.tordist rotamer library
(see the online GOLD Suite documentation for implementation details).
To facilitate the identification of common and disparate binding modes for the ligands, an inꢀhouse
clustering algorithm (Figure S7) was developed. Briefly, each ligand was flexibly docked ten times into
each of the spatiallyꢀaligned 17 5ꢀHT3 models described above. Then, the topꢀscoring solution for each
ligand at each receptor was selected. These solutions were subsequently clustered at a cutoff RMSD value
of 2.0 Å to identify different binding modes for each ligand. The highestꢀscoring solution from each
cluster (binding mode) was selected to represent a particular binding mode; this solution was referred to
as the exemplar. All exemplars for all ligands were then combined and clustered based on the RMSD of
their common phenylguanidine core to identify common binding modes among the set of compounds.
14
ACS Paragon Plus Environment

Page 15 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
The common binding modes were then manually evaluated. A common binding mode for both agonists
and antagonists was identified with a loop opening of 15 Å and selected for further study based on its
consistency with experimentallyꢀderived data, namely mutagenesis and ligand SARs. The resulting
receptor–ligand complexes were then energyꢀminimized in SYBYLꢀX 2.1 and energyꢀminimized under
the same conditions as described above for the ligands.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Correlation plots (Figure S10) were generated using SigmaPlot 13 (Systat Software, Inc.).
Functional characterization in Xenopus oocytes. Compounds 1–17 were evaluated functionally using
twoꢀelectrode voltage clamp of murine 5ꢀHT3A receptors expressed in Xenopus oocytes using previously
published procedures.41 Four independent recording stations with integral autosampling and data
collection were available for this study.
Brief methodology: Oocytes were surgically excised, then defolliculated by shaking for 90 min in 1.5%
collagenase dissolved in Ca2+ꢀfree ORꢀII medium (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl2 and 5 mM
HEPES, pH 7.4). cRNA was synthesized in vitro using an appropriate cDNA and the mMESSAGE
MACHINE® mRNA synthesis kit (Ambion). Each oocyte was injected with 50 nL of mRNA (0.3 ꢀg/ꢀL
concentration) and incubated at 19 °C for at least 24 h prior to electrophysiological recording. An oocyte
was placed in the oocyte recording chamber and perfused with NDꢀ96 buffer (96 mM NaCl, 2 mM KCl,
1.8 mM CaCl2, 1 mM MgCl2, 2 mM Phosphate, pH 7.4) at a flow rate of 20 mL/min. Two electrodes (1–2
Mꢁ filled with 3 M KCl) corresponding to the current and voltage electrodes were inserted and the
holding potential clamped at –60 mV using a Warner instruments amplifier. Ligands were prepared in
NDꢀ96 buffer then injected into the chamber. Current responses were recorded using a DATAPAC A/D
board and software. Data were analyzed by nonꢀlinear curve fitting algorithms using GraphPad Prism
software. For determination of antagonist Ki values, the IC50 was first determined by inhibition of 5ꢀHT
15
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 16 of 36
1
2
3
4
5
6
7
8
induced responses by increasing concentrations of the antagonist (compound 4). The Ki was calculated
from the IC50 using the Cheng–Prusoff relationship. A concentration of 5ꢀHT equal to the EC50 and EC90
for 5ꢀHT stimulation of the receptor was used for inhibition and agonist like experiments respectively.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Molecular biology and transfection. A full length cDNA clone corresponding to the mouse 5ꢀHT3A(b)
form42 of the receptor was isolated from a neuroblastoma N1Eꢀ115 cell line cDNA library as previously
described.43 The expression patterns of the 5ꢀHT3A and 5ꢀHT3B subunits has led to the proposal that 5ꢀ
HT3 receptors in the central nervous system are 5ꢀHT3A homomers, while those in the peripheral nervous
system are a mixture of 5ꢀHT3A homomers and 5ꢀHT3A/5ꢀHT3B heteromers.44,45 Despite the differences
in subunit composition, both types of receptors have essentially identical ligandꢀbinding properties.46
Thus, homomeric 5ꢀHT3A receptors are an appropriate model for the structure of the ligandꢀbinding
domain of native 5ꢀHT3 receptors, regardless of whether they are 5ꢀHT3A homomers or 5ꢀHT3A/5ꢀHT3B
heteromers.
