(Received for publication, November 4, 1996, and in revised form, December 27, 1996)
From the Department of Biochemistry and Molecular
Biology, Research Institute Neurosciences, Vrije Universiteit, De
Boelelaan 1083, 1081 HV Amsterdam, The Netherlands and the
§ Department of Molecular and Cellular Neurobiology,
Membrane Physiology Section, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands
A cDNA encoding a G-protein-coupled receptor
was cloned from the central nervous system of the pond snail
Lymnaea stagnalis. The predicted amino acid sequence of
this cDNA most closely resembles the Drosophila
tyramine/octopamine receptor, the Locusta tyramine receptor, and an octopamine receptor (Lym oa1) that we
recently cloned from Lymnaea. After stable expression of
the cDNA in HEK293 cells, we found that
[3H]rauwolscine binds with high affinity to the receptor
(KD = 6.2·109 M).
Octopamine appears to be the most potent naturally occurring agonist to
displace the [3H]rauwolscine binding
(Ki = 3.0·10
7 M).
Therefore, the receptor is considered to be an octopamine receptor and
is consequently designated Lym oa2. The novel receptor shares little pharmacological resemblance with Lym oa1,
indicating that the two receptors represent different octopamine
receptor subfamilies. Octopaminergic stimulation of Lym oa2
does not induce changes in intracellular concentrations of cAMP or
inositol phosphates. However, electrophysiological experiments indicate
that octopamine is able to activate a voltage-independent
Cl
current in HEK293 cells stably expressing Lym
oa2. Although opening of this chloride channel most
probably does not require the activation of either protein kinase A or
C, it can be blocked by inhibition of protein phosphorylation.
G-protein1-coupled receptors form a large superfamily of membrane receptors that can be found in all species ranging from unicellular eukaryotes to mammals and that can interact with a large variety of signals (e.g. light, odorants, Ca2+, biogenic amines, glycoprotein hormones, etc.) (1).
Bioamines like dopamine, epinephrine, norepinephrine, and octopamine all interact with specific G-protein-coupled receptors. Dopamine is present in high amounts in both vertebrate and invertebrate species. Epinephrine and norepinephrine, on the other hand, are present predominantly in vertebrates, whereas octopamine is considered to act as a major neurotransmitter in invertebrate species only. Octopamine is often referred to as the invertebrate counterpart of norepinephrine because of the structural similarity between these neurotransmitters (they differ only in the presence of a single catecholic hydroxyl group) as well as their functional similarity (both serve an important role in stress adaptation) (2). Consequently, the adrenergic and octopaminergic receptors share pharmacological as well as structural properties (3, 4).
The role of octopamine as a neurotransmitter has been studied particularly well in insects, where it has been shown to interact with at least four different octopamine receptor subtypes (3). The central nervous systems of a number of snails have also been used to study octopaminergic neurotransmission. The presence of octopamine has been demonstrated in the sea slug Aplysia (5-7), the land snail Helix (8, 9), and the pond snail Lymnaea (10). In these species, the interactions of octopamine with its receptors have been described mainly at the electrophysiological level (7, 8, 11-15). In general, application of octopamine induces a hyperpolarization of susceptible snail neurons, a process that is thought to be mediated by an increase in cAMP (2, 11, 12), leading to an increased potassium (12-14) or calcium (15) conductance. In addition, a pharmacological characterization of octopamine receptors has been described for Lymnaea.2
We have recently cloned and characterized a cDNA encoding an
octopamine receptor expressed in the brain of Lymnaea
stagnalis (16). This receptor (Lym oa1) shows moderate
homology to the Drosophila tyramine/octopamine receptor (17,
18), to the Locusta tyramine receptor (19) and to the
vertebrate -adrenergic receptors. Activation of this receptor, when
expressed in human embryonic kidney (HEK293) cells, leads to elevated
concentrations of both intracellular inositol phosphates and cAMP. The
pharmacological profile of Lym oa1 suggests that this
receptor represents a previously unknown octopamine receptor
subtype.
