A novel site on γ3 subunits important for assembly of GABAA receptors
-
Upload
nguyenkhanh -
Category
Documents
-
view
218 -
download
0
Transcript of A novel site on γ3 subunits important for assembly of GABAA receptors
1
A novel site on γ3 subunits important for assembly of
GABAA receptors*
Sarto I.‡, Klausberger T.‡, Ehya N.‡, Mayer B.§, Fuchs K.‡, and Sieghart
W.‡¶
‡Division of Biochemistry and Molecular Biology, Brain Research Institute, University of
Vienna and Section of Biochemical Psychiatry, University Clinic for Psychiatry, A-1090
Vienna, Austria and §Institute for Theoretical Chemistry and Molecular Structural Biology,
University of Vienna, A-1090 Vienna, Austria
¶To whom correspondence should be addressed: Division of Biochemistry and Molecular
Biology, Brain Research Institute, University of Vienna, Spitalgasse 4, A-1090 Vienna,
Austria, Tel.:++43-1-4277-62950; Fax: Tel.:++43-1-4277-62959; E-
mail:[email protected]
Running title: GABAA receptor assembly
* This work was supported by grant P12637-Med of the Austrian Science Fund.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
2
GABAA receptors are ligand-gated chloride channels and are the major
inhibitory transmitter receptors in the central nervous system. The majority of these
receptors is composed of two α, two β and one γ subunit. To identify sequences
important for subunit assembly, we generated C-terminally truncated and chimeric γ3
constructs. From their ability to associate with full length α1 and β3 subunits, we
concluded that amino acid sequence γ3(70-84) either directly interacts with α1 or β3
subunits or stabilizes a contact site elsewhere in the protein. The observation that this
sequence contains amino acid residues homologous to γ2 residues contributing to the
benzodiazepine binding site at the α1/γ2 interface suggested that in α1β3γ3 receptors the
sequence γ3(70-84) is located at the α1/γ3 interface. In the absence of α1 subunits this
sequence might allow assembly of β3 with γ3 subunits. Other experiments indicated that
sequences γ3(86-95) and γ3(94-107) that are homologous to previously identified
sequences important for assembly of γ2 subunits, are also important for assembly of γ3
subunits. This indicates that during assembly of the GABAA receptor, more than one N-
terminal sequence is important for binding to the same neighbouring subunit. Whether
the three sequences investigated are involved in direct interaction or stabilize other
regions involved in intersubunit contacts, has to be further studied.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
3
γ-Aminobutyric acidA (GABA)A1 receptors are the major inhibitory transmitter
receptors in the central nervous system and mediate fast synaptic inhibition by opening an
intrinsic chloride channel (1). These receptors carry binding sites for a number of
pharmacologically and clinically important drugs, such as benzodiazepines, barbiturates,
steroids, anesthetics and convulsants that modulate GABA-induced chloride ion flux by
interacting with separate and distinct allosteric binding sites (2).
The GABAA receptor is a hetero-oligomeric protein consisting of five subunits (3, 4).
So far at least 19 GABAA receptor subunits belonging to several subunit classes (six α, three
β, three γ, one δ, one ε, one π, one θ and three ρ) have been identified in the mammalian
nervous system (5, 6). Although a variety of subunits can be co-expressed within the same
neuron (7, 8), not all receptors that theoretically can be formed are actually formed in the
brain (9-11). Thus, GABAA receptor heterogeneity is limited by the temporal and spatial
expression of subunits (12) and by structural and conformational requirements during
assembly (13). The majority of GABAA receptors is composed of two α, two β and one γ
subunit (4, 11, 14-17) where the γ subunit is located between an α and a β subunit (4, 18).
The assembly of hetero-oligomeric receptors of the ligand gated ion channel
superfamily comprising the nicotinic acetylcholine (nACh) receptor, GABAA receptor,
glycine receptor, and 5-hydroxytryptamine, type 3 receptor, is a complex, multistep process
that also requires conformational changes in the involved subunits (for review see Ref. 19).
The assembly of subunits seems to occur in the endoplasmic reticulum and to involve
interaction with chaperone molecules (20, 21). The processes that lead from single subunits to
completely assembled and pharmacologically functional receptors are still a matter of debate.
For the assembly of nACh receptors a sequential mechanism was proposed (22-24), while for
the GABAA receptors it is not clear whether oligomerization can start randomly from any
possible subunit dimer or follows a strict order of assembly intermediates (25).
In the whole receptor superfamily the major determinants for intersubunit contacts
seem to be located within the N-terminal domains (26-28). For the GABAA receptor, specific
amino acid sequences have been identified in this domain that seem to mediate heteromeric
and homomeric assembly of β subunits (29) or assembly of α with β subunits (30). In
addition, the γ2 amino acid sequences γ2(91-104) and γ2(83-90) were identified as sites
important for assembly with α1 and β3 subunits, respectively (31).
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
4
The amino acid sequence of the γ3 subunit is 64% identical with that of the γ2 subunit
(32, 33) and each of these subunits contributes to the formation of the benzodiazepine binding
site that is located at the interface between the α and γ subunit (34). In the present study it
was investigated whether amino acid sequences homologous to putative assembly sites on γ2
subunits are also important for assembly of γ3 subunits. By using truncated as well as
chimeric constructs a novel amino acid sequence was identified on the γ3 subunit that either
directly interacts with α1 or β3 subunits, or stabilizes a contact site in a different region of the
protein.
In addition, it was demonstrated that sequences homologous to those previously
identified on γ2 subunits are also important in γ3 subunits for interaction with neighbouring
subunits. It was, thus, concluded that during assembly of GABAA receptors more than one N-
terminal γ3-sequence is important for forming contacts with the same adjacent subunit.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
5
EXPERIMENTAL PROCEDURES
AntibodiesThe antibodies anti-peptide α1(1-9), anti-peptide β3(1-13), anti-peptide
α1(328-382), anti-peptide β3(345-408) and anti-peptide γ3(1-35) were generated and affinity
purified as described previously (4, 9). None of these antibodies exhibited any cross-reactivity
with any other GABAA receptor subunits as demonstrated in Western blot and
immunoprecipitation studies using recombinant GABAA receptors (9).
Generation of cDNA constructsFor the generation of recombinant receptors, α1, β3 and
γ3 subunits of GABAA receptors from rat brain were cloned and subcloned into pCDM8
(Invitrogen, San Diego, CA) and subsequently into pCI (Promega, Madison, WI) expression
vectors as described previously (4). Truncated subunits were constructed by PCR
amplification using the full-length subunit as a template. The PCR-Primers contained EcoRI
and XhoI restriction sites, which were used to clone the fragments into pCI vectors. The
truncated subunits were confirmed by sequencing. Chimeras were constructed using the ”gene
SOEing” technique (35) and were cloned into pCI vectors using the EcoRI and XhoI
restriction sites of the primers.
Culture and Transfection of HEK 293 cellsTransformed human embryonic kindney
(HEK 293) cells (CRL 1573; American Type Culture Collection, Rockville, MD) were
cultured as described (4). 3 x 106 cells were transfected with 20 µg of subunit cDNA for
single subunit transfection using the calcium phosphate precipitation method (36). On co-
transfection with two or three different subunits, for each subunit 10 µg or 7 µg cDNA per
subunit was used, respectively. A total of about 20 µg cDNA per transfection and a cDNA
ratio of 1:1:1 seemed to be optimal for the expression of GABAA receptors under the
conditions used as judged by receptor binding studies in cells transfected with α1, β3 and γ3
subunits.
