2 interactions
with the carboxyl terminus of Kir3.4 inward rectifier K+ channel
subunitsExamination of agonist interactions with unnatural nicotinic acetylcholine receptor mutants
4. An Unnatural Amino Acid for Site-Specific Labelling of Membrane Proteins
6. The role of leucine residues in the M2 domain of the AChR
8. Alteration of a highly conserved vicinal disulfide in the nicotinic acetylcholine receptor.
9. Expression of the extracellular domain of the nicotinic acetylcholine receptor
10. Examination of Nicotinic Acetylcholine Receptors in Neuronal Cell Lines
Examination of agonist interactions with unnatural nicotinic acetylcholine receptor mutants
3. An Unnatural Amino Acid for Site-Specific Labelling of Membrane Proteins
6. Alteration of a highly conserved vicinal disulfide in the nicotinic acetylcholine receptor
7. Investigating a conserved pore-region residue in potassium channels with unnatural amino acids
1. A Multi-Substrate Single-File Model for Ion-Coupled Transporters
4. Ion Binding and Permeation at the GABA Transporter GAT1
1. Investigating a conserved pore-region residue in potassium channels with unnatural amino acids
2. Functional analysis of the weaver mutant GIRK2 K+ channel and rescue of weaver granule cells
4. Determining the subunit stoichiometry and arrangement of heteromultimeric potassium channels
6. Integrin Interactions with GIRK channels
7. Antagonistic modulation of an inward rectifier
K+ (GIRK) channel by 
and ß subunits of
G proteins
8. Adenovirus-mediated expression of K+ channels in primary and secondary cell cultures
1. Adenovirus-mediated expression of K+ channels in primary and secondary cell cultures
1. Functional analysis of the weaver mutant GIRK2 K+ channel and rescue of weaver granule cells
Mark .W. Nowak, Pat C. Kearney, Wenge Zhong, Scott K. Silverman, Cesar G. Labarca, Margaret E. Saks, Jeff R. Sampson, John Abelson and Dennis A. Dougherty
We have previously demonstrated the ability to introduce unnatural amino
acids into ion channels employing the stop codon suppression method in Xenopus oocytes.
Oocytes are coinjected with cRNA containing a UAG codon at the position of
interest and a suppressor tRNA with the corresponding anticodon, CUA, and chemically
acylated with the unnatural amino acid. We examined the interaction of acetylcholine
(ACh) with conserved tyrosine residues in the subunit
at positions 93, 190 and 198 of the mouse nicotinic acetylcholine receptor
(nAChR) by introducing derivatives of tyrosine and phenylalanine
(Nowak et al., Science 268, 439,1995).
Our initial study revealed the 4' OH group of tyrosine residue 93
functions as a hydrogen bond donor. We have further demonstrated that the position
of this OH group is also critical for wildtype function. Shifting the position
of the OH group at position 93
by inserting 3-OH-phenylalanine or homo-tyrosine (which extends the sidechain
one
bond length from the peptide backbone) increases the EC50 8-10-fold.
We have extended the stop codon suppression method to examine the interaction
of other agonists with nAChR. To further understand the interaction of agonists
with nAChR we compared the dose-response relationships for ACh to that for
the simple quaternary ammonium tetramethylammonium (TMA) and nicotine (NIC).
Since the millimolar TMA and NIC concentrations required for activation of
the unnatural mutant nAChRs also block the channel we introduced leucine to
serine mutations at the 9' position within the pore region of the , ß and subunits.
These mutations shift the TMA activation of nAChR toward lower agonist concentrations
(Labarca et al., Nature 376, 514, 1995) where channel block does
not occur.
We examined the activation by ACh, nicotine and TMA of AChR containing unnatural
mutations at positions 93
and 198
and Leu9'Ser mutations. EC50 values (mM) obtained from the dose-response relationships
are summarized below.
| Site | Residue | ACh | TMA | Nicotine |
| 93¹ |
Tyr | 0.14±0.03 | 21.2±4 | 8.8±0.6 |
| 3-F-Tyr | 0.077 | 14.1±2.2 | 7 | |
| 4-MeO-Tyr | 2.2±0.9 | 187±18 | 52 | |
| Phe | Phe 5.9±0.1 | 478±71 | 94±8 | |
| 198² |
Tyr | 0.38 | 83 | 27 |
| 4-MeO-Phe | 0.78 | 136 | 50 | |
| Phe | 2.9 | 407±34 | 143 | |
| p-F-Phe | 1.5 | 360 | 128 |
¹ßL9'S, L9'S
mutants
²( L9'S)2 mutants
TMA and NIC activation of the mutant nAChRS relative to the wildtype receptor
showed a similar pattern as observed for ACh activation suggesting that all
three agonists interact in a similar manner with residues 93
and 198.
These findings also indicate that an alkyl cation moiety, the common structural
feature for ACh, TMA and NIC, is a strong determining factor in agonist association
with tyrosine residues at positions 93
and 198.
Patrick C. Kearney, Haiyun Zhang, Wenge Zhong, Dennis A. Dougherty
A nonsense suppression method was employed to incorporate a total of 10 natural
and unnatural residues into the 9' position of the M2 region into the ß, ,
and 
subunits
of muscle nicotinic receptors. Many unnatural side chains produce functional
receptors at this position. The data allow for 33 pairwise comparisons of functional
properties as influenced by structural properties such as side chain chirality,
polarity, and branching Subtle changes such as the chirality of the side chain
(isoleucine vs allo-isoleucine) produce marked changes in gating kinetics,
leading to changes in the EC50 for ACh. These changes are themselves
modified by other subunits in a way that suggests physical interactions and
an ordering of the subunits. The data point to the importance of polar interactions
at the 9' position.
