5-Fluorotryptamine is a partial agonist at 5-HT 3 receptors, and reveals that size and electronegativity at the 5 position of tryptamine are critical for efficient receptor function

Antagonists, but not agonists, of the 5-HT 3 receptor are useful therapeutic agents, and it ispossible that partial agonists may also be potentially useful in the clinic. Here we show that 5-fluorotryptamine (5-FT) is a partial agonist at both 5-HT 3A and 5-HT 3AB receptors with an R max ( I max / I max 5-HT) of 0.64 and 0.45 respectively. It is about 10 fold less potent than 5-HT: EC 50 =16 and 27 μ M, and K i for displacement of [ 3 H]granisetron binding=0.8 and 1.8 μ M for 5-HT 3A and 5-HT 3AB receptors respectively. We have also explored the potencies and efficacies of tryptamine and a range of 5-substituted tryptamine derivatives. At 5-HT 3A receptors tryptamine is a weak ( R max= 0.15), low affinity (EC 50 =113 μ M; K i =4.8 μ M) partial agonist, while 5-chlorotryptamine has a similar affinity to 5-FT (EC 50= 8.1 μ M; K i =2.7 μ M) but is a very weak partial agonist ( R max =0. 0037). These, and data from 5-methyltryptamine and 5-methoxytryptamine, reveal the importance of size and electronegativity at this location for efficient channel opening.


Introduction
The 5-HT 3 receptor is a member of Cys-loop family of ligandgated ion channels, which also includes nicotinic acetylcholine, GABA and glycine receptors (Reeves and Lummis, 2002). These proteins are pentamers, and each subunit has a large extracellular N-terminal domain, four transmembrane helices (M1-M4) and an intracellular loop between M3 and M4. The binding site is located at the interface of two adjacent subunits and is formed by the convergence of three loops (A-C) from the principal subunit and another three loops (D-F) from the complementary subunit (Reeves and Lummis, 2002). Molecular details of the binding pocket have been extrapolated from the structure of the acetylcholine binding protein, which is homologous to the extracellular domain of Cys-loop receptors, and a range of amino acid residues that are important for agonist and antagonist binding have been identified (Reeves et al., 2003;Thompson et al., 2005). 5-HT 3 receptors can function as homopentamers of 5-HT 3A receptor subunits, or as heteropentamers of 5-HT 3A and 5-HT 3B receptor subunits (5-HT 3AB receptors). The incorporation of B subunits results in some changes in the biophysical characteristics of the receptor, but has little effect on the pharmacological profile (Brady et al., 2001;Davies et al., 1999;Dubin et al., 1999). 5-HT 3 receptor antagonists have been suggested to be potentially useful in treating inflammatory pain, anxiety, depression, schizophrenia, and drug abuse (Greenshaw and Silverstone, 1997), and are currently in clinical practice for the treatment of irritable bowel syndrome and emesis (Butler et al., 1988;Camilleri et al., 2000;Sanger and Nelson, 1989). It is therefore not surprising that many 5-HT 3 receptor antagonists have been developed. There are, however, fewer 5-HT 3 selective agonists. 2-methyl-5-HT and mCBPG have been widely used, and some novel compounds have been developed more recently such as benzoxazoles (Yoshida et al., 2005) and pyrroloquinoxaline-related compounds (Campiani et al., 1997). Here we explore the agonist properties of a compound closely related to 5-HT, 5-fluorotryptamine (5-FT), at both 5-HT 3A and 5-HT 3AB receptors, and compare them to the properties of 5-HT, mCBPG and tryptamine. We also explore several other 5-substituted tryptamine derivatives.
Harvested stage V-VI Xenopus oocytes were washed in six changes of ND96 (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM HEPES, pH 7.5), de-folliculated in 1.5 mg ml − 1 collagenase Type 1A for approximately 2 h. Enzyme treatment was terminated by washing in six changes of ND96 and oocytes were stored in ND96 containing 2.5 mM sodium pyruvate, 50 mM gentamicin and 0.7 mM theophylline.

Radioligand binding
This was undertaken in HEK 293 cells which provide an established and robust method of studying ligand binding. Methods were as previously described (Lummis et al., 1993), with minor modifications. Briefly, transfected HEK 293 cells were washed twice with phosphate buffered saline (PBS) at room temperature and scraped into 1 ml of ice-cold HEPES buffer (10 mM, pH 7.4) containing the following proteinase inhibitors (PI): 1 mM EDTA, 50 μg ml − 1 soybean trypsin inhibitor, 50 μg/ml bacitracin and 0.1 mM phenylmethylsulphonyl fluoride. After thawing, they were washed with HEPES buffer, resuspended, and 50 μg of cell membranes incubated in 0.5 ml HEPES buffer containing 0.5 nM [ 3 H]granisetron (a concentration approximately equivalent to the K d ); non-specific binding was determined using 10 μM quipazine. Competition binding was performed using ligand concentrations from 0.1 μM-10 mM. Reactions were incubated for at least 1 h at 4°C and terminated by vacuum filtration using a Brandel cell harvester onto GF/B filters presoaked in 0.3% polyethyleneimine. Radioactivity was determined by scintillation counting using a Beckman LS6000SC (Fullerton, California, USA). Competition binding data were analyzed by iterative curve fitting (GraphPad Prism v3.02, GraphPad Software, San Diego, California, USA), according to the equation: where B min is the lowest total binding, B max is the maximum specific binding at equilibrium, [L] is the concentration of competing ligand and IC 50 is the concentration of competing ligand that blocks half of the specific bound radioligand. K i values were estimated from IC 50 values using the Cheng-Prusoff equation: where K i is the equilibrium dissociation constant for binding of the unlabeled antagonist, [L] is the concentration of radioligand and K d is the equilibrium dissociation constant of the radioligand.

Electrophysiology
Agonist-induced currents were recorded at 22-25°C from individual oocytes using the OpusXpress system (Molecular Devices Axon Instruments, Union City, CA). 5-HT, mchlorophenylbiguanide (mCPBG), 5-FT and tryptamine (Sigma) were stored as 20-100 mM aliquots at−20°C, diluted in Ca-free ND96 buffer (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2, 5 mM HEPES, pH 7.5) and delivered to cells via the automated perfusion system of the OpusXpress. Glass microelectrodes were backfilled with 3 M KCl and had a resistance of ∼1 MΩ. The holding potential was − 60 mV. To determine EC 50 values, concentration-response data were fitted to the four-parameter logistic equation, where I max is the maximal response plateau, I min is the minimum response plateau, [A] is the log concentration of agonist and n H is the Hill coefficient, using PRISM v4. 03 software (GraphPad, San Diego, CA). Relative efficacies of the partial agonists are reported as R max = I max drug / I max 5-HT. One-way ANOVAs were performed with Dunnett's post test to determine statistical significance. Data are quoted as mean± SEM (n) unless otherwise stated.

Effects of agonists on 5-HT 3 receptor mediated currents
Application of 5-HT to Xenopus oocytes expressing 5-HT 3A or 5-HT 3AB receptors produced concentration-dependent, rapidly activating, inward currents that desensitised over the time-course of the application (Fig. 2). Plotting current amplitude against a series of 5-HT concentrations revealed EC 50 s of 1.4 μM and 3.2 μM with Hill slopes of 2.5 and 1.4 respectively ( Table 1).
Application of 5-FT to Xenopus oocytes expressing 5-HT 3A or 5-HT 3AB receptors also produced concentration-dependent, rapidly activating, inward currents, with EC 50 s of 16 μM and 27 μM and Hill slopes of 2.4 and 1.4 respectively. A  maximal concentration of 5-FT, however, did not elicit the same maximal currents as those obtained from 5-HT application in the same oocyte, indicating a partial agonist; 5-FT had a R max (I max drug / I max 5-HT) of 0.64 ± 0.03 for 5-HT 3A receptors and R max of 0.45 ± 0.04 for 5-HT 3AB receptors ( Table 2). Application of mCPBG produced concentration-dependent, rapidly activating, inward currents, with EC 50 s of 0.5 μM and 1.1 μM and Hill slopes of 2.3 and 1.6 for 5-HT 3A or 5-HT 3AB receptors, respectively. This compound had an R max of 0.74 ± 0.07 for 5-HT 3A receptors and 0.92 ± 0. 09 for 5-HT 3AB receptors.
Application of tryptamine produced concentration-dependent, rapidly activating, inward currents, but here there was little desensitisation over the time-course of the application (Fig. 2). Plotting current amplitude against a series of tryptamine concentrations revealed EC 50 s of 113 μM and 61 μM with Hill slopes of 2.5 and 1.8 for 5-HT 3A and 5-HT 3AB receptors respectively. Tryptamine had an R max of 0.15 ± 0. 06 for 5-HT 3A receptors and an R max of 0.14 ± 0. 03 for 5-HT 3AB receptors.
5-ClT was a very weak partial agonist of 5-HT 3A receptors, with an R max of 0. 0037; the size of the responses precluded data from 5-HT 3AB receptors. Despite its low R max , 5-ClT had an EC 50 (8.1 ± 0.3 μM, n = 11) that was lower than that of 5-FT (16 μM).
5-MeT was also a very weak partial agonist at 5-HT 3A receptors with an R max of 0. 0023. Dose response curves yielded an EC 50 of 60 ± 3μM (n = 3) indicating it was slightly more potent than tryptamine (EC 50 = 113 μM).
5-MeOT was unable to activate 5-HT 3 receptors at concentrations up to 10 mM.

[ 3 H]granisetron binding studies
Saturation binding studies revealed no significant difference in the affinity (K d ) of [ 3 H]granisetron between 5-HT 3A and 5-HT 3AB receptors (0.42 ± 0.15 and 0.62 ± 0.21 nM respectively, n = 3). Competition binding studies using [ 3 H]granisetron revealed displacement of specific binding in a concentrationdependent manner by all the ligands. K i s (Table 3) revealed that 5-HT, mCPBG, 5FT and tryptamine did not substantially distinguish between 5-HT 3A and 5-HT 3AB receptors.
Competition radioligand binding studies on the mutant receptors N128A, T181A and E236A, revealed no significant changes in K i values compared to WT receptors for either 5-FT or 5-HT (Table 4). E129A and T179A mutant receptors had either no specific radioligand binding, or levels were too low to obtain accurate data as previously reported (Sullivan et al., 2006).

Discussion
The data described here show that 5-FT is a partial agonist at both 5-HT 3A and 5-HT 3AB receptors, with an R max close to 0.5 and an EC 50 about 10 fold higher than 5-HT. Similarly, tryptamine is a partial agonist at both types of receptor, as previously reported for various native and recombinant 5-HT 3 receptors, including those natively expressed in N1E-115 cells, which may possess both 5-HT 3A and 5-HT 3B receptor subunits (van Hooft and Vijverberg, 1996). Tryptamine has a lower potency than both 5-HT and 5-FT (EC 50 10-100 fold higher) and a lower R max , indicating the importance of the group at the 5 position of 5-HT. Further studies on other 5-substituted tryptamine derivatives confirm this hypothesis, and also reveal the importance of size and electronegativity at this location for efficient channel opening.
Subtle differences between 5-HT 3A and 5-HT 3AB receptors have been reported by a number of authors, and were also observed in the current study. Compared to the 5-HT 3A receptor, responses from 5-HT 3AB receptors are smaller and desensitise more rapidly; EC 5O and K d values differ by ∼ 2 fold and there is an ∼ 2 fold decrease in the Hill slope of the dose response curves. There is also a difference in the efficacy of mCPBG, which acts as a partial agonist at 5-HT 3A receptors, but a full agonist at 5-HT 3AB receptors. This indicates gating characteristics of the two receptors are different, and indeed it has been established that the channel conductance is greatly increased in 5-HT 3AB receptors (Davies et al., 1999). 15.5 ± 3.5 5-Cl-tryptamine 2.7 ± 0.7 3.1 ± 1.1 5-Me-tryptamine 11. 0 ± 0.9 7.7 ± 1.1 5-MeO-tryptamine 34.9 ± 3.0 21.7 ± 2.1 Data = mean ± SEM, n =3-6.  Previous functional studies have revealed only small differences in the affinities (EC 50 and IC 50 s) of 5-HT 3A and 5-HT 3AB receptors for a range of 5-HT 3 selective ligands (Brady et al., 2001), and we observed a similar absence of selectivity for 5-HT, mCPBG, 5-FT and tryptamine in this study. These results are somewhat surprising, given that a recent study has suggested that in the heterologously expressed 5-HT 3AB receptors the subunits are in the order BABBA (Barrera et al., 2005), and, as agonist binding sites in Cys-loop receptors are constituted from two adjacent subunits, these data imply that binding interfaces would either be AB (most likely), BA or BB.
Based on the sequence alignment (Fig. 3), one would expect significant structural differences due to the different residues that would contribute to AA compared to AB/BA or BB binding sites. At present, we cannot explain why there are not larger changes in pharmacological characteristics of the AB receptor.
The new data reveal some interesting features of the binding pocket. Tryptamine is ∼ 100 fold less potent and much less efficacious than 5-HT (R max = ∼0.15), establishing the importance of the hydroxyl group. However 5-FT can significantly compensate for the lack of a hydroxyl; it is only 10 fold less potent than 5-HT and R max = ∼ 0.5. In our model of the binding Fig. 3. Alignment of 5-HT 3A and 5-HT 3B subunit sequences. Residues that have similar chemical properties are shown in grey. The binding loops that constitute the binding site are underlined.
