| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, November 23, 1994) From the
A recombinant baculovirus containing a mouse
5hydroxytryptamine
The elucidation of the three-dimensional structure of
transmitter-gated ion channels, a superfamily of receptors that
includes the nicotinic acetylcholine (nACh), ( The Torpedo nACh receptor three-dimensional
structure is known to a resolution of 9 Å (6) , but it is
likely that a structure at atomic resolution will have to await
three-dimensional crystals suitable for x-ray diffraction. Despite the
relative ease of purifying large amounts of the Torpedo receptor, no such crystals have been obtained. This could simply
be due to a structural property of this particular receptor (such
species-specific effects on crystallization are often seen with soluble
proteins). Alternatively, the pseudo rather than true symmetry of a
receptor with four different subunit types per molecule may make
crystallization in the third dimension difficult. A channel composed of
a single subunit type rather than several might therefore have a better
chance of crystallizing, in addition to making overexpression more
straightforward. Several of the cloned subunits from the
neurotransmitter-gated ion channel family form functional channels when
expressed alone in heterologous
systems(7, 8, 9, 10) . There is as
yet no evidence that any of these homo-oligomers are present in
vivo, and for certain subunits this is clearly not the case (for
example, the glycine In this study, we have overexpressed a
functional, homo-oligomeric 5-HT We have used Spodoptera frugiperda Sf9 cells, infected with recombinant Autographa californica baculovirus, to express the mouse 5-HT
Enzymatic deglycosylation was
carried out using endoglycosidase H from Boehringer Mannheim. The
reaction was carried out with 5 µl of sample in a final volume of
20 µl in 10 mM MES, pH 5.5. Enzyme (1 milliunit) was added
to 1 aliquot, and another was left as the control. The reaction mixture
was incubated at 24 °C for 3 h. Western blots and binding assays
were carried out on each sample.
After centrifugation, the Triton and NaCl concentrations
and the pH level were adjusted to 0.5%, 300 mM, and 7.5,
respectively. The supernatant was loaded at room temperature onto a
column containing GR119566X-linked Affi-Gel 15 resin(16) ,
pre-equilibrated with 10 volumes of solubilization buffer adjusted as
described. After application, the resin was washed with 30 bed volumes
of medium salt buffer (0.25% Triton, 300 mM NaCl, 10 mM HEPES, pH 7.5), followed by 10 bed volumes of low salt buffer (as
medium salt, containing 150 mM NaCl). Finally, the resin was
washed with 10 bed volumes of low salt buffer containing 0.6% CHAPS.
The receptor was eluted from the resin using 10 bed volumes of the
medium salt buffer containing 0.6% CHAPS and 0.1 mM quipazine. The eluted sample was concentrated to 0.5 ml using a 50-ml Amicon
stirred cell and Microsep 100K concentrators (Filtron) and loaded onto
a 10/30 Superose 6 column (Pharmacia Biotech Inc.), pre-equilibrated
with 0.6% CHAPS, 500 mM NaCl, 1 mM EDTA, 10 mM HEPES, pH 7.5 (no protease inhibitors). Fractions containing pure,
oligomeric 5-HT
Immunofluorescence labeling studies were carried out on
infected insect cells that had been fixed and permeabilized to
determine the cellular distribution of the receptors. Labeling was to a
cytoplasmic epitope using fluorescein-labeled anti 5-HT
Figure 1:
Immunofluorescence detection of
5-HT
Figure 2:
Western blot of samples from 5-HT
The receptor preparation
obtained from the affinity column was further purified on a Superose 6
gel filtration column. In addition to removing the few contaminating
protein species, this step served both to remove the eluting ligand
(necessary to allow binding assays) and any aggregated receptor
particles. The peak corresponding to single channel molecules was
identified by electron microscopy visualization of individual
fractions. Exchange into a detergent more suited than CHAPS for
crystallization could also be done at this point. The sample from this
step was concentrated to a final concentration of 0.2-1.0 mg/ml
by ultrafiltration. The overall yield of purified receptor was
100-200 µg of protein from 3 liters of infected Sf9 cells.