Wenge Zhong, Dennis A. Dougherty
An acidic residue, Asp 180, in the subunit
is near the binding site in the nicotinic acetylcholine receptor (nAChR) (1).
Single-channel analysis suggested that a homologous residue in the 
subunit,
Asp 175, is mainly involved in gating (2). We conducted studies on the relationships
between Asp
180, the homologous Asp
174, and the leucines at the 9' site in the M2 region of the several subunits
in the nAChR. By mutating Asp to Asn in the and subunits
and Leu to Ser at 9' in each subunit, we were able to generate 15 mutant receptors.
Electrophysiological studies provided a table of EC50 values ( µM)
shown below.
| D180N |
D174N |
D180N, D174N |
|
| wt 49 | >267 | >207 | >360 |
| ßL9'S 1.22 | 76 | 61 | >307 |
| L9'S
4.2 |
31 | 76 | >244 |
| L9'S
1.0 |
15.7 | 19.6 | 137 |
| L9'S
0.4 |
3.1 | 7.9 | 47 |
The data show that Asp to Asn mutation increases the EC50 values
while the Leu to Ser mutations decrease EC50 values. However, this
opposing effect is not subtractive, which suggests interactions between these
residues. Because the binding site and M2 regions are thought to be separated
by ~ 50 Å, these interactions must be remote and/or indirect. Furthermore ßL9'
interacts with D180
and D174
most strongly, for the D180N
and D174N
mutations suppress the decrease in EC 50 caused by ßL9'S mutations.
When comparing L9'S
and L9'S
series (third and fourth rows in the table) with the L9'S
series, linear relationships with slopes of ca. 1 were observed, suggesting
a similar mode of remote interaction between the acidic residues and the 9'
leucines of the , ,
and subunits.
On the other hand, the ßL9'S series shows a poorer correlation, implying
that different mode (s) of remote relations may be involved. Taken together,
the data suggest (a) that the acidic residues near
the binding sites in both and subunits
are interacting indirectly with the 9' Leucines in M2 in the pore region of the
nAChR, and (b) that the ß subunit may play a special role in the binding/gating
conformational change of the receptor. Further studies using single-channel recording
are underway.
References
(1) Karlin, A., et al., J. Biol. Chem., 270, 3160-3164, 1995
(2) Zhang, Y., et al., Meeting abstract, Society for Neuroscience, 21, p1582, 1995
Justin Gallivan , Dennis A. Dougherty
We have used the nonsense suppression technique to incorporate biocytin into nicotinic acetylcholine receptors at two positions in the main immunogenic region. Binding of [125I]streptavidin to intact cells shows that the biocytin is exposed at one position, but not at the other. In chemical work over the next year, this reagent will be improved with various chain lengths; and attempts will be made to detect incorporation with fluorescent streptavidin. This reagent shows promise for monitoring side chains that become exposed to external reagents during channel function. TAG mutations will be generated at several codons of interest; and fluorescence will be studied as the channel is activated in oocytes.
Margaret E. Saks, Jeffrey R. Sampson, Mark W. Nowak, Patrick C. Kearney, Fangyong Du, John N. Abelson and Dennis A. Dougherty
By studying positions in the nAChR subunit
that were more tolerant with regard to substitution, we demonstrated that our
original suppressor tRNA, MN3, led to incorporation of natural amino
acids at the mutation site along with the desired unnatural amino acid. This
is probably due to reacylation (and/or editing) of the chemically acylated
tRNA-MN3
by synthetases that are endogenous to the Xenopus oocyte. A new tRNA,
THG73, has been designed and evaluated as a vehicle for incorporating unnatural
amino acids site-specifically into proteins expressed in vivo using the
stop codon suppression technique. The construct is a modification of tRNAGln(CUA)
from Tetrahymena theromophila, which naturally recognizes the stop codon
UAG. Using electrophysiological studies of mutations at several sites of the
nicotinic acetylcholine receptor, we have established that THG73 represents a
major improvement over previous nonsense suppressors both in terms of efficiency
and fidelity of unnatural amino acid incorporation. Compared to tRNA-MN3, THG73
is at least 100-fold less likely to be acylated by endogenous synthetases of
the Xenopus oocyte. This effectively eliminates a major concern of the in
vivo suppression methodology - the undesirable incorporation of natural amino
acids at the suppression site. In addition, THG73 is 5- to 10-fold more efficient
at incorporating unnatural amino acids in the present system. Taken together,
these two advances should greatly expand the range of applicability of
the in vivo nonsense suppression methodology.
Pamela M. England, Dennis A. Dougherty
The recent demonstration that unnatural amino acids can be incorporated into ion channels expressed in Xenopus oocytes suggests a means of manipulating receptors/ion channels in vivo in novel and informative ways.
We have designed a photoreactive unnatural amino acid (o-nitrophenylglycine,
NPG) that, after incorporation into a protein, results in sequence-specific
peptide backbone cleavage of the protein upon irradiation with ultraviolet
light (350 nm). As an initial test of the efficiency of this residue, we have
incorporated NPG into the amino-terminal region (Leu47NPG) of the Shaker B
potassium channel and expressed these mutant receptors in Xenopus oocytes.