pocket (Reeves et al., 2003), the hydroxyl of 5-HT is located in a hydrophilic pocket constituted of Asn128, Glu129, Thr179, Thr181 and Glu236, and it has the potential to hydrogen bond with at least one of these residues (Fig. 4). Mutation of Asn128, Thr181 and Glu236 to Ala results in no significant changes to the 5-HT K i , suggesting that Glu129 and Thr179 are the most likely residues to contribute to hydrogen bonds. However as alanine substitutions at these positions result in poor receptor expression we cannot yet prove this hypothesis. 5-FT can be located in a similar location to 5-HT, but we believe it is unlikely that F also forms hydrogen bonds here. Fluorine is the most electronegative element, and as such it is reluctant to donate a lone pair of electrons to a hydrogen bond donor. As a result, organic fluorine (fluorine bonded to a carbon) hardly ever accepts a hydrogen bond (Dunitz, 2004). Even without a hydrogen bond, however, it appears that an electronegative atom is more favourable than no substituent at all at this location.
To further explore this region of the binding site, we examined 5-ClT, 5-MeT and 5-MeOT in 5-HT 3A receptors. 5-ClT was of similar potency to 5-FT in the functional assays (EC 50 = 8 μM) but was much less effective in opening the channel (R max = 0. 0037). 5-ClT and 5-FT bind to the receptor with similar affinities (K i s are not significantly different), demonstrating there is no relationship between K i or EC 50 and R max Thus it appears that the atom at the 5 position of tryptamine plays a critical role in the conformational changes that result in channel opening. Since both 5-FT and 5-ClT present a relatively electronegative atom at this position, we propose that the increased steric size of Cl vs. F contributes to decreased efficacy of 5-ClT. Sterics also rationalize the inefficacy of 5-MeOT, which has an electronegative element in the 5 position but is apparently too large. The data from 5-MeT also support the hypothesis that size and polarity are important; Me is a similar size to Cl, but is nonpolar, and 5-MeT is less effective at opening the channel.
The data also show that for most agonists there is a direct relationship between EC 50 and K i, with EC 50 s 13-50 fold higher than K i . This is expected, as K i values are considered to represent binding to a high affinity desensitised state. However, for 5-ClT and 5-MeT, which have very low efficacy, EC 50 is less than 5 fold higher than K i . This suggests that if agonist binding does not result in significant channel opening (R max less than 0.01), then there may be no significant entry of receptors into a high affinity state.
Partial agonists are increasingly being used to distinguish between binding and gating events at Cys-loop receptors, and 5-FT, with an R max of ∼ 0.5, will be a useful addition to this class of compounds which includes the more usually used mCBPG (R max = ∼ 0.8) and 2-methyl-5-HT (R max = ∼ 0.2). Partial agonists are also potentially useful as therapeutic agents. The most well-established role of 5-HT 3 receptors is in regulating gastrointestinal motility and the vomiting reflex, although they may play a role in many other neuronal functions. Currently, 5-HT 3 receptor antagonists are used clinically as antiemetics, and to treat irritable bowel syndrome (Butler et al., 1988;Camilleri et al., 2000;Sanger and Nelson, 1989). However, there is some evidence that these compounds also cause side effects in many patients, by inhibiting normal lower bowel function (Talley, 1992). Thus there has been an increased interest in 5-HT 3 receptor partial agonists which might control gastroenteric motility without completely blocking 5-HT 3sensitized nerve function (Sato et al., 1997;Sato et al., 1998). 5-HT 3 receptor agonists also have a potential therapeutic role as they can modulate acetylcholine release in vivo (Consolo et al., 1994), making these compounds of interest for the treatment of neurodegenerative and neuropsychiatric disorders in which cholinergic neurons are affected. Full 5-HT 3 receptor agonists, however, cause nausea and vomiting; thus partial agonists are potentially more useful for therapeutic applications in this area. Recently developed compounds, e.g those described by Yoshida et al. (2005), are probably potentially more useful as therapeutics than 5-FT, but a comparison of their actions compared to 5-FT may clarify details of their mode(s) of action.
In conclusion we have shown that 5-FT is a partial agonist at both homomeric 5-HT 3A and heteromultimeric 5-HT 3AB receptors. The data have also revealed that the atom in the 5 position of 5-HT plays an important role both in receptor binding and in subsequent channel gating. Fig. 4. 5-HT docked into a homology of the 5-HT 3 receptor (Reeves et al., 2003). A. The extracellular domains of two subunits of the 5-HT 3 receptor showing the location of the binding pocket (boxed) at their interface. B. Enlarged image of the binding site showing the proximity of the hydroxyl group of 5-HT to the hydrophilic residues Asn128, Glu129 Thr179, Thr181 and Glu236.