This paper describes the structure of a second octopamine receptor
cDNA (Lym oa2) cloned from Lymnaea. The
predicted amino acid sequence of this receptor only shows limited
similarity to Lym oa1, as well as to the insect tyramine
receptors and the -adrenergic receptors. Its pharmacological profile
clearly differs from that of previously described receptors. When the
novel receptor is stably expressed in HEK293 cells, application of
octopamine does not lead to changes in the intracellular concentration
of cAMP or inositol phosphates. Activation of Lym oa2 does,
however, induce a long lasting opening of Cl
channels.
G-protein-mediated activation of a similar Cl
current was
recently described by Postma et al. (20) in rat fibroblasts.
Their study also indicated that none of the known second messenger
system was involved in the Cl
channel activation.
Nevertheless, it was suggested that the increase in chloride
conductance was closely associated with the activation of
phosphoinositide hydrolysis. In contrast, our results clearly indicate
that this novel signaling pathway functions independently from the
activation of phospholipase C (PLC). We further show that although
protein kinase A and protein kinase C are most probably not involved in
the signaling pathway, protein phosphorylation is important for the
opening of the Cl
channel.
Central nervous systems of adult L. stagnalis bred in the laboratory (21) were dissected, the total
RNA was isolated (22), and cDNA was transcribed using oligo(dT)
primers and SuperScriptTM RNase H reverse
transcriptase (Life Technologies, Inc.). This cDNA served as a
template in a degenerate PCR strategy designed to isolate genes
encoding G-protein-coupled bioamine receptors. The oligonucleotide primers recognize stretches of conserved amino acid residues present in
transmembrane (TM) regions 6 and 7 and have been described previously
(23). PCR products were cloned and sequenced (Sequenase, Life
Technologies, Inc.), and one of the fragments revealed significant similarity to the TM6-TM7 region of adrenergic receptors. The presence
of several large cDNA clones corresponding to this initial fragment
was confirmed in a PCR-based screening of a fractionated Lymnaea central nervous system cDNA library in
-ZAP.
Isolated single plaques containing the full-length cDNA inserts
were converted into pBS-SK
phagemids (designated pBS-Lym
oa2) by in vivo excision, and the insert was
sequenced (Sequenase, Life Technologies, Inc.).
A
PCR fragment covering the 5 part of the open reading frame of pBS-Lym
oa2 was generated using a sense oligonucleotide based on
the DNA sequence around the start codon of the open reading frame and
an antisense oligonucleotide based on the sequence 3
from the
endogenous KpnI site (located at position 563; see Fig. 1).
This fragment was cloned, sequenced on both strands, and combined with
the 3
-coding region of Lym oa2 obtained as a
KpnI and XbaI fragment (XbaI site
located at position 1931; see Fig. 1). The total coding region was
cloned into pcDNA3 (Invitrogen), yielding a construct designated
pcDNA-Lym oa2.
HEK293 cells were stably transfected with pcDNA-Lym oa2 as described before (16). The level of expression of Lym oa2 was determined by measuring the binding of [3H]rauwolscine (81-85 Ci/mmol; Amersham Corp.) to membrane preparations of resistant colonies. One cell line exhibiting a Bmax of 2.2 pmol/mg was selected for further study.
Radioligand Binding AssaysThe preparation of membranes of HEK293 cells and the radioligand binding experiments were carried out as described previously (16).
Saturation isotherms were obtained by incubating membrane protein with increasing amounts of [3H]rauwolscine (0.1-45 nM). Nonspecific binding was determined by the addition of mianserin to a final concentration of 5 µM.
Competition curves were obtained by incubating membrane protein with 6 nM [3H]rauwolscine and increasing amounts of
competitor (1010-10
4 M). The
obtained data were fitted using Kaleidagraph 3.0 (Abelbeck Software) as
described previously (16).
HEK293 cells were grown to ± 50% confluency in 24-well plates, and incubated with 1 µCi of myo-[3H]inositol (18 Ci/mmol, Amersham Corp.) per ml of inositol-free Dulbecco's modified Eagle's medium (Life Technologies, Inc.) for 20-24 h. Cells were incubated with agonists in the presence of LiCl (10 mM) for 60 min at 37 °C. Cells were lysed with chloroform:methanol, and total cellular inositol phosphates were extracted using Dowex AG 1-X8 anion exchange resin as described previously (16).