The Shaker B channel is voltage dependent and rapidly inactivating.
The inactivation is mediated by a ~20 amino acid region at the amino terminus.
Truncated (
6-46)Shaker B
mutants do not exhibit rapid inactivation (ShakerBIR). Non-photolyzed oocytes
displayed wild type Shaker B currents (rapid inactivation). On the other
hand, photolyzed
oocytes displayed Shaker B currents in which the inactivation was removed
(ShakerBIR). The ShakerBIR currents are a direct result of generating
the Shaker B 1-47
mutant from the full length Shaker B Leu47NPG mutant in vivo, demonstrating
the ability of the unnatural amino acid NPG to provide peptide backbone cleavage
of a functional
ion channel expressed in Xenopus oocytes. The ability to generate sequence-specific
single cuts of the peptide backbone in functional
receptors/ion channels in vivo, combined with electrophysiology, will
allow us 1) to continue to probe structure function relationships and 2) to resolve
the time course of signalling events associated with these proteins.
Purnima Deshpande, Haiyun Zhang, and Cesar Labarca
We have continued our studies of the gating mechanism of the acetylcholine receptor (AChR) by site-directed mutagenesis of amino acid residues in the M2 transmembrane region. We have previously shown that leucine to serine mutations at the 9' position of M2 shift the dose-response relation for ACh to the left by about ten-fold for every mutated subunit; a receptor with four mutated subunits is 10,000 times more sensitive to ACh than the wild type. Each of the five 9' leucine residues seems to participate independently and symmetrically in the gating process. Spontaneous channel openings, which are very rare in the wild type receptor, increase with the number of mutated subunits and become very frequent, up to about 50% of the maximum current, in receptors with four or five mutations. In addition, AChR antagonists like curare or hexamethonium become agonists and are more effective in receptors with four mutated subunits than with two. Likewise, a weak agonist like tetramethylammonium becomes much more effective in receptors with mutated subunits.
In addition to leucine at position 9' there are several other hydrophobic
residues in M2, most of them leucines, that are conserved in all four subunits.
From our previous work and the work of others, it is reasonably well established
that residues 2', 6' and 10' face the channel in the open state. A top view
of the M2 helixes
with residues 2',6', and 10' facing the channel shows that most of the hydrophobic
residues (3', 7', 13', 16', 17') face the neighboring M2 helixes,
while 8' and 15' are on the distal side of M2 with respect to the channel,
about 180 degrees from 10' and 6' respectively.
Leucine to serine mutations at position 16' in the , ß and subunits
affects gating in a way similar to mutations at position 9', shifting the dose-response
to agonist to the left, but the effect is not multiplicative as with position
9'. The residue at position 16' in the subunit
is phenylalanine; mutation of this residue to serine shifts the dose-response
to the right and greatly increases the rate of desensitization. Leucine to
serine mutations (valine to serine in the subunit)
at position 17' shift the dose-response relation to the left in all subunits.
L17'S in the subunit
is particularly effective, shifting the response by two orders of magnitude.
The
effect is additive for + ß and
for + but
not for and ß + .
Residues at positions 8' and 15' are positioned away from the channel. Leucine
to serine mutation at position 8' in the subunit
shifts the dose-response to agonist to the right about 3-fold. At position
15 the same mutation produces a shift to the left of about 3-fold, much less
than at position 9', 16' or 17'. Analysis of other hydrophobic residue mutations
is in progress. A simple structural explanation for these results so far is
that most of the M2 leucines--and in some cases valines-old the M2s together
around
the pore by forming some kind of hydrophobic bonds between helices.
Disruption of these bonds by replacing leucine with a polar residue probably
makes the ring of M2s more relaxed and easier to open by agonists. Competitive
antagonists probably produce a small change in conformation at the binding site,
insufficient to gate the wild type receptor, but enough to open the more relaxed
ring of M2s in the mutated receptors. Leucine 9' seems to play a special role
in channel gating; perhaps this residue is very near to the channel gate or forms
the gate itself.
Haiyun Zhang, Purnima Deshpande, Cesar Labarca
Structural data for the nAChR suggest that a Leucine residue at the approximate
midpoint of the M2 transmembrane domain (the 9' position) occupies a "kink" in
each of the 5 M2 helices ( , , ß, , )
and points into the closed channel. In the open channel configuration, the
M2 helices rotate so that Leu9' side chains no longer occlude the conduction
pathway. We expressed mouse nAChRs with varying numbers (m*s) of
Leu9'Ser mutated subunits in Xenopus oocytes. The dose-response relation
shifted to the left by about 10-fold for each mutation, so that a nAChR with
four Leu9'Ser mutations is 104-fold more sensitive than wild
type.
In detailed single-channel studies of the Leu9'Ser receptors from m*s =
0 to m*s = 5 ( 2ß *, 2*ß , 2*ß *, 2*ß* *
and 2*ß* * * receptors
were tested), we found that mutated AChRs have (1) longer openings and bursts,
(2) briefer closed times, and (3) more frequent spontaneous openings. These
effects increase with m*s. For example, average channel open times
for wild type 2ß and
for 2*ß* * (m*s =
4) receptors are 1.8±0.3 ms and 40±5 ms, respectively. The spontaneous
openings, however, have the average open time of 0.24±0.02 ms for all
m*s > 0, while the spontaneous opening frequency increases about
104 fold from ~1e-5 for m*s = 1,2 to ~0.1 for m*s = 4,5.