Siteꢀdirected mutagenesis was carried out using the QuikChange system (Stratagene) or pAlter altered
states mutagenesis kit (Promega) and the entire coding region of the mutant subunit was sequenced to
ensure that only the desired mutation is present. The nomenclature used to describe mutants is amino acid
in wildꢀtype/position/substitution; e.g., W90F. Since the amino terminus of the mature 5ꢀHT3A subunit is
unknown, the amino acid numbering system used here includes the signal sequence and starts from the
initial methionine. Cultures of tsA201 cells, a derivative of the widelyꢀused HEK 293 cell line, were
maintained in DMEM medium containing 10% fetal bovine serum, 100 units/mL penicillin and 100
units/mL streptomycin. Cultures at 50–60% confluence were transfected with 10 ꢀg receptor cDNA per
100 mm dish using the Fugene transfection reagent (Roche Diagnostics). Maximal expression was
obtained 36–72 h after transfection.
Ligand binding assays. Transfected cells were scraped from dishes, washed three times with phosphateꢀ
buffered saline and resuspended and homogenized in 6 mL of 154 mM NaCl, 50 mM TrisꢀHCl, pH 7.4
16
ACS Paragon Plus Environment

Page 17 of 36
ACS Chemical Neuroscience
1
2
3
4
per 100 mm dish. The homogenate was then used in binding assays or frozen until needed. No change
5
6
was observed in either ligand affinity or B
values after freezing.
max
7
Membranes were incubated for 2 h at 37 °C in a total volume of 0.5 mL of 154 mM NaCl, 20 mM Trisꢀ
HCl, pH 7.4 containing the appropriate concentrations of the competing unlabeled ligand (e.g., mCPG; 2)
and radioligand ([3H] granisetron; PerkinElmer, 85 Ci/mmol). Binding was terminated by rapid vacuum
filtration onto GF/B filters pretreated with 50 mM TrisꢀHCl, pH 7.4, 0.2% polyethyleneimine and the
filters were washed with 10 mL of cold 50 mM TrisꢀHCl, pH 7.4 per sample. Nonspecific binding was
defined as that binding not displaced by 10 ꢀM mCPBG (1). IC50 values for the various compounds were
determined by fitting the data to Eq. 1 using a Levenberg–Marquardt algorithm in a commercially
available software package (Igor Pro, WaveMetrics, Oswego, OR):
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(Eq. 1)
[ I ]
= (1 + (
θ
) n )−1
IC 50
where
concentration [I] compared to that in the absence of antagonist, IC50 is the concentration of antagonist at
which =0.5, and n is the apparent Hill coefficient. All experiments were carried out with a
θ
is the fractional amount of [3H]granisetron bound in the presence of the antagonist at
θ
[3H]granisetron concentration equal to its experimentallyꢀdetermined dissociation constant for the
particular receptor (WT: 2.2 nM; W90F: 11 nM).
Supporting Information
C, H, N analysis for compounds 4–6 (Table S1). Doseꢀresponse curves for agonists 1–11 (Figure S1).
Inhibition of responses elicited by 2.5 µM serotonin for antagonists 12–17 (Figure S2). Sequence
alignment of the mouse and human 5ꢀHT3A LBDs (Figure S3). Amino acid differences between the m5ꢀ
HT3R and h5ꢀHT3AR (Figure S4). Modeled LBD C loop flexibility (Figure S5). Aryl–guanidinium torsion
angle preferences (Figure S6). Ligand binding mode identification workflow (Figure S7). Putative
17
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 18 of 36
1
2
3
4
5
6
7
8
binding modes for agonists not shown in main text (Figure S8). Putative binding modes for antagonists
not shown in main text (Figure S9). Relationships between affinity (pKi), potency (pEC50/pIC50) and
percent efficacy (Figure S10).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Supplementary data related to this article can be found at http://...
Abbreviations
mCPG, Nꢀ(3ꢀchlorophenylguanidine; mCPBG, LGICR, ligandꢀgated ion channel receptor; Nꢀ(3ꢀ
chlorophenyl)biguanide; QSAR, quantitative structureꢀactivity relationship; 5ꢀHT, serotonin; OBS,
orthosteric binding site; ECD, extracellular domain; MD, molecular dynamics; LBD, ligand binding
domain; SAR, structureꢀactivity relationship; nAChR, nicotine acetylcholine receptor;
AUTHOR INFORMATION
Corresponding Author
*Eꢀmail: mdukat@vcu.edu
AUTHOR CONTRIBUTIONS
M. D. conceived the idea, supervised the work and prepared the first draft of the manuscript; K. A. and S.
C. synthesized arylguanidines under the supervision of M. D.; S. K. performed electrophysiological
studies at m5ꢀHT3A receptors under the supervision of M. K. S.; D. Y. and H. L. N. performed siteꢀ
directed mutagenesis and radioligand binding studies at m5ꢀHT3A receptors under the supervision of M.