Measurements of Cyclic AMP FormationHEK293 cells were grown to ±80% confluency in 24-well plates and incubated with agonists and 300 µM 3-isobutyl-1-methylxanthine for 20 min at 37 °C. The medium was aspirated and the cells were lysed by sonication in 200 µl of ice-cold 0.1 N HCl. Then, 75 µl of neutralization buffer (230 mM NaOH, 560 mM Tris, 140 mM NaCl, 56 mM EDTA, 0.35 M HEPES) was added, and the concentration of cAMP was determined as described previously (25).
Electrophysiological ExperimentsFor electrophysiological
recordings, HEK293 cells were kept in Petri dishes (Costar) and bathed
in 140 mM NaCl, 2.8 mM KCl, 1 mM
CaCl2, 2 mM MgCl2, 10 mM HEPES, 10 mM glucose, pH 7.2, adjusted with
NaOH. The composition of standard (non-selective) pipette solution was
135 mM KCl, 1 mM CaCl2, 10 mM EGTA, 10 mM HEPES, 2 mM MgATP,
0.1 mM Tris-GTP, pH 7.2, adjusted with KOH. To test whether
currents were carried by Cl ions, KCl in the pipette
solution was replaced by K+-aspartate. Agonists were
administered by means of a pico-spritzer (General Valve, Fairfield,
U. S. A.) using a small glass pipette (tip diameter 20 µm) placed
at a distance of ~100 µm from the recorded cell. This setup enabled
rapid application of drugs (
1 s).
Whole cell voltage clamp experiments were performed using either an
Axoclamp 2A amplifier (Axon Instruments, Inc., Foster City, CA) in the
continuous single electrode voltage clamp mode, in which case cell
capacitance (~60 picofarads) was not compensated, or a Liszt EPC7
amplifier (Liszt, Darmstadt, Germany), allowing cell capacitance
compensation. Pipettes (2-6 M) were pulled on a Flaming/Brown P-87
(Sutter Instrument Co.) horizontal micro-electrode puller from Clark
GC-150 glass (Clark Electromedical Instruments, UK; seal resistance,
>2 G
). After disruption of the patch membrane, series resistance
(<6 M
) was compensated for ~70%. Measurements commenced 5 min
after access to the cell to allow equilibration with the pipette
solution. Data acquisition was controlled by a CED 1400 AD/DA converter
(Cambridge Electronics Design, Cambridge, UK) connected to an Intel
80486-based computer run with voltage clamp software developed in our
laboratory. The current recordings were filtered at 1-5 kHz, sampled
at 1 kHz, and stored on line. This system allowed simultaneous
application of voltage steps, acquisition of current recordings, and
timed application of drugs.
We have applied the
PCR technique to clone G-protein-coupled receptors that are expressed
in the brain of L. stagnalis, using degenerated primers
based on highly conserved sequences present in TM6 and TM7 of
G-protein-coupled receptors. The obtained PCR products were sequenced
and these sequences were conceptually translated. One derived amino
acid sequence showed considerable similarity to the sequence of the
corresponding region in catecholaminergic receptors. We isolated and
sequenced the corresponding full-length cDNA from a
Lymnaea central nervous system library (see Fig.
1). An open reading frame that can encode a protein of
578 amino acids is present on this cDNA (see Fig. 1). The presence
of seven hydrophobic regions characteristic for G-protein-coupled
receptors can clearly be recognized within the predicted amino acid
sequence. Within these regions, the most prominent amino acid identity
is found with the Drosophila tyramine/octopamine receptor
(49%) (17, 18), the Locusta tyramine receptor (50%) (19),
Lym oa1 (45%), and the mammalian
2-adrenoreceptors (±41%), thus confirming that the
open reading frame is likely to code for a bioamine receptor.
To further delineate the nature of this receptor
we stably expressed it in HEK293 cells. We then used the binding of
[3H]rauwolscine to membranes of these cells as a marker
to select clonal lines expressing high levels of receptor protein. One
clone, exhibiting a Bmax of 2.2 pmol/mg, was
chosen for a pharmacological characterization of the receptor. Fig.
2 shows the saturation binding curve and the
corresponding Scatchard plot of the binding of
[3H]rauwolscine. The affinity constant
(KD) of [3H]rauwolscine for the novel
receptor is 6.24 nM. The ability of several (mainly
adrenergic) compounds to displace [3H]rauwolscine binding
is presented in Table I. The rank order of potencies of
selected agonists is: (±)-p-synephrine > (±)-p-octopamine > xylometazoline > ()-norepinephrine = clonidine > epinephrine
p-tyramine > dopamine. The rank order of potencies of
selected antagonists is: rauwolscine = mianserin > phentolamine > spiperone > yohimbine > (
)-propanolol > prazosine > pindolol. Since octopamine is
the most potent naturally occurring agonist to displace the [3H]rauwolscine binding, the receptor was considered to
be an octopamine receptor and consequently called Lym
oa2.
|
We recently cloned and characterized another octopamine receptor from
Lymnaea (Lym oa1; Ref. 16). To allow for a
comparison of the binding properties of both octopamine receptors,
Table I also shows the pKi values of the same set of
ligands obtained by the displacement of [3H]rauwolscine
binding from Lym oa1. As can be seen from a Pearson correlation graph (Fig. 3) the pharmacological profiles
of Lym oa1 and Lym oa2 are considerably
different.
Signal Transduction of Lym oa2
We tested to
discover to which signal transduction pathways Lym oa2 can
be coupled. In transiently transfected HEK293 cells, as well as in four
independently isolated stable cell lines, octopaminergic stimulation
did not induce any change in the concentration of cAMP or inositol
phosphates. Also, stimulation with tyramine, epinephrine,
norepinephrine, dopamine, serotonin, or histamine did not change the
concentration of these second messengers as compared with
non-transfected HEK293 cells. Positive control experiments, however,
resulted in pronounced increases in cAMP or inositol phosphates
(stimulation of endogenous -adrenergic receptors increased cAMP
20-fold over basal levels, and stimulation of Lym oa1
receptors expressed in HEK293 cells increased inositol phosphates
45-fold over basal levels; data not shown).
Since activation of many neuronal receptors induces changes in membrane
conductances of the cells in which they are expressed, we tested
whether application of octopamine to HEK293 cells expressing Lym
oa2 elicited electrical responses in these cells.
Application of 10 µM octopamine to HEK293 cells
expressing Lym oa2 induced a large but slow increase in the
holding current of the voltage-clamped cells (Fig.
4A). The amplitude of the octopamine-induced
current decreased at more depolarized potentials. Application of
synephrine (10 µM), an agonist with a higher potency than
octopamine (see Table I), mimicked the octopamine-induced response,
whereas yohimbine (10 µM), an antagonist, inhibited the
response, thus confirming the specificity of the effect (see Fig. 4,
B and C). Application of octopamine to
non-transfected HEK293 cells did not show any effect (Fig.
7A).
In order to study the nature of the current (e.g. the
voltage dependence and ionic selectivity), we performed voltage ramp experiments. Current responses were recorded while the voltage was
continuously varied from 80 mV to +10 mV over a period of 8 s,
both in the absence and presence of octopamine (Fig.
5A). The current voltage (IV) relation of the
octopamine-induced current was obtained by subtracting the control
current from the current response in the presence of octopamine. Fig.
5B shows that octopamine activates a current response over
the whole voltage range tested and that the IV relation of the
octopamine-induced current is almost linear, indicating that the
current is voltage-independent. Comparison of the IV relations of
nonstimulated and stimulated cells revealed no or only minor
differences (<5 mV) in reversal potential (Fig. 5A). The IV
curve of the octopamine-induced current (isolated by subtraction)
showed reversal around
10 mV, which is close to the reversal
potential of chloride ions (ECl
=
2 mV under
the present ionic conditions). To test whether the octopamine response
is indeed carried by Cl
ions, we replaced KCl in the
intracellular (pipette) medium with K+-aspartate, thus
shifting ECl
to
108 mV. This change caused a dramatic
loss of the octopamine-induced inward current response (Fig. 5,
C and D). While under standard Cl
conditions (measured at
90 mV) octopamine induced an increase in
inward current of 149 ± 33 pA (n = 6), this
reversed to a negligible outward current of 13 ± 7 pA
(n = 5) with aspartate replacing Cl
in
the pipette. This result strongly suggests that octopamine activates
chloride channels.