Figure 3:
SDS-polyacrylamide gel electrophoresis
analyses of purified receptor and receptor following enzymatic
deglycosylation. The receptor sample obtained from the Superose 6
column was separated by SDS-polyacrylamide gel electrophoresis and
silver stained (panelA). Purified receptor was also
incubated with endoglycosidase H (EndoH) as
described under ``Experimental Procedures,'' and the
resulting samples were analyzed by Western blot (panelB). Molecular size markers are shown on the left of each gel.
The pharmacology of the receptor complex was also
investigated both in the membrane fraction and after the final
purification step. The dissociation constants (K
Figure 4:
Representative examples of radioligand
displacement analyses of membrane-bound and purified receptor. Two
antagonists (GR65630 and MDL72222) and two agonists (m-chloro-phenylbiguanide and 5-HT) were used at the
concentrations shown. The competition experiments were performed in
triplicate using 0.5 nM [
Figure 5:
Electron microscopy visualization of
uranyl acetate-stained soluble receptor preparations. Receptor
preparations after the gel filtration are shown both before (panelA) and after (panelB) concentration.
Examples of receptors viewed from above are circled (see
text). Examples of receptors on their sides are boxed. Views
of receptors from fields similar to that in panelA are shown enlarged to show the channel's 5-fold symmetry
more clearly (panelC). Scalebar in panelsA and B, 40
nm.
Closer analysis
of the particles seen from above should reveal the rotational symmetry
of the receptors and therefore the number of subunits, as has been done
for the native
Localization of recombinant channels
in internal membranes has been previously observed with proteins
expressed at high level using the baculovirus system. The gap junction
protein connexin-32 is found in the endoplasmic reticulum
membranes(25) , and only
Deglycosylation of active receptors did not reduce their ability to
bind ligand, demonstrating that sugars play no direct role in ligand
binding. The problem is therefore one of why only the fully
glycosylated oligomer binds ligand while partially glycosylated
oligomers do not. Three possibilities present themselves. The first is
that glycosylation is required to reach the final structure but is not
necessary to maintain it. For example, the presence of sugar residues
might alter the course of the folding pathway but not be necessary for
the thermodynamic stability of the final structure. In this case,
folding could also include oligomerization. The second possibility is
that glycosylation is simply a marker indicating the completeness of
other processing events required for activity. In other words, the
crucial event is one that occurs only after the receptor has been
completely glycosylated. If such an explanation is invoked, the change
must be one that occurs in the endoplasmic reticulum, as brefeldin A
has no detectable effect. A third possibility is that the partially
glycosylated receptors are all incorrectly folded, but this is unlikely
given the similar solubility of the various forms in detergents (see Fig. 2, lane2). Support for the first
explanation comes from analogous results reported for the epidermal
growth factor receptor and the insulin receptor. For both of these
proteins, it was reported that glycosylation was required to attain
activity but not to maintain it(31) . This was attributed to a
direct effect of glycosylation on the basis that tunicamycin blocked
receptor activation in vivo. The basic cause of the activation
was not investigated further. Tunicamycin added to the insect cell
culture medium also prevented the formation of active receptors in our
system. The simplest interpretation is again that glycosylation is
affecting activation directly. An indirect effect cannot be
discounted, however. Evidence in support of such an indirect effect
comes from experiments where the isolated insect cell membranes were
incubated with canine pancreatic microsomes. There was a reproducible
17% increase in ligand binding seen in the treated membranes compared
with controls. ( A more complete investigation of the
possible explanations will be required before a definitive answer is
known. Whether glycosylation turns out to have a direct or indirect
affect on receptor activation, it is probable that processing pathways
found only in eukaryotes are involved. One important consequence of
this is that it precludes the use of E. coli expression
systems for the expression of functional 5-HT
While the above results and other published studies of
recombinant receptors have shown that these receptors behave in much
the same way as native channels, there has been until now no direct
evidence that they assemble into similar oligomers. The images of the
homo-oligomeric 5-HT3 receptor presented here (Fig. 5) clearly
show the channel to have the same basic quaternary structure as the
well characterized Torpedo nACh receptor, with the subunits
surrounding a central pore. In addition, these recombinant 5-HT
The
recombinant 5-HT
Volume 270,
Number 11,
Issue of March 17, 1995 pp. 6056-6061
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
Receptors
Provides New Insights into Their Maturation and Structure (*)
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
(5-HT) receptor subunit cDNA
under the control of the polyhedrin promoter was shown to direct the
production of large amounts of functional 5-HT receptor in
insect cells, as assayed by Western blotting and ligand binding. After
solubilization, the receptor was purified to homogeneity by affinity
chromatography and characterized pharmacologically. The ligand binding
characteristics of the recombinant receptor were essentially identical
to those of the native receptor, both before and after purification.