In addition, single channel current amplitudes differ between wild type (6.2+0.3
pA at -100 mV in symmetrical 150 mM KCl solutions) and all m*s > 0
(8.0+0.6 pA). Because the nAChR opens more readily at more hyperpolarized membrane
potentials, we measured the relaxation time constant td of voltage jumps from
positive potentials (+40 or +60 mV) to -100 mV for the Leu9'Ser receptors.
td increases from 2.5±0.6 ms for wild type to 440+30
ms for m*s = 1 and 2, but does not increase with additional mutations.
These data suggest that both longer and more frequent openings contribute to
the >104 fold shift in EC50 between wild type and m*s = 5. A simple
mechanistic interpretation is that the Leucine to Serine mutations at the 9'
position both stabilize the open state and destabilize the closed state, and
the open state becomes more conductive.
Synthesized postsynaptic currents, produced with a piezoelectric micromanipulator
that delivered ACh pulses (>2*EC50), decayed as expected from
channel burst durations: for 2*ß* *
(m*s = 4) and 2*.ß* *
(m*s
= 2), the decay time constants are, respectively, 2,200±500 ms and 170±20
ms, >100 and >7 fold higher than for wild type. Therefore, the highly conserved
9' Leucine is crucial for appropriately brief synaptic events. There are now
several examples of inherited human diseases based on mutations in AChR M2 regions
that probably lead to excessively long channel openings.
Scott K. Silverman, Dennis A. Dougherty
Near the subunit
binding site of the nicotinic acetylcholine receptor (nAChR) is a highly conserved
vicinal disulfide, formed from the cys192 and cys193 side chains. These residues
are conserved in a variety of related receptor subunits. The disulfide may
be reversibly reduced with suitable reagents, and only when in the oxidized
(-S-S-) form does the channel bind its agonist and gate normally. The vicinal
disulfide forms an eight-membered ring, which has certain conformational preferences
that will change if the ring size is altered. We therefore substituted for
the cysteine at position 192 the one-carbon-longer homocysteine, using the
nonsense codon suppression method. This alteration expands the ring size from
eight to nine and should alter any conformational preferences of the natural
receptor. We observe that even with the homocysteine mutation, the channel
is no longer as easily affected by reduction, as expected for the larger ring
size, yet is still activated by acetylcholine with approximately the same sensitivity
as wild type. These observations appear to rule out a "conformational
switch" role for the vicinal disulfide in the function of the nAChR.
Anthony West, Dennis Dougherty
The subunits
of nicotinic acetylcholine receptors have an approximately 210 amino acid extracellular
N-terminal domain which contains the agonist-binding site. The secondary and
tertiary structure of this domain, as well as of the whole receptor, are unknown,
although Unwin has obtained a 7.5 Å resolution structure of the receptor
using electron microscopy. In hopes of obtaining high-resolution structural
information, we have been exploring various strategies for obtaining the extracellular
domain in a pure, soluble, and correctly folded state. We have been focusing
on the mouse muscle and
chicken neuronal -7
subunits After several unsuccessful attempts, we have found a method for producing
these truncated subunits
in folded state: fusion of the extracellular
domain to a glycosyl-phosphatidylinositol anchorage sequence. This causes the
protein to be produced in a lipid-linked form which can also be selectively
cleaved from the cell surface. After pilot experiments in Xenopus oocytes,
we have scaled up production by making a stably transfected CHO cell line.
By growing these cells on a large scale, we have been able to produce the extracellular
domain on a 100 µg - 1 mg scale.
Affinity purification of this material with an antibody column allows us
to obtain the protein in a relatively pure state, suitable for various biochemical
studies. Like the native receptor, this purified protein is glycosylated, has
had its N-terminal signal peptide removed, and binds -bungarotoxin
with a high affinity. Circular dichroism studies are underway to access the
secondary structure of this domain, and to compare these data to previous studies
on the whole receptor.
Mark W. Nowak, Melihat Fidan-Nowak and Yinong Zhang
To date, 8 and
3 ß neuronal nicotinic acetylcholine receptor (nAChR) subunits have been
identified. While heterologous expression of single and/or multiple subunits
in Xenopus oocytes results in functional channels, we do not fully understand
the native composition or functional role of the nAChRs present in the brain.
Expression of neuronal nAChRs in oocytes is not a suitable system for studying
the native receptors since 1) numerous functional receptors can be expressed
by coinjection
of the 8 and
3 ß nAChR subunits, 2) the properties of the expressed receptors do not
necessarily resemble those of native receptors and 3) coinjection of single and ß subunits
can result in expression of multiple receptors. For example, injection of oocytes
with 3
and ß4 mRNAs in mole ratios of 0.09:0.91, 0.5:0.5 and 0.91:0.09 yields
receptors with
EC50 values of 30 µM, 40 µM and 150 µM, respectively.
We are developing a system for examining the native composition and properties of neuronal nAChRs employing gene transfer techniques such as transfection and adenovirus-mediated infection. Specifically, we will introduce into neuronal cell lines and primary neuronal cultures antisense probes or dominant negative mutants (to knockout nAChRs) or dominant positive mutants (to alter nAChRs). We will first design antisense or mutant constructs and test them in an oocyte expression system. Following initial testing in oocytes, the antisense probes or mutants will be introduced into neuronal cell lines or primary neuronal cultures containing nAChRs. The pharmacological and electrophysiological properties of the cells before and after transfection will be characterized.