The cells were then harvested 36 h after transfection. At this time point the number of
[3H]Ro 15-1788 binding sites formed per mg protein was at its maximum for cells transfected
with α1, β3 and γ3 subunits. Results obtained, however, did not change when cells were
harvested 34 - 48 h after transfection. In addition, judged by Western blot analysis, expression
levels of full length, truncated or chimeric subunits were comparable at all harvesting times.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
6
Purification and immunoprecipitation of complete, truncated and chimeric subunitsThe
culture medium was removed from transfected HEK 293 cells and cells from 4 culture dishes
(diameter 9,4 cm) were extracted with 800 µl of a Lubrol extraction buffer (1% Lubrol PX,
0.18% phosphatidylcholine, 150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl, pH 7.4,
containing 0.3 mM phenylmethylsulfonylfluorid (PMSF), 1 mM benzamidine and 100 µg/ml
bacitracin) for 8 h at 4oC. The extract was centrifuged for 40 min at 150,000 x g at 4oC.
For Western blot analysis, the clear supernatant was incubated overnight at 4oC under
gentle shaking with 15 µg antibodies directed against the full length subunit. After addition of
Immunoprecipitin (Life Technologies, Gaithersburg, MD; preparation see (4)) and 0.5%
nonfat dry milk powder and shaking for additional 3 h at 4oC, the precipitate was washed
three times with a low salt immunoprecipitation buffer (IP low buffer) (50 mM Tris-HCl,
0.5% Triton X-100, 150 mM NaCl, and 1 mM EDTA, pH 8.0). The precipitated proteins were
dissolved in sample buffer (108 mM Tris-sulfate, pH 8.2, 10 mM EDTA, 25% (w/v) glycerol,
2% SDS and 3% dithiothreitol). SDS-PAGE and Western blot analysis with digoxygenized
antibodies were performed as described (4).
All truncated or chimeric constructs used in this study could be expressed to a comparable
extent upon single transfection into HEK cells. On co-transfection of different constructs,
however, the stability of fragments that could not bind stably to each other was reduced. This
might have been caused by proteolytic degradation due to an unstable or unproductive
interaction of the fragments. In all control experiments the extent of expression of fragments
was therefore determined in singly transfected HEK cells (13).
For [3H]muscimol binding studies, the extracts of transfected HEK cells were incubated
overnight at 4oC under gentle shaking with either 2,5µg, 10µg or 20µg γ3(1-35) antibodies,
or in a separate control experiment with a mixture of 5µg α1(1-9), 10µg β3(1-13) and 10µg
γ3(1-35) antibodies per culture dish. After addition of Immunoprecipitin and 0.5% nonfat dry
milk powder and shaking for additional 3 h at 4oC, the precipitate was washed three times
with IP low buffer. The precipitated proteins were resuspended in 50 mM Tris/citrate buffer,
pH 7.4 containing 0,1% Triton X-100, at a protein concentration in the range of 0.1 - 1 mg/ml
as measured with the BCA protein assay kit (Pierce Chemical) with bovine serum albumin as
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
7
standard. Aliquots of the suspension were then incubated for 60 min at 4oC in a total of 1 ml
of a solution containing 50 mM Tris/citrate buffer, pH 7.4, 0,1% Triton X-100, and 20nM of
[3H]muscimol (28.1 Ci/mmol, DuPont NEN) in the absence or presence of 1 mM GABA.
Precipitates were then filtered through Whatman GF/B filters, and the filters were washed
twice with 3.5 ml ice-cold 50 mM Tris/citrate buffer and were then subjected to scintillation
counting. Unspecific binding in the presence of 1 mM GABA was subtracted from total
[3H]muscimol binding, to result in specific binding (37).
Immunoprecipitation of receptors expressed on the cell surfaceThe culture medium was
removed from HEK 293 cells transfected with cDNA (21 µg per 3 x 106 cells) of GABAA
receptor subunits (cDNA ratio 1:1:1) and the cells were washed once with phosphate buffered
saline (PBS: 2.7 mM KCl, 1.5 mM KH2PO4, 140 mM NaCl, and 4.3 mM Na2HPO4, pH 7.3).
Cells were then detached from the culture dishes by incubating with 2.5 ml of 5 mM EDTA in
PBS for 5 min at room temperature. The resulting cell suspension was diluted in 6.5 ml of
cold Dulbecco´s modified Eagle medium (DMEM; Life Technologies, Gaithersburg, MD)
and centrifuged for 5 min at 1000 x g.
For Western blot analysis, the pellet from two dishes was incubated with 30 µg α1(1-9) or
γ3(1-35) antibodies in 3 ml of the same medium for 30 min at 37oC. Cells were again pelleted
and free antibodies were removed by washing twice with 10 ml of PBS buffer. Then receptors
were extracted with IP low buffer containing 1% Triton X-100 for 1 hr under gentle shaking.
Cell debris was removed by centrifugation (30 min; 150,000 x g; 4oC). After addition of
Immunoprecipitin and 0.5% nonfat dry milk powder to the clear supernatant and shaking for 3
h at 4oC, the precipitate was centrifuged for 10 min at 10,000 x g and washed three times with
IP low buffer. The precipitated proteins were dissolved in sample buffer and subjected to
SDS-PAGE and Western blot analysis using digoxygenized antibodies. Secondary antibodies
(Anti-Digoxygenin-AP, Fab fragments; Roche Diagnostics GmbH, Mannheim, Germany)
were visualized by the reaction of alkaline phosphatase with CSPD (Tropix, Bedford, MA,
USA). Protein bands were quantified by densitometry of Kodak X-Omat S films with the
Docu Gel 2000i gel documentation system using RFLP scan software (MWG Biotech,
Ebersberg, Germany). The linear range of the detection system was established by
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
8
determining the antibody response to a range of antigen concentrations following
immunoblotting. The experimental conditions were designed such that immunoreactivities
obtained in the assay were within this linear range, thus permitting a direct comparison of the
amount of antigen applied per gel lane between samples. Different exposures of the same
membrane were used to ensure that the measured signal was in the linear range of the x-ray
film.
For radioligand binding studies, the pellet from six dishes was incubated with a mixture of
20µg α1(1-9), 60µg β3(1-13) and 60µg γ3(1-35) antibodies or with 60µg γ3(1-35) antibodies
in 6 ml of DMEM medium for 45 min at 37oC. Cells were again pelleted and free antibodies
were removed by washing twice with 10 ml of PBS buffer. Then receptors were extracted
with IP low buffer containing 1% Triton X-100 for 1 h under gentle shaking. Cell debris was
removed by centrifugation (30 min; 150,000 x g; 4oC). After addition of Immunoprecipitin
and 0.5% nonfat dry milk powder and shaking for 3 hr at 4oC, the precipitate was centrifuged
for 10 min at 10,000 x g and washed three times with IP low buffer. The precipitated proteins
were resuspended in 50 mM Tris/citrate buffer, pH 7.4 containing 0,1% Triton X-100, at a
protein concentration in the range of 0.1 - 1 mg/ml and subjected to [3H]muscimol binding
assay as described above.
To verify that only receptors on the cell surface were labeled by the antibodies, parallel
samples were incubated with antibodies directed against the intracellular loop of GABAA
receptor subunits (experiments not shown). These antibodies could not precipitate any
GABAA receptor subunits under the conditions used. A possible redistribution of the
antibodies during the extraction procedure could be excluded by an experiment performed
analogous to that described (31).