Only fully glycosylated receptors bound to the ligand affinity resin,
although subsequent removal of the sugar did not affect ligand
binding. Visualization of the purified receptor using electron
microscopy showed that the receptor preparation contained a homogeneous
population of pentameric doughnut-shaped particles. The general
appearance of the recombinant homooligomeric channels was
indistinguishable from that of native 5-HT receptors.
Yields of purified receptor were of the order of 200 µg/3 liters of
original culture. The amount and homogeneity of the purified receptor
are sufficient to begin preliminary crystallization trials.
)glycine,
-aminobutyric acid type A, and 5-hydroxytryptamine (5-HT) receptors, is a major goal in molecular
neurobiology. The best characterized receptor of this type is the nACh
receptor from Torpedo electric rays, which was shown as early
as 1980 to be formed from five homologous subunits with the
stoichiometry 
(1) . These
subunits are arranged in the lipid bilayer as a pseudo pentamer, with
an aqueous pore formed in the center of the complex(2) . This
general arrangement (five subunits around a central pore) has now been
confirmed, either directly or indirectly, for most other members of the
channel family(3, 4, 5) . This accords with
their similar amino acid sequences, which additionally suggest that the
various receptors will share common elements of their tertiary
structure. subunit) (see (11) ). Such subunits
have, however, proved useful in experiments designed to elucidate
receptor structure and mechanism by mutagenesis (e.g.(12) and (13) ). The various subunits comprising
native receptors are so similar in terms of sequence that structurally
they are expected to be effectively interchangeable. Although this is
obviously a simplification (otherwise all subunits would readily form
homo-oligomeric channels), single subunits forming functional receptors
must represent some sort of consensus structure. The determination of
the three-dimensional structure of such a receptor should therefore
provide a valid model both for the receptor concerned and for the other
members of the superfamily. receptor. This receptor,
one of a number for the neurotransmitter serotonin (the others couple
to G-proteins), gates a cation-selective ion channel(14) .
Small quantities of the 5-HT receptor have been purified
from cell lines by various
groups(15, 16, 17) , and their analyses of
the purified protein seem to indicate that the native receptor is a
hetero-oligomer. Cloning, however, has led to the isolation of only one
subunit cDNA, encoding a mature polypeptide of 53.6 kDa (9) and
several slight variants (e.g.(18) ). All of these,
when expressed on their own in mammalian cells or Xenopus oocytes, form functional channels that display many, if not all,
of the characteristics of the native receptor (expression in oocytes, (9) ; expression in HEK cells,()). If and until
further subunits are cloned, the possibility remains that there is only
one subunit and that the results from purification studies have been
misleading. This study does not shed any further light on this, but any
eventual structure would inevitably be more relevant if it corresponded
to a truly native receptor. receptor
subunit with the eventual aim of crystallization for structural study.
The baculovirus system has been successfully used both for the
expression of neurotransmitter-gated ion channels (19, 20, 21, 22, 23) and
for the expression of other membrane channels and
pores(24, 25) . The high cost of scale-up with insect
cell culture means that expression levels in the milligram per liter
range are required for the isolation of the quantities of pure protein
required for structural studies. As some of the overexpression studies
show, this is a realistic goal(20, 24, 25) .
Materials
TNM-FH media was purchased in
powdered form from Sigma. [H]granisetron (80
Ci/mmol) was obtained from DuPont NEN. GR65630 and GR119566X were
synthesized by the Chemistry Research Department (Glaxo Group
Research). Quipazine dimaleate and m-chloro-phenylbiguanide
were obtained from Cookson Chemicals, and MDL 72222 was from Research
Biochemicals Inc. All other compounds were of the highest grade
available.Isolation of 5-HT
All baculoviral and insect cell manipulations were
performed essentially as described by Summers and Smith(33) .