Dominant mutations were constructed in the neuronal AChR subunits by inserting
a stop codon before the start of the second transmembrane domain (-2'TAG).
The mutant proteins contain the extracellular N-terminal domain and the first
transmembrane domain. Coinjection of oocytes with wildtype 3:ß4
mRNAs (which results in expression of a functional receptor) and ß4 -2'TAG
mutant mRNA decreased receptor expression. Injection of mutant ß4 mRNA
containing a stop codon in the N-terminus before the start of the first transmembrane
domadin (N-Term. TAG) did not effect expression of 3:ß4
receptors. Similar results were observed for the 3
-2'TAG and 3
N-Term. TAG mutants. The -2'TAG dominant mutants will be useful for knocking
out nAChRs in neuronal cells. We are currently making the -2'TAG mutants in
the
other neuronal subunits.
Dominant positive mutations have been made in which a leucine (Leu) residue
at position 9' within the pore was changed to serine (Ser). Muscle nAChRs containing
Leu9'Ser mutations (see abstract #xxx) display increased agonist sensitivity
and spontaneous channel activity. Oocytes injected with 3L9'S:ß4
expressed receptors with an EC50 for acetylcholine of 0.1-0.2 µM (wildtype
EC50 = 40 µM), a 400-fold increase in agonist sensitivity. In addition,
the leak currents of oocytes expressing the mutant receptors were > 0.5 µA
compared to 50-100 nA for oocytes expressing wildtype receptor. The large leak
currents indicate spontaneous receptor activity.
Following the testing of the antisense probes, dominant negative and dominant
positive mutants in Xenopus oocytes, mammalian constructs will be made
for transfection into neuronal cell lines. Initially, we will examine the nAChRs
in rat pheochromocytoma cells (PC12) which express 3, 5, ß2, ß3
and ß4 subunits and at least three different functional nAChRs. To determine
the composition of the nAChRs in PC12 cells we will knockout receptors by transfecting
with either the antisense probes or dominant negative mutants. To alter the
electrophysiological properties of the native receptors the PC12 cells will
be transfected with the dominant positive mutants.
Jun Li
Cyclic nucleotide-gated (CNG) channels are allosterically activated by the binding of cAMP or cGMP. We are interested in the structural basis of binding, and of subsequent conformational changes (known as "gating") leading to the opening of the ion-permeating pore. Previous studies have identified amino acid residues involved in the initial "docking" of the ligand¹, or in the "induced fit" of the ligand, which, on one hand, strengthens the binding by forming stronger interactions, and on the other hand, elicits further changes in protein structure². Most of the site-directed mutagenesis studies have been focused on hydrogen bonds or electrostatic interactions. Since adenosine and guanosine are both aromatic compounds, we decided to investigate the role of aromatic-aromatic interactions in binding and gating. We have made a series of "alanine-scanning" mutations on each of the aromatic residues in the putative cyclic nucleotide binding domain. Some of these mutants did not generate functional channels when expressed in Xenopus oocytes, while others, although expressed, showed less than 2-fold alterations in cAMP or cGMP sensitivity from the wild type values. However, one mutant, Y565A of the rat olfactory channel a subunit, displayed dramatically higher apparent sensitivity toward both cAMP and cGMP. Compared to the wild type channels, the EC50 of cAMP is lowered by 10 fold (from 80 µM to 8 µM); the EC50 of cGMP is lowered by 6-7 fold (from 2.6 µM to 0.4 µM). In structural models of this binding site, Y565 is located away from the immediate protein-ligand interface, unlikely to directly contact the ligand. One interpretation is that this mutation alters the intrinsic gating barrier of the channel, making the channel find it easier to be activated by binding to a suitable agonist. We are at present testing this hypothesis, mostly by observing the effect of this mutation on the nickel inhibition of channel activity. Since nickel inhibition is known to be primarily a gating effect, Y565A is expected to show less inhibition by nickel.
1. Altenhofen et al. (1991) PNAS 88: 9868-9872
2. Varnum et al. (1995) Neuron 15: 629-625
Alyce Su, Sela Mager, Stephen L. Mayo
Ion-coupled transporters are simulated by a model that differs from contemporary alternating-access schemes. Beginning with concepts derived from multi-ion pores, the model assumes that substrates (both inorganic ions and small organic molecules) hop (a) between the solutions and binding sites and (b) between binding sites within a single-file pore. No two substrates can simultaneously occupy the same site. Rate constants for hopping can be increased both (a) when substrates in two sites attract each other into a vacant site between them and (b) when substrates in adjacent sites repel each other. Hopping rate constants for charged substrates are also modified by the membrane field. For a 3-site model, simulated annealing yields parameters to fit steady-state measurements of flux coupling, transport-associated currents, and charge movements for the GABA transporter GAT1. The model then accounts for some GAT1 kinetic data as well. The model also yields parameters that describe the available data for the rat 5-HT transporter and for the rabbit Na+-glucose transporter. The simulations show that coupled fluxes and other aspects of ion transport can be explained by a model that includes local substrate-substrate interactions but no explicit global conformational changes.