Secondary structure predictionThe EMBL PredictProtein server (38) was used to
align γ3 sequences to homologous sequences available in the SwissProt database and then to
predict the secondary structure based on the set of aligned sequences by the PHDsec method
(39, 40). The significance of this neural network – based secondary structure prediction for
the sequences of interest was evaluated by force field calculations within a Dynamic Monte
Carlo (DMC) optimization scheme (41). In this procedure the structure optimization started
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
9
with the sequence in extended conformation (φ, ψ = 180°). Within one DMC step a single φ,
ψ or side chain dihedral angle was updated in the range [-180,180], the conformational energy
was calculated within the ECEPP/3 force field (42) and the free energy of solvation was
computed based on a continuum solvation model. 2 x 105 DMC steps were performed. The
final acceptance probability was given by a modified Metropolis criterium considering both,
conformational and solvation energies. Details on the optimization algorithm are given in
(43).
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
10
RESULTS
Truncated γ3 constructs are able to assemble with full length α1 and β3
subunitsRecently, it has been demonstrated that γ2(1-113) was the shortest γ2 fragment that
still could assemble with full length α1 and β3 subunits (31). In the present study it was
investigated whether the same holds true for fragment γ3(1-116) that is homologous to γ2(1-
113) (Fig. 1A). For this, fragment γ3(1-116) was either singly transfected into HEK cells or
co-transfected with full length α1 or β3 subunits. Expressed subunits were extracted from
these cells and immunoprecipitated either with γ3(1-35), or with α1(328-382) or β3(345-408)
antibodies, respectively. The precipitate was subjected to SDS-PAGE and Western blot
analysis using digoxygenized γ3(1-35) antibodies. Fragment γ3(1-116) contained a single
glycosylation site and migrated as two protein bands with apparent molecular mass 14 and 16
kDa that presumably represented the unglycosylated and glycosylated fragments, respectively
(Fig. 1B). As shown in Fig. 1C+D, fragment γ3(1-116) could be co-precipitated with α1 or β3
subunits from appropriately transfected HEK cells. These data indicated that binding sites for
α1 and β3 subunits are located on homologous N-terminal fragments of γ2 and γ3 subunits.
To identify amino acid sequences important for binding, the even shorter N-terminal
γ3 fragments γ3(1-84) and γ3(1-72) were generated and their interaction with full length α1
and β3 subunits was investigated after co-transfection into HEK cells. Fragments γ3(1-84)
and γ3(1-72) exhibit a molecular mass of 9.4 and 8.2 kDa, respectively, and migrated as
single protein bands (Fig. 1B) consistent with the observation that these fragments do not
contain a putative glycosylation site (33). Whereas γ3(1-84) still could be co-precipitated with
α1 and β3 subunits (Fig. 1C+D), no co-precipitation with these subunits could be detected for
the shorter fragment γ3(1-72). Although expression of fragment γ3(1-72) was reduced
compared with γ3(1-84) (Fig. 1B), a stable binding of this fragment to α1 or β3 subunits
would have been detected. The inability of α1(328-382) or β3(345-408) antibodies to co-
precipitate the fragment γ3(1-72) or to precipitate the fragment γ3(1-116) after single
transfection into HEK cells (experiments not shown) confirmed the conclusion that these
antibodies did not cross-react with the γ3 fragments used.
The observation that γ3(1-84) still could be co-precipitated with α1 and β3 subunits
was surprising, because this fragment no longer contained the sequences γ3(86-93) and γ3(94-
107) that are homologous to the sequences γ2(83-90) and γ2(91-104) that previously have
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
11
been identified to be important for binding of γ2 to β3 and α1 subunits, respectively (Fig. 1A).
This indicated that γ3 subunits possibly use sequences different from those of γ2 subunits for
assembly with α1 and β3 subunits and that these γ3 sequences are located within the N-
terminal 84 amino acids of the mature γ3 peptide.
Amino acid sequence γ3(70-84) is important for binding to α1 or β3 subunitsIn
order to identify amino acid sequences important for binding to α1 and/or β3 subunits, it was
investigated which γ3 sequences could induce binding to these subunits after incorporation
into a fragment that originally could not bind to α1 or β3 subunits. The fragment α1(1-68)
seemed to be suitable for this purpose because it is homologous to γ3(1-84), but could not be
co-precipitated with α1 or β3 subunits after co-expression in HEK cells (Fig. 2A).
Several chimeras were constructed in which the C-terminal part of the α1(1-68)
fragment was replaced by the corresponding γ3 sequence (Fig. 2A). These chimeras were
transfected into HEK cells together with full length α1 or β3 subunits and expressed
fragments were co-precipitated using α1(328-382) or β3(345-408) antibodies, respectively
(Fig 2B+C). These antibodies were directed against amino acid sequences of the intracellular
loop of α1 or β3 subunits, respectively, and only recognized full length subunits but not the
truncated chimeras. The precipitate was subjected to SDS-PAGE and the proteins were
detected using digoxygenized α1(1-9) antibodies in Western blots. The actual expression of
the chimeras was confirmed by precipitation with α1(1-9) antibodies and detection with
digoxygenized α1(1-9) antibodies (Fig. 2D).
Chim1, in which the sequence α1(35-68) of the α1(1-68) fragment was replaced by
the homologous sequence γ3(51-84), could be co-precipitated with α1 as well as with β3
subunits (Fig. 2A-C). Chim2, however, in which the sequence α1(35-53) was replaced by the
homologous sequence γ3(51-69), could not be co-precipitated with full length α1 or β3
subunits. This indicated that the sequence important for binding to α1 and β3 subunits is
located within the amino acid sequence γ3(70-84). This conclusion was confirmed by the
observation that Chim3, in which the sequence α1(54-68) of the α1(1-68) fragment was
replaced by the homologous sequence γ3(70-84), was able to bind to α1 and β3 subunits (Fig.
2A-C).
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
12
The whole sequence γ3(70-84) is necessary for efficient assembly with full length α1
and β3 subunitsTo investigate which part of the γ3(70-84) sequence is important for
assembly with α1 or β3 subunits, four additional chimeras of the α1(1-68) fragment were
constructed (Fig. 3A).
In Chim4, amino acids α1(61-68) were replaced by amino acids γ3(77-84), in Chim5
amino acids α1(57-68) were replaced by γ3(73-84), in Chim6 α1(54-59) was replaced by
γ3(70-75) and in Chim7 α1(54-62) was replaced by γ3(70-78). All these chimeras were
expressed to a similar extent after transfection into HEK cells (experiments not shown). None
of these chimeras could be co-precipitated with full length β3 subunits indicating that the
complete γ3(70-84) sequence is needed for binding to this subunit (Fig. 3B). Interestingly,
Chim6 could be co-precipitated with full length α1 subunits, but the extent of co-precipitation
was only 6±2% (mean±S.E.M., n=3) of that of Chim3 as quantified by densitometry of
protein bands (see Experimental Procedures). Binding of Chim7 to α1 subunits was stronger
than that of Chim6, but the extent of co-precipitation of Chim7 with α1 subunits was still only
19±5% (mean±S.E.M., n=3) of that of Chim3. Thus, the complete γ3(70-84) sequence was
necessary to allow binding to α1 subunits with full efficiency. Interestingly, all these
chimeras contained a single glycosylation site and thus, migrated as two peptides with a
molecular mass of 10 and 12 kDa. On co-precipitation experiments, predominantly the
glycosylated form of Chim3, Chim6 and Chim7 or Chim3 was precipitated with α1 or β3
subunits, respectively (Fig. 3B).