The 5-HT
R Recombinant
BaculovirusR cDNA open reading frame, contained in the vector
pCDM8-5HTaS, was removed as a BamHI-EcoRI fragment. This was inserted directly into
the baculoviral transfer vector pVL1393 (Invitrogen), cut with the same
enzymes, to form the vector pT147. Recombinant virus was obtained by
homologous recombination in vivo, using 5 µg of pT147 and
1 µg of linearized wild-type A. californica nuclear
polyhedrosis virus DNA (Invitrogen).Cell Culture and Infection
Cells (Sf9)
were grown in TNM-FH media supplemented with 10% fetal calf serum.
Spinner flasks were used for cultures on the scale 50-500 ml, but
for protein production culture was performed in a 50-liter fermentor
(SGI). This allowed growth at a 32-liter scale, with control of pH,
oxygen levels, and temperature. Cells were infected with virus at a
multiplicity of infection of 5-10 and then harvested after 50 h
by continuous centrifugation. Cells were stored at -70 °C.Western Blots, Confocal Imaging, and
Deglycosylation
Immunolabeling experiments were performed
using polyclonal antisera (a gift from Dr. R. M. McKernan), raised
against the M3-M4 loop of the murine 5-HT receptor
expressed in Escherichia coli(26) . Western
blots were performed essentially following the method of Towbin et
al.(27) . For the fluorescence-labeled confocal imaging,
Sf9 cells were grown directly on 22-mm glass coverslips in 6-well
dishes. 48 h after infection (multiplicity of infection, 5-10)
with either the control or the 5-HTR virus, the cells were
washed with PBS and fixed for 5 min with glutaraldehyde (0.5%). After
three further washes with PBS, the cells were incubated with the
primary antisera (1:1000 dilution in PBS, 0.1% Triton X-100, overnight,
4 °C), biotinylated secondary antibody (1:200 dilution in PBS, 3 h,
22 °C), and fluorescein isothiocyanate-avidin (1 h, 22 °C). The
final wash was with saline-free phosphate buffer. Images were obtained
on an MRC-600 confocal microscope.Radioligand Binding Assays
For
radioligand binding experiments with either membranes or solubilized
receptor, the samples were incubated in HEPES buffer (10 mM,
pH 7.4) containing 0.1 nM [H]granisetron
in a final volume of 0.5 ml for 1 h at 0 °C. Nonspecific binding
was defined as that not displaced by 100 nM GR65630. For
competition studies, 0.5 nM [H]granisetron was used; concentrations
ranged from 0.01 to 18 nM in the saturation experiments. Bound
and unbound radioligand were separated by vacuum filtration using a
Brandel cell harvester onto GF/B filters (presoaked for 1 h in 0.1%
polyethyleneimine), followed by 2 
2-ml washes in ice-cold HEPES
buffer. All assays were performed in triplicate. Experimental data were
analyzed using the programs EBDA, LIGAND, and
KINETIC(28, 29) .Receptor Purification
All buffers
contained leupeptin (5 µg/ml), pepstatin A (2.5 µg/ml), and
phenylmethylsulfonyl fluoride (1 mM), and all operations were
carried out at 4 °C unless otherwise stated. The cell pellet from 3
liters of the fermentor culture was resuspended in 40 ml of 10 mM HEPES, pH 7.0, 1 mM EDTA, and lysed by 20 strokes in a
hand-held Potter homogenizer. The sample was centrifuged at 1000
g for 15 min. This procedure was repeated three times
with the pellet from the low-speed spin. The supernatants were pooled
and re-centrifuged at 130,000 g for 45 min to pellet
the membranes. These were resuspended in a total volume of 15 ml using
the HEPES buffer. For solubilization, the membranes were added to 10
volumes of the solubilization buffer (0.1% (w/v) Triton X-100, 10%
(v/v) glycerol, 10 mM bis-Tris propane, pH 9.0), and incubated
in this buffer for 1 h on ice with stirring. The sample was then
centrifuged at 130,000 g for 1 h to remove insoluble
material.R were pooled and further concentrated using
Microsep 100K concentrators. Where detergent exchange from CHAPS was
required, this was carried out on the Superose 6 column.Electron Microscopy
Samples were adsorbed
onto freshly glowdischarged, carbon-coated copper grids, washed in 100
mM cacodylate, pH 6.8, to remove detergent, and stained with
2% aqueous uranyl acetate. Micrographs were taken at nominal
magnifications of 45,000-49,000 on Phillips 420 or CM12
microscopes using low dose units and Kodak SO163 film.