Yongwei Cao
The rat serotonin transporter is a member of the family of neurotransmitter transporters that are dependent on external Na+ ions for substrate translocation. Previous voltage-clamp¹ and single-channel² recordings revealed three distinct currents associated with the rat serotonin transporter expressed in oocytes: (1) the substrate- and Na+-dependent transport-associated current, (2) the substrate-independent leakage current, and (3) the substrate-independent and hyperpolarization-induced transient current. My recent data shows that acidic external pH differentially affected these three currents. Proton significantly potentiates the transport-associated current when the external pH is lowered to below 5.5 from 7.5. Correspondingly, the 5HT uptake is reduced at pH values below 5.5. This is probably due to the dissipation of Na driving force for 5HT uptake by the large transport-associated current at acidic pH values. Protons are highly permeable in the leakage state. The proton permeability is at least 20,000 times greater than that for Na+, K+, and Li+. This high permeability leads to several tens to several hundreds of nA of inward H+ current when external pH is lowered to 6.5 and below. The proton leakage current is also present in many other Na+-coupled neurotransmitter transporters such as dopamine and GABA transporters. It is also present is some orphan transporters, enabling us to study these transporters in the absence of known substrate and to search for a potential substrate. Unlike in the leakage state, proton is not permeable in the transient current state. Furthermore, protons dramatically inhibit the transient current. These results suggest that the proton may provide us an additional tool for dissecting different aspects of transporter function.
In addition, a site-directed mutation was found to alter the three conducting states and the coupling of substrate transport with Na+ ions. The G250E mutation in the loop between the 4th and the 5th putative transmembrane domains dramatically reduces the transport-associated current and the transient current but increases the leakage current. In addition, this mutant transporter no longer shows the Na+-coupled transport. 5HT can be taken up by oocytes expressing this mutant in the absence of extracellular Na+ or other alkaline ions (NMDG substitution). Several other mutation are now being characterized. It is our ultimate goal to combine site-directed mutagenesis analysis and the parameters revealed from electrophysiological recordings, in order to develop a structure-function model that explains how ion-coupled transporters work.
1. Mager et al.(1994) Neuron 12:845-859
2. Lin et al., (1996) submitted.
Fan Lin, Sela Mager
The mammalian 5-HT transporter is an important target for drugs of therapy and abuse. We have combined single-channel recording techniques and site-directed mutagenesis to study permeation properties of the cloned rat 5-HT transporter expressed in oocytes. We demonstrate that the permeation pathway of the transporter behaves like an ionic channel. There are at least two distinct transport-related single channel conducting states. 1) In the absence of 5-HT, the transporter displays an ohmic conductance of ~6 pS. When Na+ is replaced by Li+, the open probability of this state increases. 2) In the presence of 5-HT, the transporter appears to open a different permeation pathway. This transmitter-dependent ionic current requires the presence of Na+ and is blocked if Na+ is replaced by Li+.
We mutated asparagine177 in the third transmembrane domain to glycine. This mutation more than doubles the single-channel conductance of both states, without significantly changing the other kinetic parameters. Therefore, we conclude that this residue is located in or near the permeation pathway.
Sela Mager, Nurit Kleinberger-Doron#, Gilmor I. Keshet#, Baruch I. Kanner#
#Department of Biochemistry, Hadassah Medical School, The Hebrew University, Jerusalem 91120, Israel
These experiments measure the binding of ions and the permeation of substrates during function of the GABA transporter GAT1. GAT1 was expressed in Xenopus oocytes and studied electrophysiologically as well as with [3H]GABA flux; GAT1 was also expressed in mammalian cells and studied with [3H]GABA and [3H]tiagabine binding. Voltage jumps, Na+ and Cl- concentration jumps, and exposure to high-affinity blockers (NO-05-711 and SKF-100330A) all produce capacitive charge movements, and occlusive interactions among these 3 types of perturbations show that they all measure the same population of charges. The concentration dependences of the charge movements reveal (a) that two Na+ ions interact with the transporter even in the absence of GABA and (b) that Cl- facilitates the binding of Na-. Comparison between the charge movements and the transport-associated current shows that this initial Na+-transporter interaction limits the overall transport rate when [GABA] is saturating. However, two classes of manipulation--treatment with high-affinity uptake blockers and the W68L mutation--"lock" Na+ onto the transporter by slowing or preventing the subsequent events that release the substrates to the intracellular medium. In another contrast with normal transport, the Na+ substitutes Li+ and Cs+ do not support charge movements, but they can permeate the transporter in an uncoupled fashion. Our results (a) support the hypothesis that efficient removal of synaptic transmitter by the GABA transporter GAT1 depends upon the prior binding of Na+ and Cl+ and (b) indicate the important role of the conserved putative transmembrane domain 1 in interactions with the permeant substrates.