The sequence γ3(70-84) is important for assembly of GABAA receptors composed of
α1β3γ3 subunitsTo investigate the importance of the γ3(70-84) sequence not only for the
assembly of truncated with the full length subunits, but also for assembly of full length
subunits and receptors expressed on the cell surface, a full length γ3 chimera (γ3*) was
constructed in which the sequence γ3(70-84) was replaced by the homologous sequence
ρ1(92-106). The sequence of the ρ1 subunit shares 32,4% sequence identity with the γ3
subunit (Align; Genestream; IGH Montpellier, France) and was chosen because ρ1 was
reported not to assemble with α1 or β3 subunits (44, 45). In control experiments (Fig. 4A) it
was demonstrated that the extent of expression of the γ3* chimera was similar to that of the γ3
subunit in HEK cells. Subunits γ3 as well as γ3* migrated as protein bands with apparent
molecular mass of 43 kDa (46). The weak staining of protein bands with an apparent
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
13
molecular mass of about 51 kDa presumably was caused by a cross-reactivity of the antibody
with a HEK cell protein (Fig. 4A).
HEK cells were co-transfected with α1, β3 and γ3* or with α1, β3 and γ3
subunits. GABAA receptors expressed on the cell surface were labeled by an incubation of
intact cells with γ3(1-35) antibodies. Antibody-labeled receptors were extracted and
precipitated by addition of Immunoprecipitin. The precipitate was subjected to SDS-PAGE
and Western blot analysis using digoxygenized γ3(1-35) antibodies (Fig. 4B). The staining
intensity of the protein bands was quantified and results obtained from two experiments
indicated that the amount of γ3* subunits on the gel was only 19% and 31% of of that
precipitated from α1β3γ3 transfected cells. The Western blot was then stripped and analyzed
using digoxygenized α1(1-9) antibodies, followed by another stripping procedure as well as
detection with digoxygenized β3(1-13) antibodies. The amounts of α1 subunits (51 kDa) co-
precipitated with γ3* subunits were 23% and 28%, and those of β3 subunits (54 kDa) were
23% and 30% of those precipitated from the surface of α1β3γ3 transfected cells. The similar
reduction in α1, β3, and γ3* subunits precipitated by γ3(1-35) antibodies suggested that all
γ3* subunits on the cell surface were associated with α1 and β3 subunits. The reduction in the
amount of γ3* containing receptors formed on the cell surface indicated the importance of the
sequence γ3(70-84) for the assembly of full length subunits.
In similar experiments, GABAA receptors formed on the surface of α1β3γ3* or
α1β3γ3 transfected cells were labeled by an incubation with α1(1-9) antibodies and the
antibody labeled receptors were again extracted, precipitated by Immunoprecipitin and
subjected to SDS-PAGE and Western blot analysis using digoxygenized γ3(1-35) antibodies.
In agreement with results shown in Fig. 4B, the amount of γ3* subunits co-precipitated by the
α1(1-9) antibody was only 22±4% (mean±S.E.M., n=6, from six different transfections) of
that of the wild-type γ3 subunit (Fig. 5).
The Western blots were then again stripped and re-analyzed using digoxygenized
α1(1-9) or β3(1-13) antibodies. Results obtained revealed that comparable amounts of α1 or
β3 subunits could be detected on the surface of cells transfected with α1, β3 and γ3*, or α1,
β3 and γ3 subunits (Fig. 5). Similar results were obtained when the order of detection of
subunits was changed and Western blots were first probed with digoxygenized α1(1-9)
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
14
antibodies and after stripping were re-analyzed with digoxygenized γ3(1-35) or β3(1-13)
antibodies. The finding that in spite of the strong reduction of γ3* containing receptors
comparable amounts of α1 or β3 subunits could be detected on the surface of α1β3γ3* or
α1β3γ3 transfected cells, then seems to indicate that in addition to α1β3γ3* receptors, α1β3
receptors were formed in these cells.
This conclusion was supported by experiments investigating [3H]muscimol binding on
the cell surface. Previous studies have indicated that [3H]muscimol binding sites on the cell
surface are only formed by receptors composed of αβ or αβγ subunits (37). HEK cells
transfected with α1β3γ3 or α1β3γ3* subunit combinations were incubated either with a
mixture of α1(1-9), β3(1-13) and γ3(1-35) antibodies for labeling of αβ plus αβγ receptors, or
with γ3(1-35) antibodies for labeling of γ3 or γ3* containing receptors. Labeled receptors
were then extracted from the cells, precipitated by the addition of Immunoprecipitin, and
incubated with 20nM [3H]muscimol in the absence or presence of 1mM GABA. The mixture
of antibodies precipitated comparable amounts of specific [3H]muscimol binding sites (and
thus of total receptors) from the surface of α1β3γ3 or α1β3γ3* transfected cells (340±14 or
351±25 fmol/mg protein, respectively; mean±S.E.M., n=3). The number of [3H]muscimol
binding sites precipitated by γ3(1-35) antibodies, however, was only 96±5 or 30±4 fmol/mg
protein, respectively (mean±S.E.M., n=3). The apparently low percentage of [3H]muscimol
binding sites precipitated by γ3(1-35) antibodies from the surface of α1β3γ3 or α1β3γ3*
transfected cells at least partially was due to the relatively low avidity of these antibodies.
These antibodies, however, exhibited the same precipitation efficiency for α1β3γ3 and
α1β3γ3* receptors (experiments not shown), indicating that the difference in receptors
precipitated reflected the difference in receptors expressed on the surface of α1β3γ3 or
α1β3γ3* transfected cells. Since cell surface labeling, extraction and precipitation of
receptors as well as [3H]muscimol binding assays were performed in the same experiments
and under exactly identical conditions, these data indicate that [3H]muscimol binding to γ3*
containing receptors represented only about 31% of that to γ3 containing receptors.
Considering the variability of the methods used, these data were in good agreement with
results from Western blot experiments and suggested the formation of small amounts of
receptors composed of α1β3γ3* subunits on the cell surface of appropriately transfected cells.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
15
The electrophysiological properties of α1β3γ3* receptors formed could not be
investigated because it was not possible to unequivocally separate these low abundance
receptors from the large surplus of functional α1β3 receptors2.
Secondary structure prediction for amino acid sequence γ3(70-84)To investigate the
secondary structure of the amino acid sequence γ3(70-84), a neural network-based prediction
was performed using PHDsec (39, 40) for the sequence γ3(68-86). This procedure predicted a
low probability for the formation of an α-helix (below 1 in a probability scale [0,10]), a
higher probability (around 6) for the formation of a loop within the sequence γ3(68-72) and a
high probability (around 7) for the formation of an extended structure for the sequence γ3(73-
84) (data not shown). This prediction was essentially confirmed by results from force field
calculations within a Dynamic Monte Carlo optimization scheme. Low energy structures
exhibiting a potential energy of about 10 kcal/mol below random coil structures showed a
loop for residues γ3(70-76), and an extended structure along the residues γ3(77-84) that
potentially could form a β-sheet (data not shown). Overall, the predicted structure is probably
not sufficiently stable to form an autonomous structural element, but could gain induced
structural stability on interaction with intra- or intermolecular partner sequences.
Regions homologous to previously identified putative γ2 assembly sites are also
important for assembly of γ3 subunitsThe identification of γ3(70-84) as a sequence
important for assembly with α1 and β3 subunits does not exclude the presence of additional
sequences with similar importance in this subunit. It was thus investigated whether the
sequences γ3(86-93) and γ3(94-107) that are homologous to previously identified putative
assembly sites on γ2 subunits (Fig. 1A) (31), are also important for assembly of γ3 subunits.