M. W.; P. D. M. generated homology models and carried out docking studies. All coꢀauthors contributed
to the preparation of the manuscript.
18
ACS Paragon Plus Environment

Page 19 of 36
ACS Chemical Neuroscience
1
2
3
4
FUNDING
5
6
7
8
9
This work was supported, in part, by Jeffress Memorial Trust RG Jꢀ778 (MD), and A. D. Williams Trust
funds (MD).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACKNOWLEDGMENT
5ꢀHT3 receptor Ki values for compounds 2, 4−6 and 10 were generously provided by the National Institute
of Mental Health's Psychoactive Drug Screening Program, # HHSNꢀ271ꢀ2013ꢀ00017ꢀC (NIMH PDSP).
The NIMH PDSP is directed by Bryan L. Roth MD, PhD at the University of North Carolina at Chapel
Hill and Project Officer Jamie Driscoll at NIMH, Bethesda MD, USA. The authors are grateful to Prof.
Richard A. Glennon for discussions and proof reading of the manuscript. Samantha Casterlow was a
MARC STAR scholar at VUU and received a portion of her synthetic training from Dr. Eliseu
DeOliveira. We thank Dr. Umberto M. Battisti for his assistance with 13C NMR.
CONFICT of INTEREST
The authors report no conflict of interest.
19
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 20 of 36
1
2
3
4
References
5
6
(1) Lummis, S. C. R. (2012) 5ꢀHT3 receptors. J. Biol. Chem. 287, 40239ꢀ40245.
7
8
9
(2) Dukat, M., AbdelꢀRahman, A. A., Ismaiel, A. M., Ingher, S., Teitler, M., Gyermek, L., and Glennon,
R. A. (1996) Structure–activity relationships for the binding of arylpiperazines and arylbiguanides at 5ꢀ
HT3 serotonin receptors. J. Med. Chem. 39, 4017ꢀ4026.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(3) Dukat, M., Young, R., Darmani, N. N., Ahmed, B., and Glennon, R. A. (2000) The 5ꢀHT3 agent Nꢀ(3ꢀ
chlorophenyl)guanidine (MDꢀ354) serves as a discriminative stimulus in rats and displays partial agonist
character in a shrew emesis assay. Psychopharmacology 150, 200ꢀ207.
(4) Dukat, M., Choi, Y.ꢀN., Teitler, M., Du Pre, A., HerrickꢀDavis, K., Smith, C., and Glennon, R. A.
(2001) The binding of arylguanidines at 5ꢀHT3 serotonin receptors: a structure–affinity investigation.
Bioorg. Med. Chem. Lett. 11, 1599ꢀ1603.
(5) Glennon, R. A., Daoud, K. M., Dukat, M., Teitler, M., HerrickꢀDavis, K., Purohit, A., and Syed, H.
(2003) Arylguanidine and arylbiguanide binding at 5ꢀHT3 serotonin receptors: a QSAR study. Bioorg.
Med. Chem. 11, 4449ꢀ4454.
(6) Heesing, A., Schmaldt, W. (1978) Zum substituenteneinfluß der guanidinoꢀ und der
guanidiniumgruppe, teil I: der effekt auf die 19FꢀNMRꢀspektren von fluorbenzolen und auf die
aromatische substitution. Chem. Ber. 111, 320ꢀ334.
(7) Machnitzki, P., Tepper, M., Wenz, K., Stelzer, O., Herdtweck, E. (2000) Waterꢀsoluble phosphines.
Part XII. Pd catalyzed P–C coupling reactions: a novel synthetic route to cationic phosphines with paraꢀ
and metaꢀguanidiniumphenyl moieties. J. Organomet. Chem. 602, 158ꢀ169.
(8) Nandurkar, N. S., Bhor, M. D., Samant, S. D., Bhanage, B. M. (2007) Ultrasoundꢀassisted
regioselective nitration of phenols using dilute nitric acid in a biphasic medium. Ind. Eng. Chem. Res. 46,
8590ꢀ8596.
(9) Christiansen, W. G. (1923) NꢀMethylꢀparaꢀaminoꢀorthoꢀchlorophenol sulfate, a new photographic
developer. J. Am. Chem. Soc. 45, 2192ꢀ2194.