We then performed a limited number of initial experiments to study the
signal transduction pathway underlying the activation of the
Cl channel. The above experiments already indicated that
the process is slow; the current still increased 10 min after
application of octopamine. To see whether a phosphorylation step is
involved, we tested the effect of the nonspecific protein kinase
inhibitor H1004 on the octopamine-induced current. Fig.
6 shows that in the presence of 50 µM
H1004, octopamine did not induce or only very slightly induced the
inward Cl
current, whereas subsequent application of only
octopamine did evoke the normal response.
Interestingly, we observed that HEK293 cells stably expressing Lym
oa1 showed the same current response to octopamine as cells expressing Lym oa2 (Fig. 7B). The
effect, however, is not a general consequence of the activation of
overexpressed, heterologous receptors, since stimulation of a
Lymnaea serotonin receptor (5-HT2Lym) stably expressed in HEK293 cells (23) did not show any effect on the Cl conductance (Fig. 7C). Furthermore,
stimulation of
-adrenergic receptors that are endogenously present
in HEK293 cells did not influence the inward current (Fig.
7D). Stimulation of Lym oa1, Lym
oa2, 5-HT2Lym, and
-AR all differentially
influenced the outward current. These effects, however, were not
studied in detail.
We have used a degenerate PCR strategy to isolate cDNAs
encoding G-protein-coupled bioamine receptors that are expressed in the
central nervous system of the pond snail L. stagnalis.
Recently, we reported the cloning and expression of a serotonin
receptor (5-HT2Lym; Ref. 23) and an octopamine receptor
(Lym oa1; Ref. 16) using the same strategy. This paper
describes the isolation of a cDNA encoding a second octopamine
receptor, designated Lym oa2. The predicted amino acid
sequence of Lym oa2 exhibits the highest similarity to the
Drosophila tyramine/octopamine receptor, the
Locusta tyramine receptor, and Lym oa1, and to
the vertebrate -adrenergic receptors. Although both Lym
oa2 and Lym oa1 encode Lymnaea
octopamine receptors, their amino acid identity is only moderate,
i.e. 45% in the TM regions. An interesting difference in
the sequence of both octopamine receptors can be found in TM5. Mutagenesis studies on catecholaminergic receptors have indicated that
two conserved serine residues in this domain (Ser204 and
Ser207 in the
-adrenergic receptor) play a crucial role
in ligand binding. Supposedly, the two Ser hydroxyl groups can hydrogen
bond to the catechol hydroxyl groups of the ligand (see for instance
Refs. 26-28). In octopamine receptors there is no obvious need for
conservation of both serine residues in TM5 because the aromatic ring
of octopamine is monohydroxylated. Indeed, in Lym oa1 only
a single serine is found at the relevant position in TM5. In Lym
oa2, however, both serines are present. Interestingly, the
catecholamines (epinephrine, norepinephrine, dopamine) show a
considerably higher affinity for Lym oa2 than for Lym
oa1 (see Table I). This suggests that the serine residues
in TM5 play a role in agonist binding in the Lymnaea
octopamine receptors similar to their role in the vertebrate catecholamine receptors.
In general, agonists exhibit higher affinities for Lym oa2
than for Lym oa1, while antagonists have higher affinities
for Lym oa1 than for Lym oa2. Another
interesting difference between the binding properties of both
octopamine receptors is the opposite order of affinities for the
isoschizomers rauwolscine and yohimbine. Lym oa1 has a
higher affinity for yohimbine than for rauwolscine (pKi = 8.9 versus 7.5), while Lym
oa2 has a higher affinity for rauwolscine than for
yohimbine (pKi = 8.0 versus 6.2). Both
the pharmacological profiles of Lym oa1 and Lym
oa2 indicate a closer relationship to the -adrenergic
receptors than to the
-adrenergic receptors. In that respect it is
noteworthy to mention the much higher affinity of the
1
antagonist prazosine for Lym oa1 (pKi = 7.0) than for Lym oa2 (pKi = 4.7).
Whereas the pharmacological profile of Lym oa1 still shows a moderate similarity to that of the Drosophila
tyramine/octopamine receptor, the Locusta tyramine receptor,
and the
2-adrenergic receptors, the pharmacological
profile of Lym oa2 clearly differs from that of Lym
oa1, the tyramine receptors, the adrenergic receptors, and
the insect octopamine receptor subtypes as described in tissue preparations.