Applying an approach previously used for the identification of the respective
sequences on the γ2 subunits, the sequences γ3(86-95) and γ3(94-107) were incorporated into
the truncated fragment α1(1-100) which per se was not able to bind to α1 or β3 subunits (13,
31) (Fig. 6A). In these experiments, γ3(86-93), one of the sequences homologous to the
previously identified γ2 sequences, was slightly elongated to γ3(86-95) to avoid loss of a
glycosylation site that possibly could be important for assembly (31). The resulting chimeras
Chim8 and Chim9 were then co-transfected into HEK cells with full length α1 or β3 subunits
and co-immunoprecipitated using α1(328-382) or β3(345-408) antibodies, respectively.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
16
As shown in Fig. 6B, Chim8 could be co-precipitated by the α1(328-382) antibody
after co-transfection with α1 subunits into HEK cells, indicating that the region γ3(86-95) is
sufficient to induce interaction of the chimeric α1(1-100) fragment with α1 subunits. Chim8
contains two glycosylation sites and migrated as two peptides with apparent molecular mass
14 and 16 kDa which represented the mono- and diglycosylated fragments, respectively. The
unglycosylated peptide with apparent molecular mass 12 kDa could only be detected on
extremly long exposures (experiments not shown). Interestingly, Chim8 on co-transfection
with β3 subunits, could also be co-precipitated with β3(345-408) antibodies. Under these
conditions only the diglycosylated form was co-precipitated, possibly indicating that β3
subunits preferentially interacted with the fully glycosylated fragment (Fig. 6C). These results
are similar to those obtained previously with the α1(1-100) chimera containing the
homologous γ2 sequence (31) and indicated that the sequence γ3(86-95) can induce binding to
α1 as well as β3 subunits (Fig. 6A). Unfortunately, migration of the protein bands in the two
blots (Fig. 6B+C) cannot be directly compared because the blots were taken from different
gels.
Similar to Chim8, Chim9 could be co-precipitated by α1(328-382) antibodies on co-
transfection with α1 subunits into HEK cells (Fig. 6B). In Chim9 the glycosylation site γ3N93
is lost (Fig. 6A) and, therefore, this chimera contains only a single glycosylation site. As
shown in Fig. 6B, Chim9 migrated as two peptides with apparent molecular mass 12 and 14
kDa which represented the unglycosylated and glycosylated fragments, respectively. In
contrast to Chim8, however, Chim9 could not be precipitated by β3(345-408) antibodies from
HEK cells co-transfected with Chim9 and β3 subunits (Fig. 6C). This indicated that the
sequence γ3(94-107) cannot induce interaction with β3 subunits and is in agreement with
previous results for the homologous sequence of the γ2 subunit (31). It was, therefore,
concluded that not only the sequence γ3(70-84), but also sequences homologous to those
previously identified on the γ2 subunit, are important for assembly of the γ3 with α1 or β3
subunits.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
17
DISCUSSION
Amino acid sequence γ3(70-84) is important for assembly with α1 and β3 subunitsIn
the present study, it was investigated, whether γ3 amino acid sequences homologous to
previously identified putative binding sites on the γ2 subunit are important for assembly with
α1 and β3 subunits. It was demonstrated that the C-terminally truncated γ3 fragment γ3(1-84)
could still bind to α1 or β3 subunits in appropriately co-transfected HEK cells, as indicated
by co-immunoprecipitation with subunit-specific antibodies, although this fragment lacked
the putative binding sites found in γ2 subunits. From this it was concluded that residues
important for assembly with α1 and β3 subunits were located within the first 84 amino acids
of the mature γ3 peptide.
Sites important for binding were then identified by incorporating various γ3 sequences
into the α1(1-68) fragment. This fragment is homologous to γ3(1-84) but in contrast to the
latter fragment could not bind to α1 or β3 subunits after co-expression in HEK cells. Results
indicated that replacement of the sequence α1(54-68) of the α1(1-68) fragment by the
sequence γ3(70-84) was sufficient to induce binding of the chimeric construct to full length
α1 or β3 subunits.
To investigate whether it is possible to define distinct amino acid residues in γ3(70-84)
important for binding to α1 or β3 subunits, several additional chimeras were constructed by
incorporating parts of the γ3(70-84) sequence into α1(1-68) fragments. None of these new
chimeras could interact with β3 subunits, indicating that the complete γ3(70-84) sequence
was necessary for inducing interaction with β3 subunits. Although the sequences γ3(70-75)
and γ3(70-78) were sufficient to induce weak binding of the chimeric α1(1-68) fragment to
the α1 subunit, strong binding to the α1 subunit again could only be demonstrated with the
complete γ3(70-84) sequence. This either indicated that only the complete γ3(70-84) sequence
can form the respective contact site to α1 or β3 subunits or that the complete sequence is
necessary for stabilizing other regions elsewhere in the protein that then can interact with the
adjacent subunits.
To investigate the importance of the γ3(70-84) sequence not only for interaction of
subunits, but also for assembly of α1β3γ3 receptors a chimeric γ3 subunit (γ3*) was
constructed in which the γ3(70-84) sequence was replaced by the homologous ρ1(92-106)
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
18
sequence. The γ3* construct was then co-expressed with α1 and β3 subunits in HEK cells.
Immunolabeling and quantification of receptors on the cell surface indicated a 70-80%
reduction of receptors containing α1, β3 and γ3* subunits. The additional observation that the
amount of specific [3H]muscimol binding sites as well as of α1 or β3 subunits on the cell
surface was comparable in α1β3γ3* or α1β3γ3 transfected cells suggested that due to the
reduced efficiency of assembly of γ3* subunits, in addition to α1β3γ3* receptors, receptors
composed of α1β3 subunits had been formed. These results again confirmed the importance
of the γ3(70-84) sequence for assembly of GABAA receptors. The remaining formation of
α1β3γ3* receptors on the cell surface can be explained by the existence of additional
assembly sites on the γ3 subunit that can partially compensate for the loss of one assembly
site caused by the absence of the γ3(70-84) sequence in γ3* subunits.
Amino acid residues contributing to the benzodiazepine binding site are located within
γ3(70-84) The benzodiazepine binding site of GABAA receptors is located at the interface
between α and γ subunits and several amino acid residues have been identified on these
subunits (α1H101, α1Y159, α1T206, α1Y209 and γ2F77, γ2M130, respectively) that
contribute to this site (34, 47). Due to the high sequence homology between γ2 and γ3
subunits (33) it can be assumed that these subunits exhibit a comparable structure and
comparable interactions with neighbouring subunits. In addition, it can be assumed that amino
acid residues homologous to those forming the benzodiazepine binding site in α1β3γ2
receptors will also form this site in α1β3γ3 receptors. Interestingly, the amino acid residue
γ3F80 which is located within the sequence γ3(70-84) important for binding to α1 or β3
subunits, is homologous to γ2F77 that contributes to the benzodiazepine binding site. It is thus
highly likely that this amino acid residue also contributes to the benzodiazepine binding
pocket of α1β3γ3 receptors and is located at the α1/γ3 interface.
A neural network based prediction using PHDsec as well as a Dynamic Monte Carlo
simulation, both performed in the present study, indicated that the sequence γ3(77-84) could
form a β-sheet. This prediction is not only supported by experiments using the substituted
cysteine accessibility method (48) for the homologous sequence γ2(74-81), but also by the
recently published crystal structure of the acetylcholine binding protein (AChBP), that can be
used as a model for the structure of the extracellular part of the GABAA receptor (49).