5-HT
A cDNA clone encoding a subunit of the murine 5-HT Receptor Subunit Expression in Sf9
Cells receptor was obtained from Dr E. Kawashima (Glaxo, Geneva). After
introduction of the 5-HT receptor subunit cDNA into the A. californica nuclear polyhedrosis virus genome by double
homologous recombination, a single round of plaque purification was
sufficient to obtain three occlusion-negative viral clones, all of
which proved to be recombinant. Expression levels in the infected
insect cells were defined using radioligand binding assays and Western
blots. Initial Western blots indicated that all three viruses caused
the production of 5-HT receptor subunit polypeptides. Two
of the three isolated viruses caused the appearance of specific binding
sites for a radiolabeled 5-HT receptor antagonist in the
infected cells. Binding was originally assayed on whole cells. It was
assumed that the ligand was sufficiently lipophilic to enter the cells,
as addition of brefeldin A (25 µg/ml) to the culture medium did not
affect the observed ligand binding (brefeldin A blocks outward
transport from the endoplasmic reticulum). One of these two clones was
selected for further work. Scatchard plots indicated expression levels
of 0.4-2 mg of protein/liter of culture on a small scale
(50-500 ml), with approximately half this for cells grown in a
fermentor.R
subunit antisera raised against the region M3-M4(26) . Confocal
imaging of these cells revealed strong fluorescent labeling (Fig. 1), with a low background in the control cells (infected
with a virus expressing connexin-32). Western blots of the insect cell
membranes using the anti-5-HTR antisera showed multiple
bands (see Fig. 2, lane1), with three major
species running at molecular masses of approximately 43, 49, and 56
kDa. Numerous other bands were visible both above and below these, the
higher ones presumably due to incomplete dissociation of subunits in
the SDS. The lower bands can be attributed either to proteolysis or
incomplete translation products. Control Western blots showed no
cross-reactivity of the antisera with insect proteins. The receptor
could be solubilized using various detergents. Irrespective of the
solubilization conditions used, the pattern of polypeptides seen in
Western blots remained the same (Fig. 2, lanes1 and 2). Ligand binding assays using
[H]granisetron were used to estimate the
efficiency of different solubilization procedures, with the efficacy of
a variety of buffers, pHs, salts, and detergents being examined before
optimal conditions were determined. Solubilization in 0.1% Triton
X-100, 10 mM bis-Tris propane, 10% glycerol, pH 9.0, proved
most effective, typically solubilizing over 70% of the binding sites
present.
receptors in baculovirus-infected Sf9 insect cells.
The cells were infected with either a control virus (producing
connexin-32) (panelA, leftside)
or the 5-HT receptor virus (panelA, rightside, and panelB). Cells
were prepared and labeled with the anti 5-HT antisera as
described under ``Experimental Procedures.'' Both images in A are shown at the same magnification and image gain. Scalebar in panelA, 40 µm; scalebar in panelB, 20
µm.
receptor affinity purification. Samples were separated by
SDS-polyacrylamide gel electrophoresis and transferred onto a
nitrocellulose membrane. Anti 5-HTR antisera (1:1000
dilution) was used in the primary incubation, and biotinylated
anti-rabbit IgG was used in the secondary incubation. Horseradish
peroxidase coupled to avidin was added as the tertiary label. Samples
loaded were as follows: lane1, membrane fraction
from baculovirus-infected insect cells; lane2,
protein solubilized from membranes in 0.1% Triton; lane3, flow-through from affinity column after loading
solubilized sample; lane4, combined washes from
affinity resin; lane5, eluent from column. Only the
samples in lanes1 and 2 were from
equivalent volumes. Molecular size markers were stained with Ponceau S,
and their positions are shown on the left.