Michael W. Quick¹*, Janis L. Corey²†, Norman Davidson²
¹Neurobiology Research Center and Department of Physiology and Biophysics, University of Alabama at Birmingham, Birmingham AL 35294-0021 and ²SIBIA, La Jolla, CA
Recent evidence indicates that several members of the Na+-coupled transporter family are targeted to a PKC-regulated secretory pathway and that PKC affects transport activity by redistribution of transporters between cytoplasmic vesicles and the plasma membrane. We elucidate components of this process for the GABA transporter GAT1 expressed in Xenopus oocytes using a combination of uptake assays, immunoblots, and electrophysiological measurements of membrane capacitance, transport-associated currents, and GAT1-specific charge movements. Co-injection of rat brain mRNA with GAT1 permits modulation at high transporter expression levels; this modulation is eliminated by injecting antisense synaptophysin oligonucleotides. The synaptophysin requirement suggested an interaction between GAT1 and components of the vesicular docking/fusion apparatus, so we used botulinum toxins to inactivate proteins involved in vesicle exocytosis. Inactivation of SNAP-25 and syntaxin-related proteins reduced basal GAT1 expression and prevented PKC-induced translocation. Inactivation of vesicular proteins synaptobrevin and cellubrevin prevented PKC-induced translocation but did not change basal GAT1 expression. Intracellular calcium mimicked the translocation seen with PKC. In addition, CaM kinase inhibition reduced basal GABA uptake and eliminated the PKC-induced increase in transport but did not alter the subcellular distribution of GAT1 or its PKC-induced translocation, suggesting an inactive transporter state that requires CaM kinase for function. Thus, functional expression and modulation of GAT1 in oocytes occurs via a pathway that uses components common to vesicle secretion, and suggests a relationship between factors that control neurotransmitter secretion and the components necessary for neurotransmitter uptake.
Scott K. Silverman, Dennis A. Dougherty
We have investigated a highly conserved tyrosine residue in the potassium channel pore region "signature sequence," TXXTXGYG. This residue is present in both of the G-protein-coupled inward-rectifier potassium channel subunits GIRK1 and GIRK4, and both subunits are required to form a functional heteromultimeric channel (see Silverman et al., below). We have substituted for the tyrosine residue using the nonsense codon suppression method in Xenopus oocytes. While the tyrosine residue in GIRK1 may be replaced with a variety of natural and unnatural amino acids of varying chemical structure (e.g., 1- and 2-naphthylalanine, valine) without much effect on channel expression, the corresponding residue in GIRK4 is much less tolerant of substitution. Even minor perturbation from tyrosine to 4-methoxyphenylalanine (O-methyltyrosine) drastically diminishes the functional channel expression. Further studies are in progress to clarify the different contributions of the GIRK1 and GIRK4 subunits to the potassium channel.
Paulo Kofuji, Magdalena Hofer¹, Kathleen J. Millen¹, James Millonig¹ and Mary E. Hatten¹
¹Laboratory of Developmental Neurobiology. The Rockefeller University. New York, NY 10021
The weaver (wv) mutation in mouse prevents the differentiation of cerebellar granule cells. Although proliferation is normal in wv/wv granule cells, the neurons never extend neurites or migrate along glial fibers, the hallmarks of differentiation in these cells. During the cerebellar development, wv/wv granule cells remain in the external granule cell layer where they eventually die. The massive loss of mature granule cells in the adult disrupts the cerebellar laminar structure. Thus there is no internal granule cell layer and the molecular layer which is normally composed in part of granule cell parallel fibers is considerably reduced in size. Chimeric analysis and cell culture experiments have demonstrated that the mutation acts within granule cells. Besides the cerebellar disfunction, the weaver mouse is also characterized by the loss of dopaminergic neurons in substantia nigra in a manner reminiscent of Parkinson's Disease.
Surprisingly, in weaver mouse, a point mutation was identified in the GIRK2 gene which encodes a G protein-gated inwardly rectifying K+ channel. GIRK2 was mapped to distal chromosome 16 near the weaver locus and was shown to be expressed in the developing cerebellum although the specific expression pattern of GIRK2 during cerebellar development has not been shown. The mutation found in weaver GIRK2 (wvGIRK2) exchanges a glycine residue for a serine in the pore-forming domain. This amino acid is conserved among all K+ channels but an effect of the alteration on GIRK2 function has not been demonstrated.
We have investigated the consequence of the weaver mutation on GIRK2 function first in a heterologous system and then in cerebellar granule cells. Expression of wvGIRK2 in Xenopus oocytes demonstrates that the mutant channel is highly permeable to Na+ and K+ ions. Moreover, the mutation causes the loss of G protein gating, rendering the channel constitutively active. In situ hybridizations demonstrate that GIRK2 as well as GIRK1 and GIRK3 mRNAs are highly expressed in cerebellar granule cells at the time when these cells are migrating away from the external germinal layer and extending the neurites to form the parallel fibers. In purified granule cells cultured in vitro, robust G protein-activated currents are observed in wild-type but not in wv/wv cells, instead, wv/wv cells display a higher current leakage in Na+ solution, consistent with the results obtained in Xenopus oocytes. This leakage could be prevented by open channel blockers such as QX-314 and MK-801.
Remarkably, QX-314 and MK-801 when added in weaver granule cells cultured in vitro promoted the rescue of the wv/wv phenotype, i.e. the cells extended long neurites and displayed TAG-1 immunoreactivity. Thus, Na+ flux through the wvGIRK2 channel underlies the failure of granule cell to differentiate. We are currently engaged in determining the impact of weaver mutation in substantia nigra and other CNS structures. These studies should clarify the role of G protein-gated channel currents and membrane hyperpolarization in neuronal development.
Scott K. Silverman, Dennis A. Dougherty
Recent work has established that the weaver phenotype in mouse is caused by a pore-region mutation in the G-protein-coupled inward-rectifier potassium channel subunit GIRK2. A single missense mutation (GYG to SYG) alters the functional channel so that it is no longer potassium selective, but instead allows sodium to permeate. We are interested to know if we can transfer this phenotype to related potassium channel subunits by making the homologous mutation in their pore regions. We have begun by investigating this weaver mutation in the homologous GIRK1 and GIRK4 subunits. The preliminary data show that the channels containing a mutated GIRK4 subunit still function yet are no longer potassium selective, while a similar mutation in GIRK1 reduces channel function overall.