Altogether this supports the validity of the theoretical calculations performed in this study and
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
19
suggests that every second amino acid residue in this region might be involved in intra- or
intermolecular contacts, whereas the other residues are water accessible (48) and some of
them form part of the benzodiazepine binding pocket. In addition, the crystal structure of the
AChBP indicates that amino acid residue W53, that is homologous to γ3F80 is located at the
minus side of a dimer interface of the AChBP that corresponds with the α1/γ3 interface of the
GABAA receptors. Since γ3F80 might be solvent accessible like the homologous γ2F77
residue (48) and not be buried in the contact zone between the two subunits, nearby residues
might be involved in subunit/subunit interactions.
Interestingly, a second amino acid residue T45 of the AChBP homologous to γ3N72
contributes to the other side of the interface of the subunit, referred as plus side (49). The
observation that incorporation of the sequence γ3(70-73) into α1(1-68) induced a weak
binding to α1, but not to β3 subunits, then possibly indicates that amino acid residues within
this region might allow accommodation of the α1 subunit at the other interface of the γ3
subunit (plus side) when only α1 and γ3 subunits are present. Interestingly, the sequence
γ3(70-73) seems not to be important for binding of β3 subunits. This is surprising, because γ3
subunits should accommodate an α1 subunit at its minus and a β3 subunit at its plus side (13).
It is possible, however, that γ3 subunits use different contacts for interaction with α1 and β3
subunits at the plus side. A similar observation has been made previously for the α1 subunit,
in which distinct plus sides seem to be important for binding to β3 or γ2 subunits (13).
It has to be considered, however, that the AChBP forms homopentamers, while the
receptor studied in this work is composed of α1, β3 and γ3 subunits. It is, thus, possible that
the plus and minus sides of the GABAA receptor γ3 subunit are located slightly different from
those of the AChBP (49) and that the whole sequence γ3(70-84) is located at the α1/γ3
interface and exclusively contributes to the minus side. The observation that incorporation of
γ3(70-84) into the α1(1-68) fragment not only induced binding to α1 but also to β3 subunits
might then indicate that β3 subunits can interact with the minus side of γ3 subunits when no
α1 subunits are available. In the presence of both α1 and β3 subunits, however, the minus
side seems to exclusively assemble with α1 subunits because of energetic and/or steric
reasons. Alternatively, the sequence γ3(70-84) might stabilize an assembly site for β3
subunits at the plus side of γ3 subunits in the absence or presence of α1 subunits.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
20
At least two distinct γ3 sequences are important for assembly with α1 subunitsThe
identification of γ3(70-84) as a novel site important for assembly of GABAA receptors does
not exclude the existence of additional assembly sites on γ3 subunits. In other experiments it
was, therefore, investigated whether the sequences γ3(86-95) and γ3(94-107), that are
homologous to previously identified putative intersubunit contact sites on γ2 subunits (31),
are important for binding to α1 or β3 subunits. These sequences were incorporated into the
α1(1-100) fragment which per se could not bind to α1 or β3 subunits. Incorporation of the
sequence γ3(94-107) induced binding to α1 but not to β3 subunits, indicating that this
sequence facilitates a selective assembly with α1 subunits. These data are in agreement with
the finding that the homologous site on the γ2 subunit also is important for specific assembly
with α1 subunits (31). The observation that amino acid residues S75 and P77 of the AChBP
contribute to the minus side (that is homologous to the α/γ interface of GABAA receptors),
supports the conclusion that the corresponding GABAA receptor residues γ3S103 and
γ3M105 are located at the α1/γ3 interface. The sequence γ3(94-107), or part of it, might thus
be directly involved in forming the α1/γ3 interface.
Incorporation of the sequence γ3(86-95) into the α1(1-100) fragment induced binding
to α1 as well as β3 subunits. These results again are consistent with results obtained for the
homologous sequence on the γ2 subunit (31). Based on the AChBP structure, the γ3(86-95)
region does not appear to be in a position to directly interact with the α1 or β3 subunits, but
this sequence contains several amino acid residues highly conserved in the whole receptor
superfamily (50). Such residues might be important for stabilizing the fold of the subunit and
thus, might stabilize the conformation of assembly sites elsewhere in the protein.
Thus, interestingly, at least two amino acid sequences (γ3(70-84 and γ3(94-107)) have
been identified on the γ3 subunit in the present study that seem to be located at the same
interface (49) and either can directly bind to α1 subunits or stabilize such binding sites. This
is the first time that a GABAA receptor subunit has been shown to contain more than one N-
terminal sequence important for contacts with the same neighbouring subunit during assembly
of the receptor. The formation of multiple intersubunit contacts in this receptor superfamily is
supported by the crystal structure of the AChBP that indicates that there are seven
topologically distinct regions on a subunit forming contacts with the same adjacent subunit
(49). Further experiments will have to be performed to clarify whether the sequences
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
21
investigated in the present study are only important for primary contacts of subunits or also
for contacts in the completely assembled receptor.
In different subunits homologous regions might be important for assemblyAlthough
several different putative assembly sites have been identified so far on GABAA receptor
subunits (13, 29-31, 50), it is not known whether in different subunits homologous sequences
are important for interaction with neighbouring subunits. Here, it was demonstrated that
sequences γ3(86-95) and γ3(94-107) that are homologous to previously identified sequences
important for assembly of γ2 subunits, are also important for assembly of γ3 subunits. In
addition, we have demonstrated that the γ2 sequence homologous to the newly identified
γ3(70-84) sequence seems also to be important for assembly of γ2 subunits3. These data for
the first time demonstrate that in different subunits of the same subunit class homologous
sequences are important for assembly.
Interestingly, it has been demonstrated previously that α1(58-67), a sequence
homologous to part of the newly identified γ3(70-84) sequence (see Fig. 6A for sequence
comparison), mediated binding to β3 subunits (30). This finding confirms the importance of
the newly identified sequence γ3(70-84) for assembly and supports the conclusion that even
in subunits belonging to different subunit classes homologous sequences are important for
intersubunit contacts. It is highly likely, however, that different combinations of subunits
exhibit slightly different interactions (13), ensuring binding selectivity and a fixed receptor
stoichiometry.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
22
AcknowledgmentsWe want to thank Dr. Roman Furtmüller for preliminary
electrophysiological experiments, Susanne Karall and Elisabeth Dögl for excellent technical
support and Nahid Fathi for cloning the full length γ3 subunit.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
23
REFERENCES
1. Macdonald, R. L., and Olsen, R. W. (1994) Annu. Rev. Neurosci. 17, 569-602
2. Sieghart, W. (1995) Pharmacol. Rev. 47, 181-234
3. Nayeem, N., Green, T. P., Martin, I. L., and Barnard, E. A. (1994) J. Neurochem. 62,
815-818
4. Tretter, V., Ehya, N., Fuchs, K., and Sieghart, W. (1997) J. Neurosci. 17, 2728-2737
5. Barnard, E. A., Skolnick, P., Olsen, R. W., Möhler, H., Sieghart, W., Biggio, G.,
Braestrup, C., Bateson, A. N., and Langer, S. Z. (1998) Pharmacol. Rev. 50, 291-313
6. Bonnert, T. P., McKernan, R. M., Farrar, S., Le Bourdellès, B., Heavens, R. P., Smith, D.
W., Hewson, L., Rigby, M. R., Sirinathsinghji, D. J., Brown, N., Wafford, K. A., and