(10) Hughes, J. L., Liu, R. C.; Enkoji, T., Smith, C. M.; Bastian, J. W., Luna, P. D. (1975) Cardiovascular
activity of aromatic guanidine compounds. J. Med. Chem. 18, 1077ꢀ1088.
(11) Ihara, M., Matsuda, K., Shimomura, M., Sattelle, D. B., and Komai, K. (2004) Super agonist actions
of clothianidin and related compounds on the SADβ2 nicotinic acetylcholine receptor expressed in
Xenopus laevis oocytes. Biosci. Biotechnol. Biochem. 68, 761ꢀ763.
(12) Brown, L. A., Ihara, M., Buckingham, S. D., Matsuda, K., and Sattelle, D.B. (2006) Neonicotinoid
insecticides display partial and super agonist actions on native insect nicotinic acetylcholine receptors. J.
Neurochem. 99, 608ꢀ615.
20
ACS Paragon Plus Environment

Page 21 of 36
ACS Chemical Neuroscience
1
2
3
4
5
(13) Langmead, C. J., and Christopoulos, A. (2013) Supraꢀphysiological efficacy at GPCRs: superstition
or super agonists? Br. J. Pharmacol. 169, 353ꢀ356.
6
(14) Price, K. L., Bower, K. S., Thompson, A. J., Lester, H. A., Dougherty, D. A., and Lummis, S. C. R.
(2008) A hydrogen bond in Loop A is critical for the binding and function of the 5ꢀHT3 receptor.
Biochemistry 47, 6370ꢀ6377.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(15) Thompson, A. J., Price, K. L., Reeves, D. C., Chan, S. L., Chau, P.ꢀL., and Lummis, S. C. R. (2005)
Locating an antagonist in the 5ꢀHT3 receptor binding site using modeling and radioligand binding. J. Biol.
Chem. 280, 20476ꢀ20482.
(16) Hazai, E., Joshi, P., Skoviak, E. C., Suryanarayanan, A., Schulte, M. K., and Bikadi, Z. (2009) A
comprehensive study on the 5ꢀhydroxytryptamine3A receptor binding of agonists serotonin and mꢀ
chlorophenylbiguanidine. Bioorg. Med. Chem. 17, 5796ꢀ5805.
(17) Joshi, P. R., Suryanarayanan, A., Hazai, E., Schulte, M. K., Maksay, G., and Bikádi, Z. (2006)
Interactions of granisetron with an agonistꢀfree 5ꢀHT3A receptor model. Biochemistry 45, 1099ꢀ1105.
(18) Maksay, G., Bikádi, Z., and Simonyi, M. (2003) Binding interactions of antagonists with 5ꢀ
hydroxytryptamine3A receptor models. J. Recept. Signal Transduction Res. 23, 255ꢀ270.
(19) Reeves, D. C., Sayed, M. F. R., Chau, P.ꢀL., Price, K. L., and Lummis, S. C. R. (2003) Prediction of
5ꢀHT3 receptor agonistꢀbinding residues using homology modeling. Biophys. J. 84, 2338ꢀ2344.
(20) Steward, L. J., Boess, F. G., Steele, J. A., Liu, D., Wong, N., and Martin, I. L. (2000) Importance of
phenylalanine 107 in agonist recognition by the 5ꢀhydroxytryptamine3A receptor. Mol. Pharmacol. 57,
1249ꢀ1255.
(21) Thompson, A. J., Lochner, M., and Lummis, S. C. R. (2008) Loop B is a major structural component
of the 5ꢀHT3 receptor. Biophys. J. 95, 5728ꢀ5736.
(22) Thompson, A. J., and Lummis, S. C. R. (2006) 5ꢀHT3 receptors. Curr. Pharm. Des. 12, 3615ꢀ3630.
(23) Thompson, A. J., Padgett, C. L., and Lummis, S. C. R. (2006) Mutagenesis and molecular modeling
reveal the importance of the 5ꢀHT3 receptor Fꢀloop. J. Biol. Chem. 281, 16576ꢀ16582.
(24) Yan, D., and White, M. M. (2005) Spatial orientation of granisetron in the ligandꢀbinding site of the
5ꢀHT3 receptor. Mol. Pharmacol. 68, 365ꢀ371.
(25) Barnes, N. M., Hales, T. G., Lummis, S. C. R., Peters, J. A. (2009) The 5ꢀHT3 receptor—The
relationship between structure and function. Neuropharmacology 56, 273–284.
(26) Du, J., Lü, W., Wu, S., Cheng, Y., Gouaux, E. (2015) Glycine receptor mechanism elucidated by
electron cryoꢀmicroscopy. Nature 526, 224–229.
(27) Michaelson, S. D., Paulsen, I. M., Kozuska, J. L., Martin, I. L., Dunn, S. M. J. (2013) Importance of
recognition loops B and D in the activation of human 5ꢀHT3 receptors by 5ꢀHT and metaꢀ
chlorophenylbiguanide. Neuropharmacology 73, 398–403.
21
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 22 of 36
1
2
3
4
5
6
(28) Rohde, L. A. H., Ahring, P. K., Jensen, M. L., Nielsen, E. Ø., Peters, D., Helgstrand, C., Krintel, C.,
Harpsøe, K., Gajhede, M., Kastrup, J. S., Balle, T. (2012) Interꢀsubunit bridge formation governs agonist
efficacy at nicotinic acetylcholine α4β2 receptors: A unique role of halogen bonding revealed. J. Biol.
Chem., 287, 4248–4259.
7
8
9
(29) Althoff, T., Hibbs, R. E., Banerjee, S., Gouaux, E. (2014) Xꢀray structures of GluCl in Apo states
reveal a gating mechanism of Cysꢀloop receptors. Nature 512, 333–337.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(30) Yuan, S., Filipek, S., Vogel, H. (2016) A gating mechanism of the serotonin 5ꢀHT3 Receptor.
Structure 24, 816–825.
(31) Hassaine, G., Deluz, C., Grasso, L., Wyss, R., Tol, M. B., Hovius, R., Graff, A., Stahlberg, H.,
Tomizaki, T., Desmyter, A., Moreau, C., Li, X.ꢀD., Poitevin, F., Vogel, H., Nury, H. (2014) Xꢀray
structure of the mouse serotonin 5ꢀHT3 receptor. Nature 512, 276–281.
(32) Lynagh, T., Pless, S. A. (2014) Principles of agonist recognition at Cysꢀloop receptors. Front.
Physiol. 5, 160.
(33) Kawasaki, Y., Chufan, E. E., Lafont, V., Hidaka, K., Kiso, Y., Amzel, L. M., Freire, E. (2010) How
much binding affinity can be gained by filling a cavity? Chem. Biol. Drug Des. 75, 143–151.
(34) Fischer, F. R., Schweizer, W. B., Diederich, F. (2007) Molecular torsion balances: Evidence for
favorable orthogonal dipolar interactions between organic fluorine and amide groups. Angew. Chem. Int.
Ed. Engl. 46, 8270–8273.
(35) Hof, F., Scofield, D. M., Schweizer, W. B., Diederich, F. (2004) A weak attractive interaction
between oganic fluorine and an amide group. Angew. Chem. Int. Ed. Engl. 43, 5056–5059.
(36) Müller, K., Faeh, C., Diedrich, F. (2007) Fluorine in pharmaceuticals: Looking beyond intuition.
Science 317, 1881–1886.
(37) Kesters, D., Thompson, A. J., Brams, M., van Elk, R., Spurny, R., Geitmann, M., Villalgordo, J. M.,
Guskov, A., Danielson, U. H., Lummis, S. C. R., Smit, A. B., Ulens, C. (2013) Structural basis of ligand
recognition in 5ꢀHT3 receptors. EMBO Rep. 14, 49–56.
(38) Hansen, S. B., Sulzenbacher, G., Huxford, T., Marchot, P., Taylor, P., Bourne, Y. (2005) Structures
of Aplysia AChBP complexes with nicotinic agonists and antagonists reveal distinctive binding interfaces
and conformations. EMBO J. 24, 3635–3646.
(39) Sander, T., Bruun, A. T., Balle, T. (2010) Docking to flexible nicotinic acetylcholine receptors: A
vaidation study using the acetylcholine bindingprotein. J. Mol. Graph. Model. 29, 415–424.
(40) Jones, G., Willett, P., Glen, R. C., Leach, A. R., Taylor, R. (1997) Development and validation of a
genetic algorithm for flexible docking. J. Mol. Biol. 267, 727–748.
(41) Joshi, P. R., Suryanarayanan, A., and Schulte, M. K. (2004) A vertical flow chamber for Xenopus
oocyte electrophysiology and automated drug screening. J. Neurosci. Methods 132, 69ꢀ79.
22
ACS Paragon Plus Environment

Page 23 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
(42) Hope, A. G., Downie, D. L., Sutherland, L., Lambert, J. J., Peters, J. A., and Burchell, B. (1993)
Cloning and functional expression of an apparent splice variant of the murine 5ꢀHT3 receptor A subunit.