Receptor Purification to Homogeneity
The
solubilized receptor preparation was affinity purified using a
5-HT receptor antagonist (GR119566X) coupled to Affi-Gel
15. The Western blot of the eluted fraction revealed a single band of
56 kDa (Fig. 2, lane5). Ligand binding
assays conducted on the sample applied to the resin and on the column
flow-through indicated that >90% of the active receptors were
binding to the resin. Surprisingly, the flow-through from the column (lane3) appeared to be largely unchanged, with
little depletion of the 56-kDa polypeptide apparent. Passage of the
solubilized sample through the column reduced the intensity of the
56-kDa band by less than 20% (estimated from densitometer scans
relative to the 49-kDa band). Therefore, only a small fraction of the
polypeptides visible in Western blots (and therefore also the confocal
images) were active in terms of ligand binding. It also means that the
absolute level of protein expression (including inactive receptors), is
much higher than the estimates obtained from Scatchard plots. During
affinity chromatography, the detergent was exchanged from 0.1% Triton
to 0.6% CHAPS to enable concentration of the eluted sample without
concomitant concentration of the detergent.Characterization of Purified Receptor
Preparation
The purified receptor appeared as a single band
in silver-stained gels and Western blots (Fig. 3, A and B), with a faint band at a higher molecular weight
corresponding to the subunit dimer. The precise nature of this 56-kDa
purified polypeptide in relation to the other major species present in
the membranes was investigated further. To determine if the 56-kDa
polypeptide was glycosylated, the purified receptor was incubated with
endoglycosidase H. There was a shift in apparent molecular mass from
56 to 49 kDa in the labeled band on a Western blot (Fig. 3B). The size of the deglycosylated receptor
corresponded to the molecular weight of the major immunoreactive
polypeptide in the membranes (Fig. 2, lane1).
Single point binding assays on the deglycosylated protein and controls
revealed that, contrary to expectations, there was no loss of binding
sites upon removal of sugar residues. In contrast, when N-linked glycosylation in the infected insect cells was
blocked by the addition of 5 µg/ml tunicamycin to the culture
medium, no ligand binding was observed in crude membrane preparations
(data not shown).
)
for the ligand [H]granisetron were 0.37 ±
0.07 nM (n = 5) for the membrane fraction and
0.57 ± 0.06 nM (n = 3) for the purified
receptor. These were comparable with the value of 0.3 ± 0.06
nM reported for the receptor in rat cortical
membranes(30) . The values for maximal binding (B
) were 10.9 ± 1.1 pmol/mg protein for
the membranes and 4.2 ± 2.3 nmol/mg for the purified receptor.
This gave an overall purification factor of approximately 380-fold. All
of the Scatchard plots could be fitted using a single site model.
Displacement analyses for [H]granisetron with
various unlabeled agonists and antagonists (Fig. 4, A and B) gave the expected order of potency and apparent
affinities compared with native membranes (Table 1).
H]granisetron
on either membrane fractions of the baculovirus-infected insect cells (panelA) or the purified recombinant 5-HT receptor (panelB). Apparent affinities (K
) are given in Table 1. Binding
is shown as specific counts normalized to 100% at maximal binding. For
clarity in panelB, further data points at 100% are
not shown.
Electron Microscopy
Visualization
of the negatively stained solubilized receptor by electron microscopy (Fig. 5, A and B) was used to assay the
quality of receptor preparations and determine the gross structural
features of the channel. The views of the particles were related to
what is known of the structure of the nACh receptor. Many of the
particles (some being circled in Fig. 5A) appear as
toroidal rings (doughnut shaped) with a stain-filled center. These
rings have an apparent diameter of 70 Å. This view corresponds to
the receptor seen from above or below the bilayer (molecules in
negative stain being seen in projection). Other particles (boxed in Fig. 5A) can be seen as two parallel strands
separated by stain with dimensions of 70 140 Å. These can
be interpreted as side views of the receptor, with the pore being
visible along the whole length of the molecule. Other particles
represent tilted (or distorted) views between these two extreme
orientations. Fig. 5B shows the purified receptor at a
higher concentration (estimated at 0.2 mg/ml), where it again appears
homogenous and shows little or no aggregation. Storage of the receptor
at this protein concentration in 0.6% CHAPS at 4 °C for over 3
months resulted in no changes in the appearance of the isolated
receptor particles or in the level of aggregation.