Whiting, P. J. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 9891-9896
7. Wisden, W., Laurie, D. J., Monyer, H., and Seeburg, P. H. (1992) J. Neurosci. 12, 1040-
1062
8. Fritschy, J.-M., Benke, D., Mertens, S., Oertel, W. H., Bachi, T., and Möhler, H. (1992)
Proc. Natl. Acad. Sci. U.S.A. 89, 6726-6730
9. Jechlinger, M., Pelz, R., Tretter, V., Klausberger, T., and Sieghart, W. (1998) J.
Neurosci. 18, 2449-2457
10. Nusser, Z., Sieghart, W., and Somogyi, P. (1998) J. Neurosci. 18, 1693-1703
11. Sieghart, W., Fuchs, K., Tretter, V., Ebert, V., Jechlinger, M., Höger, H., and Adamiker,
D. (1999) Neurochem. Int. 34, 379-385
12. Fritschy, J.-M., and Möhler, H. (1995) J. Comp. Neurol. 359, 159-194
13. Klausberger, T., Sarto, I., Ehya, N., Fuchs, K., Furtmüller, R., Mayer, B., Huck, S., and
Sieghart, W. (2001) J. Neurosci. 21, 9124-9133
14. Somogyi, P., Fritschy, J.-M., Benke, D., Roberts, J. D. B., and Sieghart, W. (1996)
Neuropharmacol. 35, 1425-1444
15. Im, W. B., Pregenzer, J. F., Binder, J. A., Dillon, G. H., and Alberts, G. L. (1995) J. Biol.
Chem. 270, 26063-26066
16. Chang, Y., Wang, R., Barot, S., and Weiss, D. S. (1996) J. Neurosci. 16, 5415-5424
17. Farrar, S. J., Whiting, P. J., Bonnert, T. P., and McKernan, R. M. (1999) J. Biol. Chem.
274, 10100-10104
18. Baumann, S. W., Baur, R., and Sigel, E. (2001) J. Biol. Chem. 276, 36275-36280
19. Green, W. N., and Millar, N. S. (1995) Trends Neurosci. 18, 280-287
20. Gelman, M. S., Chang, W., Thomas, D. Y., Bergeron, J. J. M., and Prives, J. M. (1995) J.
Biol. Chem. 270, 15085-15092
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
24
21. Connolly, C. N., Krishek, B. J., McDonald, B. J., Smart, T. G., and Moss, S. J. (1996) J.
Biol. Chem. 271, 89-96
22. Green, W. N., and Claudio, T. (1993) Cell 74, 57-69
23. Green, W. N. (1999) J. Gen. Physiol. 113, 163-170
24. Keller, S. H., and Taylor, P. (1999) J. Gen. Physiol. 113, 171-176
25. Klausberger, T., Ehya, N., Fuchs, K., Fuchs, T., Ebert, V., Sarto I., and Sieghart, W.
(2001) J. Biol. Chem. 276, 16024-16032
26. Verrall, S., and Hall, Z. W. (1992) Cell 68, 23-31
27. Kreienkamp, H. J., Maeda, R. K., Sine, S. M., and Taylor, P. (1995) Neuron 14, 635-644
28. Griffon, N., Büttner, C., Nicke, A., Kuhse, J., Schmalzing, G., and Betz, H. (1999)
EMBO J. 18, 4711-4721
29. Taylor, P.M , Thomas, P., Gorrie, G. H., Connolly, C. N., Smart, T. G., and Moss, S. J.
(1999) J. Neurosci. 19, 6360-6371
30. Taylor, P. M., Connolly, C. N., Kittler, J. T., Gorrie, G. H., Hosie, A., Smart, T. G., and
Moss, S. J. (2000) J. Neurosci. 20, 1297-1306
31. Klausberger, T., Fuchs, K., Mayer, B., Ehya, N., and Sieghart, W. (2000) J. Biol. Chem.
275, 8921-8928
32. Baer, K., Essrich, C., Benson, J. A., Benke, D., Bluethmann, H., Fritschy, J.-M. and
Lüscher, B. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 12860-12865
33. Herb, A., Wisden, W., Lüddens, H., Puia, G., Vicini, S., and Seeburg, P. H. (1992) Proc.
Natl. Acad. Sci. U.S.A. 89, 1433-1437
34. Sigel, E., and Buhr, A. (1997) Trends Pharmacol. Sci. 18, 425-429
35. Horton, R. M. (1993), Totowa, New Jersey: Humana Press, 251-261
36. Chen, C. A., and Okayama, H. (1988) Biotechniques 6, 632-638
37. Zezula, J., Slany, A., and Sieghart, W. (1996) Eur. J. Pharmacol. 301, 207-214
38. Sander, C., and Schneider, R. (1991) Proteins 9, 56-68
39. Rost, B., and Sander, C. (1993) Proc. Natl. Acad. Sci..U.S.A. 90, 7558-7562
40. Rost, B., and Sander, C. (1994) Proteins 19, 55-72
41. Kirkpatrick, S., Gelatt, C., and Vecchi, M. (1983) Science 220, 498-516
42. Nemethy, G., Gibson, K. D., Palmer, K. A., Yoon, C. N., Paterlini, G., and Zagari, A.
(1992) J. Phys. Chem. 96, 6472-6484
43. Mayer B., and Marconi G. (2000) J. Comput. Chem. 21, 170 – 181
44. Hackam, A. S., Wang, T. L., Guggino, W. B., and Cutting, G. R. (1997) NeuroReport 8,
1425-1430
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
25
45. Hackam, A. S., Wang, T. L., Guggino, W. B., and Cutting, G. R. (1997) J. Biol. Chem.
272, 13750-13757
46. Tögel, M., Mossier, B., Fuchs, K., and Sieghart, W. (1994) J. Biol. Chem. 269, 12993-
12998
47. Smith, G. B., and Olsen, R. W. (1994) J. Biol. Chem. 269, 20380-20387
48. Teissére, J. A., and Czajkowski, C. (2001) J. Neurosci. 21, 4977-4986
49. Brejc, K., van Dijk, W. J., Klaassen R. V., Schuurmans, M., van Der Oost, J., Smit, A. B.,
and Sixma, T. K. (2001) Nature 411, 269-276
50. Srinivasan, S., Nichols, C. J., Lawless, G. M., Olsen, R. W., and Tobin, A. J. (1999) J.
Biol. Chem. 274, 26633-26638
1 The abbreviations used are: GABA, γ-aminobutyric acid; nACh, nicotinic acetylcholine;
HEK, human embryonic kidney; PAGE, polyacrylamide gel electrophoresis; IP low buffer,
low salt immunoprecipitation buffer; PBS, phosphate buffered saline; DMEM, Dulbecco´s
modified Eagle medium; DMC, Dynamic Monte Carlo; AChBP, acetylcholine-binding
protein. 2 (R. Furtmüller, unpublished experiments) 3 (Sarto et al., in preparation)
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
26
FIGURE LEGENDS
FIG. 1. Co-immunoprecipitation of truncated γ3 fragments with full length α1 and β3
subunits. A, schematic drawing of C-terminal sequences of γ2 and γ3 fragments. The
sequence of the fragments is indicated by the amino acid numbers given in parentheses. 1
represents the first amino acid of the mature subunit. The sequences shown are indicated by
the numbers above the amino acid single letter code representing the position of the
respective amino acid residue. The previously identified putative γ2 assembly sites for α1 or
β3 subunits as well as the homologous sites on the γ3 subunit are boxed. + indicates co-
precipitation, and - indicates absence of co-precipitation of these constructs with full length
α1 or β3 subunits. B, in control experiments the truncated γ3 fragments were transfected into
HEK cells, extracted with Lubrol extraction buffer and immunoprecipitated using γ3(1-35)
antibodies. The precipitate was subjected to SDS-PAGE and the constructs were detected
using digoxygenized γ3(1-35) antibodies. The protein fragment γ3(1-116) exhibited an
apparent molecular mass of 14 and 16 kDa (the un- and monoglycosylated peptide,
respectively), whereas fragments γ3(1-84) and γ3(1-72) migrated as bands of 9.4 kDa and 8.2
kDa, respectively. C, HEK cells expressing the respective C-terminally truncated fragments
together with full length α1 subunits were extracted, proteins were precipitated with α1(328-
382) antibodies and subjected to Western blot analysis using digoxygenized γ3(1-35)
antibodies. D, HEK cells expressing the C-terminally truncated fragments together with full
length β3 subunits were extracted, proteins were precipitated with β3(345-408) antibodies
and subjected to Western blot analysis using digoxygenized γ3(1-35) antibodies. All
experiments were performed three times with comparable results.