Eur. J. Pharmacol. 245, 187ꢀ192.
7
8
9
(43) Yan, D., Schulte, M. K., Bloom, K. E., and White, M. M. (1999) Structural features of the ligandꢀ
binding domain of the serotonin 5HT3 receptor. J. Biol. Chem. 274, 5537ꢀ5541.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(44) Morales, M., and Wang, S.ꢀD. (2002) Differential composition of 5ꢀhydroxytryptamine3 receptors
synthesized in rat CNS and peripheral nervous system. J. Neurosci., 22, 6732ꢀ6741.
(45) van Hooft, J. A., and Yakel, J. L. (2003) 5ꢀHT3 receptors in the CNS: 3B or not 3B? Trends
Pharmacol. Sci. 24, 157ꢀ160.
(46) Brady, C. A., Stanford, I. M., Ali, I., Lin, L., Williams, J. M., Dubin, A. E., Hope, A. G, and Barnes,
N. M. (2001) Pharmacological comparison of human homomeric 5ꢀHT3A receptors versus heteromeric 5ꢀ
HT3A/3B receptors. Neuropharmacology 41, 282ꢀ284.
23
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 24 of 36
1
2
3
4
FIGURE CAPTIONS
5
6
7
8
9
Figure 1. Doseꢀresponse curves for 5ꢀHT and agonists 7−10. The relative response (Yꢀaxis) represents
the current resulting from exposure of oocytes expressing m5ꢀHT3A receptors to the test compound
divided by the Imax value determined for 5ꢀHT. The dashed line indicates the normalized 5ꢀHT Imax value
of 1.0. The efficacies of agonists 7−9 are significantly different from 5ꢀHT (shown in Table 3). The
efficacy of agonist 10 is not statistically different from 5ꢀHT.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Representative traces of agonist 2 (A), superꢀagonist 9 (B), partial agonist 11 (C), and
antagonist 15 (D) in comparison to responses obtained for a control concentration of 5ꢀHT. For A−C, the
control responses are shown at 5ꢀHT concentrations equal to the 5ꢀHT EC90. For inhibition by antagonist
15 the 5ꢀHT concentration for both the control and the inhibited responses is equal to the 5ꢀHT EC50 value
(D). The representative trace shown for each compound is the concentration producing a near Imax
response for 2, 9 and 11 and Imin response for the antagonist 15.
Figure 3. A) Superimposed putative binding modes of superꢀagonist 9 (magenta carbon atoms), full
agonist 2 (cyan), weak partial agonist 11 (white), and antagonist 15 (red) in the orthosteric binding site of
the m5ꢀHT3 receptor. Ligands are rendered as ballꢀandꢀstick figures. The backbone ribbon, carbon atoms
and amino acid labels of the principal subunit are colored tan; those of the complementary subunit are
yellow. Residues of the orthosteric binding site are rendered as spaceꢀfilling, except for C loop residues
F226 and I228, which are rendered as capped sticks for clarity. Important residues peripheral to the
orthosteric binding site are also rendered as capped sticks. Selected Hꢀbond interactions are indicated by
green dashed lines. The calculated distances for the h5ꢀHT3A intraꢀ and interꢀsubunit Hꢀbond interactions
shown in panel 3A are as follows: D189(OD1)–Y143(OH), 2.74 Å; T186(OG1)–Y143(OH), 3.13 Å;
24
ACS Paragon Plus Environment

Page 25 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
D229(OD1)–R92(NH1), 2.94 Å; D229(OD1)–R92(NH2), 2.78 Å; E129(OE2)–T179(N), 2.78 Å;
E236(OE2)–K238(NZ), 2.49 Å. Individual binding modes are shown for B) superꢀagonist 9, C) full
agonist 2, D) weak partial agonist 11, and E) antagonist 15. C loop residues F226 and I228 are not shown
for clarity in panels B to E.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
25
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 26 of 36
1
2
3
4
Scheme 1. Synthesis of arylguanidines 4–6, and 10.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (i) HNO3, 1,2ꢀdichloroethane, sonication; (ii) (a) Na2S2O4, NaOH, H2O, reflux;
(b) HCl, EtOAc; (iii) (a) NH2CN, EtOH, reflux; (b) NH4NO3, H2O.
26
ACS Paragon Plus Environment

Page 27 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1.
27
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 28 of 36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2.
28
ACS Paragon Plus Environment

Page 29 of 36
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3.
29
ACS Paragon Plus Environment