-aminobutyric acid, type A receptor(5) . Fig. 5C shows particles similar to those seen in Fig. 5A, magnified photographically. These appear
clearly 5-fold symmetric, although given the limitations imposed by
negative stain, confirmation of this would require analysis of a larger
number of particles (see ``Discussion'').
The 5-HT
5-HT Receptors Appear to Remain in
Internal Membranes receptors are normally
located in the plasma membrane of neurons. The location of the
recombinant receptor in insect cells was investigated using confocal
imaging of permeabilized cells (Fig. 1). When expressed in
insect cells it appears that the majority of the receptors fail to
reach the cell surface. The pattern of labeling seen in confocal
imaging is most consistent with a location in internal (cytoplasmic)
membranes. Labeling of intact cells was not possible because the
antisera used was raised against a cytoplasmic epitope. The granular
nature of the labeling suggests that the receptor might be primarily
located in vesicles of some sort. The low proportion of active
receptors relative to the total amount of receptor detected by antibody
labeling makes it impossible to discount the possibility that there are
some receptors on the surface.0.1% of the expressed glycine
receptor subunit can be detected on the surface as functional
channels(20) . The internal buildup of recombinant protein,
both in these cases and with the 5-HT receptor, can be
attributed either to overloading of the transport machinery or to a
cellular response to the potential toxicity of the recombinant
proteins. In the latter case, it is interesting to speculate that the
lack of success in expressing such membrane channel proteins in
prokaryotic organisms such as E. coli may be due to their lack
of alternative membranes in which to store them.Glycosylation Is Central to Receptor
Maturation
The isolation of a single protein band from the
affinity resin was unexpected, given the number of immunoreactive bands
in the crude extract. The receptor oligomers formed from such a
heterogeneous mixture might be expected to be heterogeneous themselves.
Even if the full-length glycosylated subunit is the only form with
binding activity, an oligomer containing at least one such subunit
might still be expected to attach to the resin. This obviously does not
occur, as no shorter (unglycosylated) forms appear in the
affinity-purified sample. Even after purification, there still appears
to be significant amounts of a polypeptide corresponding to the
full-length glycosylated subunit in the flow-through (Fig. 2, lane3), whereas 90% of the active receptor binds to
the resin. This suggests that mixed receptors do occur but that they do
not bind either to the affinity resin or to free ligand. )This microsome-mediated increase suggests
the involvement of endoplasmic reticulum-linked enzyme systems other
than those for N-linked glycosylation. This is because N-linked glycosylation is normally thought of as an entirely
cotranslational event. Any post-translational increase in receptor
activity must presumably therefore be caused by other enzymes present
in the microsome preparation. receptor
subunits and possibly also for the other subunits from the superfamily. Recombinant Homo-oligomeric Receptors Resemble Their
Native Counterparts
The recombinant receptor, both in
membranes and after purification, showed similar ligand binding
characteristics to the native receptor (see Fig. 4and Table 1). One piece of information not accessible from studies of
native membranes is the B, and thereby the
stoichiometry, of the purified molecule. If our preparation is assumed
to be completely homogenous, then a figure of 4.2 nmol/mg protein
corresponds to a binding species of 238 kDa. This is very close to the
figure of 280 kDa that would be expected if the receptor was
pentameric and there was one binding site per molecule. It is tempting
to speculate that this might explain why only fully glycosylated
receptors were purified, with five correctly modified and assembled
subunits required before binding is seen. It should be noted that the
error in the B (due primarily to the errors in
protein estimation) is sufficient to allow for target sizes down to 90
kDa (at 3 s). receptors are indistinguishable from native 5-HT receptors purified from the neuroblastoma cell line
NG108-15 and viewed by electron microscopy. These channels have
been analyzed and shown to be 5-fold symmetric. (
)This
confirms that the recombinant protein assembles in the same way as the
native channel. It is possible that small differences between the
hetero- and homo-oligomeric forms might be seen at higher resolution.