FIG. 2. γ3(70-84) is important for binding to α1 and β3 subunits. A, C-terminal sequences
of the γ3(1-84) and α1(1-68) fragments as well as of different chimeras are shown
schematically. Amino acid sequences of the γ3 subunit are boxed. HEK cells were co-
transfected with these constructs together with full length α1 or β3 subunits, extracted
proteins were precipitated with α1(328-382) (designed as α1L-Ab) or β3(345-408) antibodies
(designed as β3L-Ab), respectively, and subjected to SDS-PAGE and Western blot analysis
using digoxygenized α1(1-9) antibodies (designed as α1N-Ab). A possible co-
immunoprecipitation was investigated as described under „Results“. + indicates co-
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
27
precipitation, and - indicates absence of co-precipitation of these constructs with the
respective full length subunits. B+C, Western blots demonstrating co-precipitation of
different chimeras with full length α1 or β3 subunits. All chimeric constructs contained a
single glycosylation site. The precipitated constructs, however, migrated predominantly as the
glycosylated form with an apparent molecular mass of 12 kDa. The protein with apparent
molecular mass 51 kDa in B represents the full length α1 subunit. All experiments were
performed three times with comparable results. D, Western blots demonstrating comparable
expression of chimeric fragments. HEK cells expressing the chimeric fragments alone were
extracted, proteins were precipitated with α1(1-9) antibodies (designed as α1N-Ab), and
subjected to SDS-PAGE and Western blot analysis using digoxygenized α1N-Ab antibodies.
FIG. 3. Co-immunoprecipitation of chimeric constructs with full length α1 or β3
subunits. A, the C-terminal sequences of Chim3, of the α1(1-68) fragment, as well as of
different chimeras are shown schematically. Amino acid sequences of the γ3 subunit are
boxed. HEK cells were co-transfected with these constructs together with full length α1 or β3
subunits, extracted proteins were precipitated with α1(328-382) or β3(345-408) antibodies,
respectively, and subjected to SDS-PAGE and Western blot analysis using digoxygenized
α1(1-9) antibodies. A possible co-immunoprecipitation was investigated as described under
„Results“. +++ indicates strong co-precipitation, + indicates weak co-precipitation, +/-
indicates very weak co-precipitation and - indicates absence of co-precipitation of these
constructs with the respective full length subunits. B, Western blots demonstrating co-
precipitation of chimeric constructs with full length α1 or β3 subunits. All chimeric
constructs contained a single glycosylation site, but all co-precipitated constructs migrated
predominantly as the glycosylated form with an apparent molecular mass of 12 kDa. All
experiments were performed three times with comparable results.
FIG. 4. Western blot analysis of GABAA receptors labeled on the surface of HEK cells
using γ3(1-35) antibodies. A, In control experiments, HEK cells were co-transfected with
α1, β3, and γ3* or α1, β3, and γ3 subunits. The γ3* chimeric construct was a full length γ3
subunit in which the sequence γ3(70-84) was replaced by the homologous sequence ρ1(92-
106) of the ρ1 subunit. Proteins were extracted from cells, immunoprecipitated by γ3(1-35)
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
28
antibodies and subjected to Western blot analysis using digoxygenized γ3(1-35) antibodies. γ3
subunits as well as γ3* subunits were expressed to a comparable extent and exhibited an
apparent molecular mass of 43 kDa. The proteins with apparent molecular mass of about 51
kDa were labeled due to a cross-reactivity of the antibody with HEK cell proteins. B, HEK
cells were co-transfected with α1, β3, and γ3*, or with α1, β3, and γ3 subunits. GABAA
receptors expressed on the surface were immunolabeled by an incubation of intact cells with
γ3(1-35) antibodies, and were then extracted, precipitated by Immunoprecipitin and subjected
to SDS-PAGE and Western blot analysis using digoxygenized γ3(1-35) antibodies. Blots were
then stripped and re-analyzed with digoxygenized α1(1-9) antibodies. Then blots again were
stripped and re-analyzed with digoxygenized β3(1-13) antibodies. The experiment was
performed twice with comparable results. α1, β3, and γ3 subunits migrated as protein bands
of 51 kDa, 54 kDa and 43 kDa, respectively.
FIG. 5. Western blot analysis of GABAA receptors labeled on the surface of HEK cells
using α1(1-9) antibodies. HEK cells were co-transfected with α1, β3, and γ3*, or with α1,
β3, and γ3 subunits. GABAA receptors expressed on the surface were immunolabeled by an
incubation of intact cells with α1(1-9) antibodies, and were then extracted, precipitated by
Immunoprecipitin and subjected to SDS-PAGE and Western blot analysis using
digoxygenized γ3(1-35) antibodies. Blots were then stripped and re-analyzed with
digoxygenized α1(1-9) antibodies. Then blots again were stripped and re-analyzed with
digoxygenized β3(1-13) antibodies. Comparable results were obtained when the order of
detection of subunits was changed. These experiments were performed six times with
comparable results. α1, β3, and γ3 subunits migrated as protein bands of 51 kDa, 54 kDa and
43 kDa, respectively. The average amount of γ3* subunits detected with γ3(1-35) antibodies
in the α1-precipitate of receptors present on the cell surface was 22±4% (n=6) of γ3 subunits.
The intensity of γ3* subunit staining shown on this blot represents 15% of γ3 subunit staining.
FIG. 6. Amino acid sequences γ3(86-95) and γ3(94-107) are also important for assembly
of γ3 subunits. A, The C-terminal sequences of the α1(1-100) and γ3(1-116) fragments as
well as of different chimeras are represented schematically. The position of amino acid
residues is given by numbers on top of the single letter code and amino acid sequences of the
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
29
γ3 subunit are boxed. HEK cells were co-transfected with these constructs together with full
length α1 or β3 subunits, extracted proteins were precipitated with α1(328-382) or β3(345-
408) antibodies, respectively, and subjected to SDS-PAGE and Western blot analysis using
digoxygenized α1(1-9) antibodies. A possible co-immunoprecipitation was investigated as
described under „Results“. + indicates co-precipitation, and - indicates absence of co-
precipitation of these constructs with the respective full length subunits. B, Western blots
demonstrating co-precipitation of chimeric constructs with full length α1 subunits. C,
Western blots demonstrating co-precipitation of chimeric constructs with full length β3
subunits. Migration of the protein bands in Fig. 6B+C cannot be directly compared because
the blots were taken from different gels. All experiments were performed three times with
similar results.
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Werner SieghartIsabella Sarto, Thomas Klausberger, Noosha Ehya, Bernd Mayer, Karoline Fuchs and
3 subunits important for assembly of GABAA receptorsγA novel site on
published online June 13, 2002J. Biol. Chem.
10.1074/jbc.M203597200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on April 16, 2018
http://ww
w.jbc.org/
Dow
nloaded from