Stimulation of Lym oa2 did not lead to the activation of
the classical signal transduction pathways mediated by adenylyl cyclase and PLC. The receptor is, however, clearly able to transduce signals since we found that HEK293 cells expressing Lym oa2 showed
octopamine-induced changes in their membrane conductance. More
specifically, we observed a slow but large increase in the holding
current of voltage-clamped HEK293 cells upon application of octopamine.
At potentials below the Cl equilibrium potential, this
response was observed as an increase in inward current, most likely
caused by an efflux of Cl
ions. The process underlying
the opening of the Cl
channels may involve protein
phosphorylation, since the (nonselective) protein kinase inhibitor
HA1004 inhibited the inward current. Alternatively, the
Cl
channels involved may need to be in a phosphorylated
state to be able to open in response to the octopamine stimulus.
Because the IV relationship of the current response proved to be linear and because no changes in inositol phosphates were observed upon octopamine application, the Cl
channels involved are
suggested to be both voltage-independent and
Ca2+-independent. Additional effects of octopamine on
(voltage-dependent) outward currents were also observed but
were not pursued in detail.
Interestingly, stimulation of HEK293 cells expressing the other
Lymnaea octopamine receptor, Lym oa1, resulted
in a similar outward Cl current. In contrast to Lym
oa2, Lym oa1 has previously been shown to
activate both PLC and adenylyl cyclase. To examine whether the effect
on the Cl
channel might be secondary to the effects on
adenylyl cylase and PLC or might represent an independent signaling
route, we tested the effect of activation of two other receptors,
expressed in HEK293 cells. Stimulation of these receptors,
i.e. a Lymnaea serotonin receptor or an
endogenous
-adrenergic receptor, led to activation of PLC and
adenylyl cyclase, respectively. The subsequent rise in levels of
diacylglycerol or cAMP will activate protein kinase C or protein kinase
A, respectively. Stimulation of these receptors did, however, not show
any effect on the inward current, suggesting that protein kinase C or
protein kinase A is not involved in the phosphorylation process
underlying Cl
channel activation.
It remains to be investigated whether stimulation of Lym
oa2 in Lymnaea neurons will also affect the
Cl conductance. Activation of octopamine receptors on
neurons of other snails has been described to lead to outward potassium
currents (12-14) and inward calcium currents (15). Early studies on
Aplysia octopamine receptors (29) have shown that
stimulation of neurons with octopamine induces the specific
phosphorylation of a particular unidentified protein. In this case,
however, it was suggested that the process might involve protein kinase
A, since elevating the intracellular concentration of cAMP produced a
similar effect.
Recently, it was shown that activation of the G-protein-coupled
lysophosphatidic acid receptor present on fibroblasts leads to a long
lasting depolarization of these cells due to an efflux of chloride ions
(20). As was found in our studies, the signal transduction pathway
responsible for opening the chloride channel was suggested to be
independent of known second messengers. The opening of this particular
chloride channel could also be induced by activation of the thrombin
receptor, the endothelin receptor, and the neurokinin A receptor. Since
these receptors (as well as the lysophosphatidic acid receptor) all
couple to PLC, it was suggested that the signaling pathway leading to
the increased Cl conductance is closely associated with
phosphoinositide hydrolysis (20).
All available data suggest that the chloride channel in HEK293 cells
that we have found to be G-protein-activated is highly similar to the
channel described by Postma et al. (20). Our data, however,
exclude the option that phosphoinositide hydrolysis is important in the
signaling pathway leading from the activated receptor to the opening of
the chloride channel, since stimulation of Lym oa2 does not
result in any change in intracellular concentrations of inositol
phosphates. Also, activation of a serotonin receptor that has been
shown to be coupled to the activation of PLC does not influence the
outward Cl current. The G-protein-mediated activation of
the chloride channel described by Postma et al. in Rat-1
fibroblasts (20) and by ourselves in HEK293 cells must proceed via an
as yet unknown signaling pathway. Here we show that protein
phosphorylation is important in this pathway but that protein kinase A
and protein kinase C are not involved. The identification of the
relevant kinase and the further examination of the signal transduction
pathway will be the topics of future investigations.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U62770[GenBank].