Any such differences should not affect the utility of recombinant
homo-oligomers as structural models for the native channel.Crystallization Trials as the Next
Step
The eventual aim of this work is the crystallization
and structure determination of the 5-HT receptor. There are
several main requirements before crystallization of any protein can
reasonably be considered. First, there must be enough purified protein
to produce a solution with a final protein concentration in the
milligram per milliliter range. Second, the preparation must be
homogeneous. Overall purity is often crucial for the success of
crystallization experiments, but more subtle variations in the protein
preparation can have an impact. For example, glycosylation can be a
potential source of heterogeneity, and it is possible that the sugar
residues will have to be removed before suitable crystals are obtained,
as has been observed with lactoferrin(32) . The final
consideration is one of stability. This is perhaps most important with
an oligomeric integral membrane protein, where there are the dual
problems of oligomer stability and nonspecific aggregation. receptor prepared here satisfies all of
the above requirements. The purified receptor had an upper solubility
limit of 1 mg/ml in the conditions described, and preparation of
reasonable (200 µl) quantities of such a solution was possible from
3-liter preparations at the expression levels achieved. Working with
homo-oligomeric rather than hetero-oligomeric channels removes one
potential source of heterogeneity, but there are still others that will
need considering, such as glycosylation. If it proves necessary to
remove the sugar moieties, then enzymatic deglycosylation of the
purified receptor will be the only option, as in vivo inhibitors of glycosylation prevent the formation of active
receptor. As for stability, there was no visible deterioration in the
sample over a period of months. The receptor preparation described here
therefore demonstrates the main characteristics required for eventual
crystallization.
), 5-hydroxytryptamine;
5-HTR, 5-hydroxytryptamine receptor; MES,
2-(N-morpholino)ethanesulfonic acid; CHAPS,
3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic
acid; PBS, phosphate-buffered saline.)))
We thank Eric Kawashima (Glaxo, Geneva) for supplying
the pCDM8-5HT3aS clone, Gavin Kilpatrick (Glaxo, Stevenage) for
the ligand GR119566X, and Ruth McKernan (Merck, Sharp, and Dohme,
Harlow) for providing the anti-5HT antisera. We are also
very grateful to Stephen Hunt for help with the confocal microscopy and
Jill Chirnside for much invaluable advice on running the fermentor.
Finally, we thank Chris Staples for keeping lab work going while the
manuscript was written, Rameen, Reinhard, Jane, and Olga for useful
comments on its content, and Nigel Unwin, whose suggestions throughout
the project have been crucial to its success.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Ilegems, H. M. Pick, C. Deluz, S. Kellenberger, and H. Vogel Noninvasive Imaging of 5-HT3 Receptor Trafficking in Live Cells: FROM BIOSYNTHESIS TO ENDOCYTOSIS J. Biol. Chem., December 17, 2004; 279(51): 53346 - 53352. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Hapfelmeier, C. Tredt, R. Haseneder, W. Zieglgansberger, B. Eisensamer, R. Rupprecht, and G. Rammes Co-expression of the 5-HT3B Serotonin Receptor Subunit Alters the Biophysics of the 5-HT3 Receptor Biophys. J., March 1, 2003; 84(3): 1720 - 1733. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Steward, F. G. Boess, J. A. Steele, D. Liu, N. Wong, and I. L. Martin Importance of Phenylalanine 107 in Agonist Recognition by the 5-Hydroxytryptamine3A Receptor Mol. Pharmacol., June 1, 2000; 57(6): 1249 - 1255. [Abstract] [Full Text]
|
|
|
|
|
|
A. E. Dubin, R. Huvar, M. R. D'Andrea, J. Pyati, J. Y. Zhu, K. C. Joy, S. J. Wilson, J. E. Galindo, C. A. Glass, L. Luo, et al. The Pharmacological and Functional Characteristics of the Serotonin 5-HT3A Receptor Are Specifically Modified by a 5-HT3B Receptor Subunit J. Biol. Chem., October 22, 1999; 274(43): 30799 - 30810. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Coultrap, H. Sun, T. E. Tenner Jr., and T. K. Machu Competitive Antagonism of the Mouse 5-Hydroxytryptamine3 Receptor by Bisindolylmaleimide I, a "Selective" Protein Kinase C Inhibitor J. Pharmacol. Exp. Ther., July 1, 1999; 290(1): 76 - 82. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |
