1. *Molecular Discovery Research, GlaxoSmithKline, Harlow, Essex, UK
Neurology and Gastrointestinal Centre of Excellence for Drug Discovery, GlaxoSmithKline, Harlow, Essex, UK
5-HT3 receptors are therapeutically important members of
the superfamily of Cys-loop ligand-gated ion channels which
includes the nicotinic acetylcholine receptors (nAChRs),
GABAA, GABAC and glycine receptors (Reeves and Lum-
mis 2002; Peters et al. 2005). The 5-HT3 receptor exists as a
pentameric ring of subunits that form an integral ion channel
along the central axis. To date, functional channels composed
of a homomeric assembly of 5-HT3A (Maricq et al. 1991;
Miyake et al. 1995); or heteromeric assembly of 5-HT3A and
5-HT3B subunits (5-HT3AB) (Davies et al. 1999) have been
described. Each exhibits distinct channel characteristics
(single channel conductance, rectification), but little differ-
ence in pharmacology, at least in terms of responses to 5-
hydroxytryptamine (5-HT) or 5-HT3 receptor antagonists
(Reeves and Lummis 2002; Peters et al. 2005), although
small differences in responses to tubocurarine or picrotoxin
are observed (Davies et al. 1999; Brady et al. 2001; Das and
Dillon 2003). These findings do not, however, explain the
apparent pharmacological diversity of responses mediated by
the 5-HT3 receptor within native tissues.
Received August 18, 2008; revised manuscript received/accepted
October 28, 2008.
Address correspondence and reprint requests to Martin Gunthorpe,
Neurology and Gastrointestinal Centre of Excellence for Drug Discov-
ery, GlaxoSmithKline, New Frontiers Science Park (North), Third
Avenue, Harlow, CM19 5AW, UK.
E-mail: Martin_J_Gunthorpe@gsk.com
Abbreviations used: 5-HT, 5-hydroxytryptamine; CFP, cyan fluorescent
protein; CHO, Chinese hamster ovary; DRG, Dorsal Root Ganglion; GI,
gastrointestinal; HA, Hemagglutinin; TM, transmembrane; WT, wild type.
Abstract
The 5-HT3 receptor is a member of the ‘Cys-loop’ family of
ligand-gated ion channels that mediate fast excitatory and
inhibitory transmission in the nervous system. Current evi-
dence points towards native 5-HT3 receptors originating from
homomeric assemblies of 5-HT3A or heteromeric assembly of
5-HT3A and 5-HT3B. Novel genes encoding 5-HT3C, 5-HT3D,
and 5-HT3E have recently been described but the functional
importance of these proteins is unknown. In the present study,
in silico analysis (confirmed by partial cloning) indicated that 5-
HT3C, 5-HT3D, and 5-HT3E are not human–specific as previ-
ously reported: they are conserved in multiple mammalian
species but are absent in rodents. Expression profiles of the
novel human genes indicated high levels in the gastrointesti-
nal tract but also in the brain, Dorsal Root Ganglion (DRG)
and other tissues. Following the demonstration that these
subunits are expressed at the cell membrane, the functional
properties of the recombinant human subunits were investi-
gated using patch clamp electrophysiology. 5-HT3C, 5-HT3D,
and 5-HT3E were all non-functional when expressed alone.
Co-transfection studies to determine potential novel hetero-
meric receptor interactions with 5-HT3A demonstrated that the
expression or function of the receptor was modified by 5-HT3C
and 5-HT3E, but not 5-HT3D. The lack of distinct effects on
current rectification, kinetics or pharmacology of 5-HT3A
receptors does not however provide unequivocal evidence to
support a direct contribution of 5-HT3C or 5-HT3E to the lining
of the ion channel pore of novel heteromeric receptors. The
functional and pharmacological contributions of these novel
subunits to human biology and diseases such as irritable
bowel syndrome for which 5-HT3 receptor antagonists have
major clinical usage, therefore remains to be fully determined.
Keywords: 5-HT3, cloning, electrophysiology, ion channel,
pharmacology, serotonin.
J. Neurochem. (2009) 108, 384–396.
JOURNAL OF NEUROCHEMISTRY | 2009 | 108 | 384–396 doi: 10.1111/j.1471-4159.2008.05775.x
384 Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
2. The term ‘5-HT3–like’ receptor was introduced to describe
certain responses to 5-HT3 receptor antagonists which did
not fit the expected activity. For example, in one study, all 5-
HT3 receptor antagonists tested potently inhibited the von
Bezold-Jarisch cardiovascular reflex induced by 5-HT in
anaesthetised rats, but some of these were able to reduce a
visceral pain reflex in the same species (Banner et al. 1995).
This and other differences in responses to 5-HT3 receptor
antagonists, always described in native tissues (in vitro and
in vivo), are still not adequately explained (Sanger 1995).
Although there are many possible explanations, including
non-selectivity of action, the discovery of further subunits of
the 5-HT3 receptor now provides an opportunity to revisit the
proposal that ‘5-HT3-like’ receptors might exist, with
pharmacology distinct from the accepted 5-HT3A and
5-HT3AB receptor forms that have been extensively charac-
terized (Reeves and Lummis 2002).
In 2003, three human genes encoding novel 5-HT3C, 5-
HT3D and 5-HT3E receptor subunits were reported (Karnov-
sky et al. 2003; Niesler et al. 2003). The genes for these
subunits are located on human chromosome 3q27 whilst
those for 5-HT3A and 5-HT3B are on chromosome 11q23.
Further, the reported mRNA expression of the 5-HT3C-E
subunits tended to show a peripherally-restricted pattern,
including high levels in the gastrointestinal (GI) tract (Niesler
et al. 2003). Intriguingly, their orthologues seem to be absent
in rodents (Karnovsky et al. 2003; Niesler et al. 2003).
Together, these data suggest that 5-HT3C-E subunits are
formed for a specific purpose within the peripheral nervous
systems in some species, including humans, which is both
distinct from the more general functions of receptors formed
by 5-HT3A and 5-HT3B.
To gain insight into their functional properties, and
possible physiological roles, we used RT-PCR from cDNA
libraries to further elucidate the distribution of the mRNA’s
encoding 5-HT3C-E in human tissues. We discovered mam-
malian orthologues of 5-HT3C, 5-HT3D and 5-HT3E by
in silico searches of genomic sequence and PCR cloning
from ferret and rabbit cDNA libraries. We report definitive
evidence that the genes are not human specific, as previously
suggested (Niesler et al. 2003) but, rather, appear to be lost
specifically in the rodent lineage. Following the cloning of 5-
HT3C-E from human tissue sources, we show that C-terminal
tagged subunits localise to the plasma membrane and, using
whole-cell patch clamp electrophysiology, we assessed the
functional and pharmacological properties of the subunits as
candidate homomeric and heteromeric channels.
Experimental procedures
Sequence searches for subunit orthologues
To explore species distribution NCBI databases were homology
searched using BLAST. Sequences predicted to encode HTR3C,
HTR3D and HTR3E were identified and peptide sequences predicted
using genewise (Birney et al. 2004). Phylogenetic analysis was
performed using neighbour joining methodology implemented in
PAUP (v4b10 for UNIX) (Wilgenbusch and Swofford 2003).
Cloning of ferret and rabbit orthologues
cDNA encoding 5-HT3C and 5-HT3E was PCR-amplified from
libraries derived from ferret tissues using degenerate primers based
on dog sequence. Products from lung, spleen, liver and uterus were
cloned, sequenced and subjected to homology searches. A partial
fragment of rabbit HTR3C cDNA was isolated from rabbit colon and
small intestine libraries by RT-PCR using degenerate primers based
on an alignment of human, dog and elephant HTR3C DNA
sequences. The 5¢end of the fragment was extended by 5¢-rapid
amplification of cDNA ends using rabbit colon and small intestine
SMART rapid amplification of cDNA ends (Clontech, Mountain
View, CA, USA) cDNA libraries.
Cloning of human 5-HT3C, 5-HT3D and 5-HT3E cDNAs
HTR3C, HTR3D and HTR3E were amplified by RT-PCR from
human intestine, kidney and colon Marathon-Ready cDNAs (BD
Transduction Laboratories, Lexington, KY, USA), respectively,
sequenced and subcloned into expression vectors. Appropriate
PCR primers were used for each variant to replace the native stop
codon with a 7–8 glycine spacer followed by Hemagglutinin (HA)
or FLAG tag and a newly created stop codon immediately after the
tag (Invitrogen, Carlsbad, CA, USA). Tagged subunits were
subcloned into the gateway destination vector pFastBacMam and
fully sequenced.
Distribution of mRNA in human tissues
The distribution of the HTR3A–E subunits was assessed by TaqMan
RT-PCR (Holland et al. 1991) using transcript-specific primers and
probes (see Table S1). The mRNA based masterplates were derived
from human tissues obtained from multiple vendors and adhered to
GlaxoSmithKline (Brentford, UK) standards set for ethical use of
human tissue.
Tissue culture and transient transfections
Chinese Hamster Ovary (CHO)-K1 or Human Embryonic Kidney
293 (HEK293) cell lines were grown in D-MEM/F12 or D-MEM
media, respectively, supplemented with 10% foetal bovine serum
and 2 mM L-glutamine and maintained in a humidified incubator at
37°C with 5% CO2. Transient transfections were carried out using
lipofectamine 2000 (Invitrogen) according to the manufacturer’s
instructions. After 4 h in the presence of the lipofectamine/DNA
mix, the medium was replaced with the appropriate complete
medium and the cells were placed back in the incubator for another
40–48 h. For electrophysiological studies however, cells were
trypsinised and re-plated at a density of 4 · 104
cells/mL on sterile
polylysine-coated glass coverslips (BD Biosciences) prior to study
24–48 h post-transfection.
Electrophysiological investigation of subunit function
CHO or HEK293 cells were transiently transfected with 5-HT3
subunits, singly or in heteromeric combinations, with green
fluorescent protein (GFP) included as a visual marker of successful
transfection [in some experiments cyan fluorescent protein (CFP,
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 385
3. Clontech) was included as a control for effects potentially resulting
from an additional translational burden on the cell], and the
properties of the resultant channels investigated using standard
whole-cell patch-clamp methods. Recordings were performed at
room temperature (20–24°C) using an Axopatch 200B amplifier
controlled via the pClamp9 software suite (Axon Instruments,
Union City, CA, USA). The intracellular and extracellular solutions
used were (mM): CsCl 140, MgCl2 4, EGTA 10, HEPES 10 and
NaCl 130, KCl 5, CaCl2, MgCl2 1, Glucose 30, HEPES 25,
respectively (both pH 7.3). Cells were voltage-clamped at )60 mV,
unless otherwise stated, and agonists and antagonists were applied
to cells via an automated fast-switching solution exchange system
(SF-77B, Warner Instruments, Hamden, CT, USA). Current-voltage
relationships were established using a voltage-ramp protocol ()60
to +60 mV, 1.2 mV/s, 100 ms). Data were analysed using Clampfit
(Axon Instruments), Excel (Microsoft) and Origin (Microcal).
Unless otherwise stated, data are presented as mean ± SEM and
Student’s t-test was used to assess statistical significance (p < 0.05
being regarded as significant).
Indirect immunocytochemical analysis of tagged 5-HT3 subunits
Cells were analysed in suspension 48 h post-transfection using
standard immunocytochemistry under non-permeabilized condi-
tions. All staining steps were carried out on ice to minimise
internalisation. Immunostaining was carried out using primary rabbit
polyclonal anti-HA or FLAG tag (Sigma, St Louis, MO, USA) and
secondary Alexa fluor488-coupled goat anti-rabbit IgG (Invitrogen)
antibodies. Cell nuclei were stained with Hoechst reagent according
to the manufacturer’s instructions (Invitrogen). Cells were mounted
in Immu-Mount (Thermo Shandon, Waltham, MA, USA) and
analyzed using epifluorescence (Zeiss-Axioskop 2, Maple Grove,
MN, USA) and laser scanning (Leica TCS SP, Bannockburn, IL,
USA) microscopes.
Results
Prediction and detection of species orthologues
Coding sequences for the putative orthologues of the 5-
HT3C, and 5-HT3E subunits from dog and chimpanzee
were predicted from genomic DNA and their orthology
confirmed by phylogenetic analysis (Fig. 1). The chimpan-
zee and dog 5-HT3D orthologues were predicted to include
stop codons in their open reading frames. This, together
with the long branch-lengths and incongruent position of
dog 5-HT3D in the phylogenetic analysis suggests that
5-HT3D may be a pseudogene in these species. Mouse and
rat genomic sequences were BLASTed for sequences
homologous to 5-HT3C, 5-HT3D and 5-HT3E. No sequence
more homologous than the mouse and rat 5-HT3A and
5-HT3B genes were found. Rodent orthologues of 5-HT3C-E
were also searched for in the syntenic region of the mouse
and rat chromosomes (Using ENSEMBL AlignSplice-
View http://www.ensembl.org/index.html) but no coding
sequences were returned.
To explore the species distribution of the subunits,
mRNA encoding 5-HT3C and 5-HT3E was RT-PCR ampli-
fied from cDNA derived from ferret tissues. Products of
similar size were seen for colon and lung: the HTR3C
product was 78% identical to dog HTR3C whilst the
sequence from the HTR3E product was 96% identical to
dog HTR3E (Fig. S1). In addition, HTR3C was partially
cloned from rabbit colon yielding a sequence 73% identical
to human HTR3C at the protein level (Fig. S1). Amplifi-
cation of HTR3C from guinea pig tissues was attempted but
HTR3A human
HTR3A chimp
HTR3A dog
HTR3A rat
HTR3A mouseHTR3B ratHTR3B mouse
HTR3B chimp
HTR3B dog
HTR3B human
HTR3D chimp
HTR3D human
HTR3E chimp
HTR3E human
HTR3D dog
HTR3C dog
HTR3C human
HTR3C chimpHTR3E dog
Fig. 1 Neighbour-joining phylogenetic tree,
showing the relationships between the de-
duced peptide sequences of the novel hu-
man 5-HT3 subunits and their orthologues
in chimpanzee, mouse, rat and dog. Chim-
panzee and dog sequences were predicted
from genomic sequence whereas human,
mouse and rat were based on cDNA.
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
386 | J. D. Holbrook et al.
4. extensive efforts failed, suggesting the subunits are also
absent from this species.
Cloning and mRNA distribution of the novel subunits and
their splice variants
In human tissues, we confirmed the existence of the long and
short splice variants of HTR3A previously reported (Hope
et al. 1993; Downie et al. 1994; Belelli et al. 1995); the
longer variant (HTR3AL) has an insertion of 32 aa relative to
the shorter variant (HTR3AS; utilised in the functional
studies in this paper) within the predicted transmembrane
(TM) 3 region (Fig. S2). In addition a putative third variant
of HTR3A (HTR3Aext) was found, represented in the
sequence databases (bt007204) with an alternative upstream
translation start site that would extend the N-terminus by six
amino acids (Fig. S2). The mRNA tissue distribution of
HTR3AS and HTR3AL were assayed by TaqMan (Figs 2
and S3) indicating a relatively high level of expression in
DRG, with both variants also clearly detectable in the GI
tract (colon, duodenum, ileum and jejunum), lymph nodes
and tonsils.
There are three known splice variants of HTR3B (Tzvet-
kov et al. 2007) which affect the translation initiation site
and therefore the N-terminus of the protein. One of these
HTR3A
0
100
200
300
400
HTR3B
0
50
100
HTR3C
0
100
200
300
400
0
100
200
300
HTR3D
100
200
300
0
Normalisedabundance
Mammarygland
Salivarygland
HTR3E
Adipose-O
Adrenal
Aorta-A
Aorta-T
Artery-CA
Artery-CO
Artery-F
Artery-P
Bladder-D
Bladder-DO
Blood
Bonemarrow
Brain-CB
Brain-FL
Brain-HY
Brain-OL
Brain-PL
Brain-TL
Brain-T
Breast
Cecum
Cervix
Colon-A
Colon-D
DRG-L
DRG-T
Gallbladder
Heart-LA
Heart-LV
Heart-MV
Heart-RA
Heart-RV
Heart-S
Kidney-C
Kidney-M
Liver
Lung
Lymphnode
Muscle
Pancreas
Pituitary
Prostate
Rectum
Skin
Duodenum
Ileum
Jejenum
Spinalcord-C
Spinalcord-T
Stomach
Stomach-C
Synovium
Testis
Thyroid
Tonsil
Trachea
Urethra
Uterus
Vein-S
Vein-V
Adipose-Q
Fig. 2 Expression profiles of 5-HT3 receptor subunits in human tis-
sues assessed by TaqMan RT-PCR. Relatively high levels of the novel
subunits are detected in the DRG, GI tract and brain but subtype
differences are evident (see text & supplementary information). The
horizontal axis identifies the tissue from which the RNA was extracted
and the vertical axis indicates copies of mRNA detected per 50
nanogram of total RNA.
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 387
5. variants (HTR3B_V2) lacks the b1-b2 loop and therefore is
unlikely to be functional and could operate as a dominant-
negative. Taqman analysis of the longest variant (HTR3B)
identified a relatively ubiquitous expression pattern with
substantial expression in DRG, brain, kidney, liver, lung,
testis and the gastrointestinal tract (Fig. 2) whereas the
shorter variant (HTR3B_V1) resulted in a brain-specific
expression pattern (Fig. S3). The longest variant was used
for the functional studies in this paper.
Initial attempts to clone human 5-HT3C resulted in the
identification of multiple PCR products of different sizes
from a range of tissues (Fig. S4). In particular, products from
intestine and colon were larger than bands from liver and
lung, sequence verification demonstrated that the only full
length variant was from human intestine with other variants
likely to arise by exon deletion. We note that two of the
variants detected in human lung lack the Cys-loop region and
hence are likely to be non-functional. TaqMan analysis of
HTR3C using primers to the Cys-loop containing exon 5,
which should detect all functional variants, identified highest
expression levels in the DRG, lung and duodenum (Fig. S3 –
HTR3C-2) whereas primers designed to exons 2 and 6
uncovered more robust expression in the GI tract especially
the cecum as well as DRG and lung. The longer ‘GI tract
variant’ was characterised further in this study (Fig. 2).
There are two known splice variants of 5-HT3D, the longer
variant (HTR3DL, Fig. S2) has a 5¢ extension in the first
coding exon, an inserted second exon and then a shorter exon
3. Therefore, the two variants have very different peptide
sequences at the N-terminus. The shorter variant (HTR3Ds,
Fig. S2) seems to be the most frequently reported and was
used in this study. Its mRNA expression levels are low in
most tissues but expression is detectable in DRG and a
variety of other tissues (Fig. 2). Primers designed to detect
the longer variant returned very low expression levels
(Fig. S3). Other known variants lack the Cys-loop region
and so are predicted to be non-functional. 5-HT3E also has
three known variants that differ in coding exons 1, 2 and 3
(Fig. S2). Variant 3 is not represented in the public sequence
databases but was cloned within this study. Given that it was
the most readily cloned and contains all functional motifs it
was used in the further studies described in this paper.
TaqMan analysis demonstrated that it was ubiquitously
expressed across a wide range of tissues including the GI
tract and was present at especially high levels in the DRG,
brain and pituitary gland (Fig. 2).
Functional characterisation of novel 5-HT3 receptor
subunits
To assess whether the newly identified subunits form
receptors at the plasma membrane, C-term HA- or FLAG-
tagged constructs were generated. Based on the structure of
the homologous nicotinic acetylcholine receptor (Unwin
2005) and the 3D model of the 5-HT3 receptor (Lummis
et al. 2005), we hypothesized that tags positioned at the
C-term, rather than N-term, would cause least interference
with receptor function and provide greater accessibility to
antibodies for immunocytochemical analysis. Indeed, in
independent experiments, a 5-HT3A-lumio C-term tagged
construct gave rise to robust 5-HT-gated responses indistin-
guishable from the wild type (WT) receptor (C. Gill,
unpublished observations). Mammalian expression vectors
for 5-HT3B-HA, 5-HT3C-FLAG, 5-HT3D-FLAG and 5-HT3E-
FLAG were therefore generated with an extra seven or eight
glycine residues included between the receptor sequence and
tag to allow for maximum flexibility and accessibility to
antibodies at the extracellular surface (no additional amino
acids were introduced within the 5-HT3 coding sequence).
Immunocytochemical analysis of the various tagged
subunits using standard fluorescent microscopy approaches
indicated expression at the cell periphery in a subset of cells
(Fig. 3a). This staining pattern is as expected for the
transient transfection approach taken and is consistent with
plasma membrane localisation of each of the tagged variants.
This result was confirmed by confocal microscopy showing
a circumferential staining (although not continuous in
certain instances), consistent with membrane-bound expres-
sion for each variant (Fig. 3b). Only background staining
was seen in neighbouring untransfected cells (Fig. 3a), or in
cells transfected with the empty vector alone (data not
shown). These data suggest that at least a proportion of the
subunits assemble as receptors at the plasma membrane.
(a) (b)
5-HT3C
5-HT3D
5-HT3E
5-HT3B
Fig. 3 Immunocytochemical analysis of the tagged 5-HT3 receptor
subunits 5-HT3B-HA, 5-HT3C-FLAG, 5-HT3D-FLAG and 5-HT3E-FLAG
in CHO cells. (a) Conventional fluorescence microscopy. Panels show
representative fields of view identifying the respective 5-HT3 receptors
(green fluorescence; left panels) and Hoechst nucleic acid staining
(blue fluorescence; right panels) overlaid to visualise neighbouring
non-transfected cells. (b) Laser scanning confocal microscopy. Rep-
resentative confocal images showing positive staining for HA or FLAG
tag. Arrows in (a) indicate plasma membrane-bound immunoreactive
signals.
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
388 | J. D. Holbrook et al.
6. Attempts to further characterise the different homomeric
5-HT3 receptors at the protein level were made by western
blotting using corresponding anti-tag antibodies (see
Fig. S5). 5-HT3B-HA, 5-HT3C-FLAG and 5-HT3E-FLAG
all showed abnormal electrophoretic profiles, consisting of
multiple distinctive bands with molecular weights lower than
their predicted molecular weights of 50–60 kDa, potentially
indicating that additional mechanisms or accessory proteins
may be required for achieving or maintaining receptor
integrity.
To determine if the novel 5-HT3 subunits were capable of
forming functional homomeric receptors, we expressed each
subunit individually in CHO cells using green fluorescent
protein as a marker of transfection and used whole-cell patch
clamp electrophysiology to examine 5-HT-evoked responses.
In all experiments, a comparatively high test concentration of
5-HT was applied (50-fold EC50 at 5-HT3A) and main-
tained for at least 20 s, and parallel experiments with 5-HT3A
were used as a positive control. Robust responses were
readily identifiable in 5-HT3A transfected cells in response to
30 lM 5-HT. In contrast, no functional responses to 100 lM
5-HT were detected in WT cells or those transfected with 5-
HT3B or any of the novel subunits (Fig. 4) indicating that the
novel genes do not form functional 5-HT-gated homomeric
receptors.
To begin to assess the ability of the novel subunits to form
heteromeric receptors we first compared the 5-HT responses
of cells transiently transfected with either 5-HT3A or 5-
HT3A + 5-HT3B to define suitable conditions for the clear
identification of known heteromeric receptors. Transfection
of cells with 5-HT3A + 5-HT3B led to currents with a similar
overall magnitude and appearance to 5-HT3A receptors
(Fig. 5), however, these responses activated significantly
more rapidly (Rise time 1758 ± 139 compared to
899 ± 38 ms, for 5-HT3A versus 5-HT3AB receptors, respec-
tively) and showed linear rather than inward rectification
[Rectification ratio (I+60mV/I)60 mV) = 1.02 ± 0.04 (n = 5)
rather than 0.34 ± 0.01(n = 4); reversal potentials were
unchanged; Fig. 5] indicative of the formation of functional
heteromeric 5-HT3AB receptors (Davies et al. 1999; Das and
Dillon 2003). Following a recent report demonstrating that
mouse 5-HT3AB receptors were less sensitive to picrotoxin
blockade than 5-HT3A (Das and Dillon 2003), we also
examined this aspect of human receptor pharmacology. We
found that picrotoxin inhibits the human 5-HT3A receptor
maximally at 300 lM (99 ± 0.3% block) yet only resulted in
84 ± 2% block in cells transfected with 5-HT3A + 5-HT3B
(Fig. 5d). Overall, these data are consistent with the expres-
sion of functional human 5-HT3AB receptors providing
confidence in the methods employed to detect novel 5-HT3
receptor heteromerisation; our results also suggest that
picrotoxin is a less potent inhibitor at the human compared
to mouse 5-HT3 receptors (Das and Dillon 2003, 2005).
Thus, we next compared recordings from cells transiently
transfected with either 5-HT3A or 5-HT3A + 5-HT3C. 5-HT-
evoked responses in cells transfected with 5-HT3A + 5-HT3C
were similar in appearance but significantly smaller than
currents recorded in cells transfected with 5-HT3A alone
(Fig. 6); this difference was apparent in a comparison of all
cells studied or the subset of responding cells only. Further
examination of the properties of the 5-HT responses
including the response rise-time, current-voltage relationship
or degree of blockade by picrotoxin did not, however,
indicate any other clear signature suggestive of the occur-
rence of a novel heteromeric receptor with distinct biophys-
ical or pharmacological properties (Fig. 6).
The 5-HT3C experiments contained a DNA control,
namely inclusion of an equivalent amount of empty
pcDNA3.1 vector in the 5-HT3A transfection to equate to
the amount of 5-HT3C used in the co-transfection experiment
(see legend to Fig. 6). To provide additional confidence in
the conclusion that the reduction in 5-HT3A function by 5-
HT3C was a specific effect of this subunit we also incorpo-
rated a further ‘translational control’ in parallel with our
experiments with 5-HT3D. In these studies, conducted in
HEK293 cells, CFP was included in place of the test subunit
(a) (b)
5-HT3A
5-HT3B
5-HT3C
5-HT3D
5-HT3E
30 µM 5-HT
20 s
2 nA
20 s
1 nA
100 µM 5-HT 100 µM 5-HT
20 s
100 µM 5-HT
2 nA
20 s
100 µM 5-HT
20 s
1 nA2 nA
Fig. 4 Functional assessment of 5-HT3 receptors expressed in CHO
cells. Whole-cell patch clamp recordings of cells transiently trans-
fected with the subunit of interest were visually identified using GFP
and test responses to 30 or 100 lM 5-HT were recorded to assess
functionality. (a) robust 5-HT3A responses were readily identifiable
upon application of 30 lM 5-HT. Seven of eight cells responded with a
mean inward current of 1677 ± 1166 pA (Range 164–4086 pA). (b) No
functional responses were detected in cells transfected with 5-HT3B or
any of the novel subunits. Mean currents were 1 ± 5, 0 ± 1, 1 ± 4, and
2 ± 2 pA for 5-HT3B, 5-HT3C, 5-HT3D, and 5-HT3E, respectively (Range
was £ 10 pA for all recordings; n ‡ 4 for each subunit).
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 389
7. to determine if an increased translational load on the cell
because of the additionally expressed protein may impact 5-
HT3A expression. In each case, the responses to 5-HT were
not significantly different and there were no apparent
differences in current kinetics or current-voltage relationships
(Fig. 7). Responses detected in cells from all three groups
also showed similar sensitivity to 30 lM picrotoxin (Fig. 7).
We therefore conclude that 5-HT3D has no detectable effect
on 5-HT3A function and, based on the CFP data, any effects
detectable on peak current with any of the novel test subunits
are unlikely to be because of the effects of altered DNA or
protein load on the cell.
Our final set of heteromer experiments conducted with 5-
HT3E (Fig. 8) indicated a clear trend towards a decrease in
peak current on co-transfection of the 5-HT3E subunit.
Comparison of the current responses indicated that there
were no apparent difference in kinetics and current-voltage
relationships did not indicate any alteration in rectification
ratio or reversal potential. Furthermore, no change in
picrotoxin sensitivity was detected (Fig. 8).
Discussion
In this study we have used a range of in silico and in vitro
approaches to assess the roles of the novel 5-HT3C-E
subunits, with particular emphasis on human physiology.
All three genes encode the four TM domains found in the 5-
HT3A and 5-HT3B subunits whereas the cysteine loop region
(a defining motif of the mammalian ‘Cys-Loop’ family) is
conserved in 5-HT3C and 5-HT3E but absent from 5-HT3D
likely rendering this subunit non-functional. Similarly, the
orthologous 5-HT3D sequences from the chimpanzee and dog
were predicted to code for open reading frames including
stop codons, which suggests that in these species 5-HT3D is a
pseudogene.
Contrary to Niesler et al. (2003) but in line with the
Southern Zooblot finding of Karnovsky et al. (2003), we
found that 5-HT3C and 5-HT3E are not human specific.
However, we confirm the apparent loss of these genes from
the rodent lineage. This is particularly interesting because
the rodent digestive system differs dramatically from that
3 µM 5-HT
4 nA
3 s
I = 0 pA
V
–60 mV
+60 mV
I
V
5-HT3A
5-HT3A + 3B
Control
3 µM 5-HT
300 µM
picrotoxin
Wash
6 s
2 nA
5-HT3A
200 pA
4 s
INormalised
–60 –40 –20 20 40 60
–1.0
–0.5
0.5
1.0
5-HT3A
5-HT3A + 3B
Vh mV
[Picrotoxin]
0
20
40
60
80
100
%responseinsensitive
toblockadebypicrotoxin
5-HT
*
*
**
**
**
3A
5-HT3A + 3B
300 µM30 µM 100 µM
P 0.05
P 0.01
(a) (b)
(c) (d)
Fig. 5 Characterisation of heteromeric 5-HT3AB receptors. Whole-cell
patch clamp recordings of CHO cells transiently transfected with either
5-HT3A or 5-HT3A + 5-HT3B (1 : 1 ratio; GFP included as a positive
marker of transfection) were used to ascertain hallmarks of receptor
heteromerisation. (a) Current-voltage relationships were established
using a ramp protocol as shown, which was timed to coincide with the
peak of the 5-HT-gated current. (b) Current-voltage relationships
established in cells transfected with 5-HT3A were inwardly rectifying
[I+60 mV/I)60 mV = 0.34 ± 0.01 (n = 4)], whereas those transfected with
5-HT3A + 5-HT3B were linear [I+60mV/I)60 mV = 1.02 ± 0.04 (n = 5)].
Reversal potentials were not significantly different: )2.7 ± 0.6 for
5-HT3A and )5.3 ± 0.5 mV for 5-HT3B. (c) 5-HT responses showed
differential sensitivity to picrotoxin: % block was near maximal with
300 lM picrotoxin versus 5-HT3A receptors (99 ± 0.3%) yet only
resulted in 84 ± 2% (n ‡ 3; p 0.01) block in cells transfected with
5-HT3A + 5-HT3B. (d) The occurrence of a significant (p 0.05)
picrotoxin-insensitive component of 5-HT activated-currents, over the
concentration range 30–300 lM, is consistent with the expression of
heteromeric 5-HT3AB receptors (Das and Dillon 2003).
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
390 | J. D. Holbrook et al.
8. of human. Rodents are unable to vomit (a mechanism
activated via 5-HT3 receptors in other species; (Malik et al.
2006) and excrete faeces as dry pellets, in contrast to most
other mammals; rabbits excrete faecal pellets but exhibit
coprophagia to fully digest and absorb the water content
e.g. (Ebino et al. 1993). These differences in GI physiology
could be related to the loss of the 5-HT3C and 5-HT3E
genes within the rodent lineage. Clearly, the lack of a
full compliment of 5-HT3 receptor subunits suggests
that conclusions drawn from studies on 5-HT3 receptor
function in these species will not necessarily translate to
humans.
We were unable to PCR-amplify guinea-pig orthologues of
5-HT3C and 5-HT3E subunits, however, a small amount of
guinea-pig genomic sequence (AAKN01387346), which is
closely homologous to 5-HT3C does exist in public databases
and hence may represent a pseudogene. It therefore seems
likely that functional 5-HT3C and 5-HT3E subunits were lost
before separation of the guinea-pig and muroid rodent
lineages. It is noteworthy that our identification of coding
sequences for 5-HT3C and 5-HT3E in rabbits, ferrets and dog
identifies potential laboratory species for future studies into
the functions of these genes.
In humans, we confirmed that 5-HT3C and 5-HT3E mRNA
is expressed in tissues of the GI tract. However, we also note
that the distribution of 5-HT3C transcripts may be more
skewed in favour of the gut than previously suggested
(Niesler et al. 2003) indicating a potentially prominent role
of 5-HT3C and 5-HT3E here. Such an hypothesis is supported
by a recent report of high and approximately equal levels of
5-HT3C and 5-HT3E mRNA in human duodenum mucosa,
compared with low levels of 5-HT3A and 5-HT3B expression;
the authors argued that, in this tissue, the 5-HT3CE heteromer
represents the predominant 5-HT3 receptor (van Lelyveld
et al. 2008). Selective 5-HT3 receptor antagonists are used to
treat emesis (Costall and Naylor 2004; Thompson and
Lummis 2006) and irritable bowel syndrome (De and Tonini
2001; Spiller 2004). Fasching et al. (2008) showed a
significant association between a coding polymorphism in
5-HT3C and response to anti-emetic 5-HT3 antagonists
5-HT3A 5-HT3A + 3C
10 µM 5-HT 10 µM 5-HT
2 s
100 pA
2 s
100 pA
5-HT3A
10 µM 5-HT
Control 30 µM
picrotoxin
Wash
100 pA
2 s
5-HT3A + 3C
Control 30 µM
picrotoxin
Wash
2 s
50 pA
*+
Vh mV
INormalised
5-HT3A
5-HT3A + 3C
–60 –40 –20 20 40 60
–1.0
–0.5
0.5
1.0
0
100
200
300
400
500
600
Peakcurrrent(pA)
All cells
Responding cells
5-HT3A
5-HT3A + 3C
n = 23 n = 20n = 27 n = 28
(a) (b)
(c) (d)
Fig. 6 Functional assessment of the potential for heteromerisation
between 5-HT3A and 5-HT3C. Whole-cell patch clamp recordings of
CHO cells transiently transfected with either 5-HT3A (1 : 3 ratio with
1 lg 5-HT3A cDNA + 3 lg empty pcDNA3.1 vector + 0.4 lg GFP) or
5-HT3A + 5-HT3C (1 : 3 ratio; 1 lg 5-HT3A + 3 lg 5-HT3C + 0.4 lg
GFP) were conducted to investigate the occurrence of potential sig-
natures of heteromeric channel assembly. (a,b) Currents recorded in
cells transfected with 5-HT3A + 5-HT3C were similar in appearance but
significantly smaller than currents recorded in cells transfected with 5-
HT3A alone [significance was apparent between a pairwise compari-
son of all cells (+, black bars) and responding cells only (*, grey bars)].
(c) Current-voltage relationships did not indicate any significant dif-
ference between responses detected in cells transfected with 5-HT3A
[Erev = )4.5 ± 0.6 mV (n = 9) I+60 mV/I)60 mV = 0.50 ± 0.04 (n = 9)] or
5-HT3A + 5-HT3C [Erev = )3.9 ± 1.8 mV (n = 7); I+60 mV/I)60 mV = 0.45
0.07 (n = 7)]. (d) Responses detected in cells transfected with 5-HT3A
and 5-HT3A + 5-HT3C showed similar sensitivity to 30 lM picrotoxin: %
Block was 52 ± 3 (n = 4) 58 ± 4 (n = 3), respectively.
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 391
9. during chemotherapy (Fasching et al. 2008). As the 5-HT3C,
5-HT3D and 5-HT3E genes are closely situated on chromo-
some 3 it is also possible that polymorphisms in 5-HT3D or
5-HT3E, contribute to the genetic association. It is of interest
that the expression of each of the 5-HT3 subunits is also high
in the DRG, providing a functional link between the gut and
the spinal cord. Potentially, therefore, the existence of novel
5-HT3 receptor subunits within these tissues offers the
possibility of identifying improved therapies for GI condi-
tions.
Using immunocytochemical approaches to study C-termi-
nally tagged subunits we demonstrated that each of the
5-HT3B-E variants can independently reach the plasma
membrane. This finding contrasts with reports from others
(Boyd et al. 2002, 2003; Niesler et al. 2007) using more
commonly applied N-terminal tagging approaches (Boyd
et al. 2002, 2003) that indicate that with the exception of
5HT3A, the other variants do not appear to reach the plasma
membrane unless co-expressed with 5-HT3A. Based on our
confocal microscopy data, functional validation of a 5HT3A-
lumio C-term tagged construct and the evaluated structural
data leading us to take this novel approach (Lummis et al.
2005; Unwin 2005), we therefore suggest that the C-terminus
is more accessible to antibodies at the cell surface and far
away from the pore region enabling efficient HA and FLAG
tag labelling with minimal effects on receptor function.
Although other possible explanations for these differing
results exist, such as endogenous expression of 5HT3A
receptors (Boyd et al. 2002, 2003; Sun et al. 2003; Quirk
et al. 2004; Niesler et al. 2007); N.B. we did not detect
functional expression of 5HT3A in WT cells), differential
expression of chaperones such as RIC-3 (Castillo et al. 2005;
Cheng et al. 2005, 2007), or components of receptor
glycosylation (Simon and Massoulie 1997; Quirk et al.
2004) or Protein Kinase C pathways (Sun et al. 2003) in the
difference host systems used, direct comparative studies to
(a) (b)
(c) (d)
0
100
200
300
400
500
600
5-HT3A + 3D
5-HT3A
+ CFP
Peakcurrent(pA)
5-HT3A
n = 20 n = 24 n = 20n = 18 n = 23 n = 18
10 µM 5-HT
5-HT3A 5-HT3A + 3D 5-HT3A + CFP
10 µM 5-HT10 µM 5-HT
200 pA
4 s
200 pA
4 s
5-HT3A + 3D
5-HT3A + CFP
5-HT3A
Control 30 µM picrotoxin Wash
Control 30 µM picrotoxin Wash
Control 30 µM picrotoxin Wash
–60 –40 –20 20 40 60
–1.0
–0.5
0.5
1.0
Vh mV
INormalised
5-HT3A
5-HT3A + 3D
5-HT3A
+ CFP
10 µM5-HT
All cells
Responding cells
Fig. 7 Functional assessment of the potential for heteromerisation
between 5-HT3A and 5-HT3D. Whole-cell patch clamp recordings of
HEK cells transiently transfected with either 5-HT3A or 5-HT3A + 5-
HT3D (1 : 3 ratio; 1 lg 5-HT3A cDNA and 3 lg 5-HT3C cDNA; GFP also
included) were conducted to investigate the occurrence of potential
signatures of heteromeric channel. In a third transfection group CFP
was included in place of 5-HT3D to act as a translational control (1 : 3
ratio; 1 lg 5-HT3 and 3 lg CFP; GFP also included). (a,b) Comparison
between peak amplitude currents from all cells (black bars) or
responding cells only (grey bars) were not significantly different be-
tween those cells transfected with 5-HT3A + 5-HT3D to those trans-
fected with 5-HT3A alone. There were no differences in current
kinetics. Inclusion of CFP in place of 5-HT3D also failed to alter current
kinetics or peak amplitude. (c) Current-voltage relationships did not
indicate any significant difference between responses detected in cells
from the three transfection groups: 5-HT3A [Erev = )2.8 ± 0.9 mV
(n = 5) I+60 mV/I)60 mV = )0.44 ± 0.04 (n = 5)] or 5-HT3A + 5-HT3D
[Erev = )4.6 ± 2.0 mV (n = 4) I+60 mV/I)60 mV = )0.50 ± 0.06 (n = 4)]
or 5-HT3D + CFP [Erev = )1.0 ± 1.9 mV (n = 5) I+60 mV/I)60 mV =
)0.50 ± 0.03 (n = 5)]. (d) Responses detected in cells from all three
groups showed similar sensitivity to picrotoxin: % block was 70 ± 4%,
70 ± 4% and 69 ± 5%, respectively (n = 5).
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
392 | J. D. Holbrook et al.
10. ascertain if N-terminal tagging approaches have a direct
effect on 5-HT3 receptor translocation are now warranted.
We used whole-cell patch clamp electrophysiology to
determine if 5-HT3C, 5-HT3D and 5-HT3E can form func-
tional homomers or distinct heteromeric receptors with
5-HT3A. Our deduced human 5-HT3A receptor properties
are consistent with previous work (Maricq et al. 1991;
Miyake et al. 1995; Reeves and Lummis 2002). The same
conditions did not provide evidence for functional expression
of 5-HT3B homomers, as expected, and additionally demon-
strate that 5-HT3C, 5-HT3D and 5-HT3E are incapable of
forming functional 5-HT-gated ion channels. These findings
are in agreement with (Niesler et al. 2007) who employed a
non-direct aequorin Ca2+
-based cellular assay to assess
functionality, and we suggest are most likely because of the
impact of differences in key amino acids in the TM2 channel
lining domains of 5-HT3A and the other subunits. For
example, the ‘polar ring’ (2¢ position in TM2 (Peters et al.
2005) is normally occupied by a serine or threonine residue
by members of the Cys-loop family and yet the novel
subunits bear a proline in this position, a non-conservative
change that may have a large impact on the structure of this
part of the channel, altering ionic conduction and/or selec-
tivity. Similarly, the -1¢ ‘intracellular ring’ normally bears an
acidic residue such as glutamate in 5-HT3A and the majority
of cationic Cys-loop receptors yet 5-HT3C, 5-HT3D and
5-HT3E all have asparagines at this position (Fig. S2).
Interestingly, however, the 5¢ lysine residue, a feature of the
5-HT3 TM2 domain implicated in desensitisation, is retained
in all of the subunits (Gunthorpe et al. 2000). Given that our
immunocytochemistry studies defined localization of the 5-
HT3C, 5-HT3D and 5-HT3E subunits at the plasma membrane,
the possibility that other (novel) agonists or co-agonists may
be required for activation requires further consideration.
Other possibilities are that the particular splice variants we
assayed are non-functional or accessory subunits may be
required for these novel subunits to attain functionality.
Studies to unambiguously define the occurrence of novel
heteromeric 5-HT3 receptors are not straightforward. The
required co-expression of the test subunit with 5-HT3A
(a) (b)
(c) (d)
0
400
800
1200
1600
5HT3A + 3E
Peakcurrent(pA)
5-HT3A
n = 26 n = 23 n = 21 n = 20
10 µM 5-HT 10 µM 5-HT
5-HT3A + 3E5-HT3A
200 pA
2 s
–60 –40 –20 20 40 60
–1.0
–0.5
0.5
1.0
INormalised
Vh mV
5-HT3A
5-HT3A + 3E
5-HT3A +3 E
Control 30 µM picrotoxin Wash
Control 30 µM picrotoxin Wash
5-HT3A
200 pA
2 s
10 µM 5-HT
All cells
Responding cells
Fig. 8 Functional assessment of the potential for heteromerisation
between 5-HT3A and 5-HT3E. Whole-cell patch clamp recordings of
CHO cells transiently transfected with either 5-HT3A or 5-HT3A + 5-
HT3E (1 : 3 ratio; 1 lg 5-HT3A + 3 lg 5-HT3E + 0.4 lg GFP) were
conducted to investigate the occurrence of potential signatures of
heteromeric channel assembly. (a,b) Currents recorded in cells
transfected with 5-HT3A + 5-HT3E showed a trend towards smaller
peak amplitudes compared to those recorded from cells transfected
with 5-HT3A alone. This difference was not significant following com-
parison between all cells (black bars) or responding cells only (grey
bars). There were no apparent differences in current kinetics. (c)
Current-voltage relationships did not indicate any significant difference
between responses detected in cells transfected with 5-HT3A
[Erev =)0.6 ± 0.6 mV (n = 12) I+60 mV/I)60 mV = )0.42 ± 0.04 (n = 12)]
or 5-HT3A + 5-HT3E [Erev = )1.7 ± 1.2 mV (n = 9) I+60 mV/I)60 mV =
)0.47 ± 0.08 (n = 9)]. (d) Responses detected in cells transfected
with 5-HT3A and 5-HT3A + 5-HT3E showed similar sensitivity to
picrotoxin: % block was 79 ± 5% and 74 ± 4%, respectively (n = 9).
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 393
11. means that there is a background level of homomeric 5-HT3A
receptor activity to deal with before the contributions of any
additional heteromeric receptors can be resolved. The design
of our comparative studies was therefore carefully controlled
with the same amount of 5-HT3A DNA used in single or
double transfections, such that the effects of addition of the
second subunit were tested upon a uniform background.
Parallel assessments of 5-HT3A were also made to control for
any variation in transfection efficiency or expression and the
studies employed high 5-HT concentrations to minimise
potential contributions of changes in agonist affinity to the
results obtained. Furthermore, our DNA (empty vector) and
‘translational’ (protein load) control experiments provided
confidence that any effects on 5-HT3A function were
specifically because of an interaction with 5-HT3A. Using
this approach we successfully validated the known hetero-
meric interaction between 5-HT3B and 5-HT3A (Davies et al.
1999; Das and Dillon 2003). Next we used this approach to
characterise the effects of 5-HT3C, 5-HT3D and 5-HT3E on 5-
HT3A receptor properties. Like 5-HT3B, the predicted
channel lining TM2 and ‘Membrane Associated’ (MA)
stretch (Kelley et al. 2003) regions of the novel subunits
differed markedly to those of 5-HT3A so we hypothesized
that incorporation of these subunits into 5-HT3 receptors
would yield channels with markedly altered properties.
However, with the exception of a significant decrease in
the mean current detectable for 5-HT3C, the profiles of 5-HT3
receptor responses were not altered by expression of the
novel subunits. The effect of 5-HT3C on the mean 5-HT
current suggests a specific effect on 5-HT3 receptor conduc-
tance or expression. In this respect, it is noteworthy that of
the novel subunits, 5-HT3C exhibits the greatest disruption of
the Membrane Associated stretch with a number of the key
arginine residues identified in 5-HT3A (Kelley et al. 2003)
not being represented in the 5-HT3C sequence (Fig. S2). Our
findings therefore provide a potential explanation for the
30% decrease in apparent maximum efficacy (Emax) of 5-
HT mediated Ca2+
responses in 5-HT3AC expressing cells
noted by (Niesler et al. 2007). These authors also describe
effects on Emax for the contribution of the 5-HT3E
(described in Fig. S2 as HTR3E_V2) (44% increase) and
5-HT3Ea (Fig. S2 as HTR3E_V1) (57% decrease) variants.
Our studies utilised a further 5-HT3E variant, V3, and hence
although they also indicate a potential interaction, the
direction and magnitude of effects differ and cannot be
reconciled based on the available information. Further
studies are needed to understand this apparently divergent
regulation by 5-HT3E variants.
Overall, the lack of any overt change in any measured
receptor properties beyond peak current response in the
heteromer experiments means that the present data do not
provide definitive evidence for the contribution of 5-HT3C and
5-HT3E as channel lining subunits. Niesler et al. (2007)
reported changes in Emax and Bmax as well as immunopre-
cipitation evidence that the novel subunits form novel
heteromeric receptors. Clearly, in the absence of the direct
evidence of changes in receptor properties sought here,
whether or not the novel subunits contribute to 5-HT3 receptor
diversity remains to be determined. The effects of novel
subunits on Emax, Bmax and peak current are of course
suggestive of an interaction between the proteins in the
recombinant host cell and further studies are now warranted to
examine these effects in greater detail and understand their
relevance to native tissue biology. We also note that the
potential for subunit heteromerisation and interaction need not
be bounded by receptor subfamilies (van Hooft et al. 1998),
hence many additional possibilities need to be addressed to
enable the full potential for the novel 5-HT3 subunits to
contribute to receptor diversity to be appreciated. Such
considerations need to pay close attention to the agonist
employed and the potential requirement for additional subunits
and/or accessory proteins to reconstitute or regulate function.
In conclusion, we have provided compelling data demon-
strating the conservation of 5-HT3C and 5-HT3E through
mammalian evolution. Indeed, the maintenance of their
mRNA expression in human GI tissue is suggestive of a role
in moderating the gut’s response to 5-HT. Our demonstration
that these subunits are non-functional when expressed alone
indicates that their primary role may be to modulate or regulate
the responses of other 5-HT or more distant Cys-loop receptor
relatives. Indeed, recent discoveries of single nucleotide
polymorphisms in HTR3B linked to major depressive disor-
der, in particular, highlight the key roles of such subunits as
modulators of receptor function and neurotransmitter signal-
ling (Krzywkowski et al. 2008). Further appreciation of the
contributions of these subunits to 5-HT receptor biology and
human physiology, may allow us to improve upon current
therapeutics targeting this receptor system.
Acknowledgement
We are grateful to David Murray and Daniel Hoston for construction
of the masterplates used for the TaqMan expression analysis.
Supporting information
Additional Supporting Information may be found in the online
version of this article:
Figure S1 Alignment of the novel HTR3 subunits found in
different species.
Figure S2 Multiple alignment of deduced peptide sequences of
human HTR3 subunit splice variants.
Figure S3 Additional TaqMan profiles of 5-HT splice variants.
Figure S4 RT-PCR and schematic of HTR3C cloning.
Figure S5 Expression analysis of the tagged 5-HT3 receptor
subunits 5-HT3B-HA, 5-HT3C-FLAG and 5-HT3E-FLAG in CHO
cells.
Table S1 Transcript-specific primers and probes used for
TaqMan analysis of 5-HT3 receptor subunits.
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
394 | J. D. Holbrook et al.
12. Please note: Wiley-Blackwell are not responsible for the content
or functionality of any supporting materials supplied by the authors.
Any queries (other than missing material) should be directed to the
corresponding author for the article.
References
Banner S. E., Carter M. and Sanger G. J. (1995) 5-Hydroxytryptamine3
receptor antagonism modulates a noxious visceral pseudoaffective
reflex. Neuropharmacology 34, 263–267.
Belelli D., Balcarek J. M., Hope A. G., Peters J. A., Lambert J. J. and
Blackburn T. P. (1995) Cloning and functional expression of a
human 5-hydroxytryptamine type 3AS receptor subunit. Mol.
Pharmacol. 48, 1054–1062.
Birney E., Clamp M. and Durbin R. (2004) GeneWise and Genomewise.
Genome Res. 14, 988–995.
Boyd G. W., Low P., Dunlop J. I., Robertson L. A., Vardy A., Lambert
J. J., Peters J. A. and Connolly C. N. (2002) Assembly and cell
surface expression of homomeric and heteromeric 5-HT3 recep-
tors: the role of oligomerization and chaperone proteins. Mol. Cell.
Neurosci. 21, 38–50.
Boyd G. W., Doward A. I., Kirkness E. F., Millar N. S. and Connolly
C. N. (2003) Cell surface expression of 5-hydroxytryptamine type
3 receptors is controlled by an endoplasmic reticulum retention
signal. J. Biol. Chem. 278, 27681–27687.
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 5-HT3A receptors versus heteromeric 5-
HT3A/3B receptors. Neuropharmacology 41, 282–284.
Castillo M., Mulet J., Gutierrez L. M., Ortiz J. A., Castelan F., Gerber S.,
Sala S., Sala F. and Criado M. (2005) Dual role of the RIC-3
protein in trafficking of serotonin and nicotinic acetylcholine
receptors. J. Biol. Chem. 280, 27062–27068.
Cheng A., McDonald N. A. and Connolly C. N. (2005) Cell surface
expression of 5-hydroxytryptamine type 3 receptors is promoted by
RIC-3. J. Biol. Chem. 280, 22502–22507.
Cheng A., Bollan K. A., Greenwood S. M., Irving A. J. and Connolly
C. N. (2007) Differential subcellular localization of RIC-3 isoforms
and their role in determining 5-HT3 receptor composition. J. Biol.
Chem. 282, 26158–26166.
Costall B. and Naylor R. J. (2004) 5-HT3 receptors. Curr. Drug Targets
CNS Neurol. Disord. 3, 27–37.
Das P. and Dillon G. H. (2003) The 5-HT3B subunit confers reduced
sensitivity to picrotoxin when co-expressed with the 5-HT3A
receptor. Brain Res. Mol. Brain Res. 119, 207–212.
Das P. and Dillon G. H. (2005) Molecular determinants of picrotoxin
inhibition of 5-hydroxytryptamine type 3 receptors. J. Pharmacol.
Exp. Ther. 314, 320–328.
Davies P. A., Pistis M., Hanna M. C., Peters J. A., Lambert J. J.,
Hales T. G. and Kirkness E. F. (1999) The 5-HT3B subunit is a
major determinant of serotonin-receptor function. Nature 397,
359–363.
De P. F. and Tonini M. (2001) Irritable bowel syndrome: new agents
targeting serotonin receptor subtypes. Drugs 61, 317–332.
Downie D. L., Hope A. G., Lambert J. J., Peters J. A., Blackburn T. P. and
Jones B. J. (1994) Pharmacological characterization of the apparent
splice variants of the murine 5-HT3 R-A subunit expressed in
Xenopus laevis oocytes. Neuropharmacology 33, 473–482.
Ebino K. Y., Shutoh Y. and Takahashi K. W. (1993) Coprophagy
in rabbits: autoingestion of hard feces. Jikken Dobutsu 42, 611–
613.
Fasching P. A., Kollmannsberger B., Strissel P. L. et al. (2008) Poly-
morphisms in the novel serotonin receptor subunit gene HTR3C
show different risks for acute chemotherapy-induced vomiting after
anthracycline chemotherapy. J. Cancer Res. Clin. Oncol. 134,
1079–1086.
Gunthorpe M. J., Peters J. A., Gill C. H., Lambert J. J. and Lummis S. C.
(2000) The 4¢lysine in the putative channel lining domain affects
desensitization but not the single-channel conductance of re-
combinant homomeric 5-HT3A receptors. J. Physiol. 522 Pt 2,
187–198.
Holland P. M., Abramson R. D., Watson R. and Gelfand D. H. (1991)
Detection of specific polymerase chain reaction product by
utilizing the 5¢–3¢ exonuclease activity of Thermus aquaticus
DNA polymerase. Proc. Natl Acad. Sci. USA 88, 7276–
7280.
van Hooft J. A., Spier A. D., Yakel J. L., Lummis S. C. and Vijverberg
H. P. (1998) Promiscuous coassembly of serotonin 5-HT3 and
nicotinic alpha4 receptor subunits into Ca(2+)-permeable ion
channels. Proc. Natl Acad. Sci. USA 95, 11456–11461.
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 5-HT3 receptor A subunit.
Eur. J. Pharmacol. 245, 187–192.
Karnovsky A. M., Gotow L. F., McKinley D. D. et al. (2003) A cluster
of novel serotonin receptor 3-like genes on human chromosome 3.
Gene 319, 137–148.
Kelley S. P., Dunlop J. I., Kirkness E. F., Lambert J. J. and Peters J. A.
(2003) A cytoplasmic region determines single-channel conduc-
tance in 5-HT3 receptors. Nature 424, 321–324.
Krzywkowski K., Davies P. A., Feinberg-Zadek P. L., Brauner-Osborne
H. and Jensen A. A. (2008) High-frequency HTR3B variant
associated with major depression dramatically augments the sig-
naling of the human 5-HT3AB receptor. Proc. Natl Acad. Sci. USA
105, 722–727.
van Lelyveld N., Linde J. T., Schipper M. and Samsom M. (2008)
Candidate genotypes associated with functional dyspepsia.
Neurogastroenterol. Motil. 20, 767–773.
Lummis S. C., Beene D. L., Lee L. W., Lester H. A., Broadhurst R. W.
and Dougherty D. A. (2005) Cis-trans isomerization at a proline
opens the pore of a neurotransmitter-gated ion channel. Nature
438, 248–252.
MalikN.M., Moore G.B., SmithG.,Liu Y. L.,Sanger G.J. andAndrewsP.
L. (2006) Behavioural and hypothalamic molecular effects of the
anti-cancer agent cisplatin in the rat: a model of chemotherapy-re-
lated malaise? Pharmacol. Biochem. Behav. 83, 9–20.
Maricq A. V., Peterson A. S., Brake A. J., Myers R. M. and Julius D.
(1991) Primary structure and functional expression of the 5HT3
receptor, a serotonin-gated ion channel. Science 254, 432–437.
Miyake A., Mochizuki S., Takemoto Y. and Akuzawa S. (1995)
Molecular cloning of human 5-hydroxytryptamine3 receptor:
heterogeneity in distribution and function among species. Mol.
Pharmacol. 48, 407–416.
Niesler B., Frank B., Kapeller J. and Rappold G. A. (2003) Cloning,
physical mapping and expression analysis of the human 5-HT3
serotonin receptor-like genes HTR3C, HTR3D and HTR3E. Gene
310, 101–111.
Niesler B., Walstab J., Combrink S. et al. (2007) Characterization of the
novel human serotonin receptor subunits 5-HT3C,5-HT3D, and
5-HT3E. Mol. Pharmacol. 72, 8–17.
Peters J. A., Hales T. G. and Lambert J. J. (2005) Molecular determinants
of single-channel conductance and ion selectivity in the Cys-loop
family: insights from the 5-HT3 receptor. Trends Pharmacol. Sci.
26, 587–594.
Quirk P. L., Rao S., Roth B. L. and Siegel R. E. (2004) Three putative
N-glycosylation sites within the murine 5-HT3A receptor sequence
affect plasma membrane targeting, ligand binding, and calcium
Ó 2008 The Authors
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Characterisation of novel 5-HT3 receptor subunits | 395
13. influx in heterologous mammalian cells. J. Neurosci. Res. 77, 498–
506.
Reeves D. C. and Lummis S. C. (2002) The molecular basis of the
structure and function of the 5-HT3 receptor: a model ligand-gated
ion channel (review). Mol. Membr. Biol. 19, 11–26.
Sanger G. J. (1995) Preclinical differences in 5-HT3 receptor antagonist
characteristics, in Serotonin and the Scientific basis of Anti-emetic
Therapy (Reynolds D. J. M., Andrews P. L. R. and Davis C. J.,
eds), pp. 155–163. Oxford Clinical Communications, Oxford.
Simon S. and Massoulie J. (1997) Cloning and expression of acetyl-
cholinesterase from Electrophorus. Splicing pattern of the 3¢ exons
in vivo and in transfected mammalian cells. J. Biol. Chem. 272,
33045–33055.
Spiller R. C. (2004) Irritable bowel syndrome. Br. Med. Bull. 72, 15–29.
Sun H., Hu X. Q., Moradel E. M., Weight F. F. and Zhang L. (2003)
Modulation of 5-HT3 receptor-mediated response and trafficking by
activation of protein kinase C. J. Biol. Chem. 278, 34150–34157.
Thompson A. J. and Lummis S. C. (2006) 5-HT3 receptors. Curr.
Pharm. Des. 12, 3615–3630.
Tzvetkov M. V., Meineke C., Oetjen E., Hirsch-Ernst K. and Brockm-
oller J. (2007) Tissue-specific alternative promoters of the seroto-
nin receptor gene HTR3B in human brain and intestine. Gene 386,
52–62.
Unwin N. (2005) Refined structure of the nicotinic acetylcholine
receptor at 4A resolution. J. Mol. Biol. 346, 967–989.
Wilgenbusch J. C. and Swofford D. (2003) Inferring evolutionary
trees with PAUP*. Curr. Protoc. Bioinformatics Chapter 6,
Unit.
Journal Compilation Ó 2008 International Society for Neurochemistry, J. Neurochem. (2009) 108, 384–396
Ó 2008 The Authors
396 | J. D. Holbrook et al.