Scott K. Silverman, Dennis A. Dougherty
G-protein-coupled inward-rectifier potassium channels (GIRK's) are multimers of at least two different kinds of subunit. Recent studies on other inward rectifiers have demonstrated that they are most likely tetramers, but the stoichiometry and arrangement of the GIRK subunits have yet to be investigated. We have designed experiments to determine these properties for the channel formed from the GIRK1 and GIRK4 subunits. We prepared multimeric GIRK constructs (in which more than one channel subunit is directly connected) and determined their functional expression in Xenopus oocytes either alone or with coexpressed monomer subunits. For example, a trimeric construct consisting of GIRK1, GIRK4, and GIRK1 concatenated in that order (1-4-1) expresses poorly alone or when coexpressed with the GIRK1 monomer, but shows a large (~6-fold) increase in expression level when coexpressed with the GIRK4 monomer. From a number of such experiments we have established that the functional channel is a 2+2 tetramer of GIRK1 and GIRK4 subunits. The data are less clear on the arrangement of these subunits but suggest that more than one arrangement is viable.
2 interactions
with the carboxyl terminus of Kir3.4 inward rectifier K+ channel
subunitsCraig A. Doupnik, Carmen W. Dessauer#, Alfred G. Gilman#, Tina Iverson
#Department of Pharmacology, University of Texas Southwestern Medical School, Dallas, TX 75235
The direct interaction of recombinant Gß1 2 proteins
with the carboxyl terminal domain of a G protein-gated inward rectifier K channel
subunit, Kir3.4 (GIRK4), was measured in real time using microsensor chip technology.
The carboxyl terminus of Kir3.4 (a.a.186-420) was expressed in bacteria as
a glutathione-S-transferase (GST) fusion protein, GST-Kir3.4ct. GST-Kir3.4ct
was immobilized to the surface of a microsensor chip by high affinity binding
of the GST domain to a covalently attached anti-GST antibody. The association
and dissociation rates of Gß1 2 dimers
with the immobilized Kir3.4ct domain were temporally resolved as a change in
refractive index detected by surface plasmon resonance. Specific binding of
Gß1 2 dimers
to Kir3.4ct was characterized by a dissociation rate (kd)
of ~0.004 s-1. Association kinetics were dominated by a concentration-independent
component (time constant ~ 50 s) which complicates models of binding and may
indicate conformational changes during binding of Gß1 2 to
Kir3.4ct. These studies demonstrate that Gß1 dimers
directly interact with the Kir3.4 channel subunit, and suggest interesting
details in the interaction with the major cytosolic carboxyl terminal domain.
Biosensor-based experiments such as these will complement electrophysiological
studies towards understanding the molecular basis of G protein interactions
with Kir channels and other ion channel proteins.
Jancy McPhee
A three amino acid motif, arginine-glycine-aspartate (RGD), at the proposed extracellular surface, just carboxy terminal to the first transmembrane domain, is well-conserved on all known GIRK subunits. This sequence is the same as the ligand frequently recognized by another family of transmembrane proteins called integrins. Integrins often recognize an RGD sequence found on extracellular matrix molecules; accordingly, they have been associated with functions such as cell-substrate adhesion or cell-cell adhesion. Interestingly, they have also been implicated in neuronal excitability in the hippocampus, neurite outgrowth, and stretch-induced neurotransmitter release. It is conceivable, therefore, that integrins may regulate ion channel activity. Because of the presence of a classic integrin ligand recognition site, RGD, on all GIRK subunits, we decided to examine whether GIRK activity, localization or expression was affected directly by interaction with integrin. Our initial approach is to express the muscarinic acetylcholine receptor and a combination of GIRK1 and GIRK4 subunits by RNA injection in Xenopus laevis oocytes. The combination of these two GIRK subunits forms a heteromultimeric channel complex that can be activated by application of acetylcholine acting through the exogenously coexpressed muscarinic receptor and an endogenous oocyte G-protein. We are determining whether native oocyte integrin interacts with the heterologously expressed channel in two ways. First, we are establishing whether long-term preincubation or short-term perfusion of RGD-containing peptides, which compete with integrin-ligand interactions, alter GIRK currents. As a control, we are also analyzing the effects of RGE-containing peptides, which do not interfere with integrin-ligand interactions, on GIRK activity. Our second approach involves mutagenesis of the extracellular GIRK RGD site to RGE. This mutation should be sufficient to disrupt GIRK channel-integrin interactions if they exist. Preliminary data suggest that the GIRK1-RGE/GIRK4-RGE double mutant complex produces no detectable current. On the other hand, the GIRK1-RGE/GIRK4 complex consistently expresses at a level 3-4 times lower than that produced from the same amount of WT GIRK1/GIRK4 RNA. Even more strikingly, the GIRK1/GIRK4-RGE complex expresses at about 10 fold lower levels than the double WT GIRK subunit complex. Voltage jump relaxations steps clarify that, in the one mutant/one WT subunit combinations described above, the mutant subunits are actually being expressed and that observed current is not merely due to the expression of the WT subunit. Because of these preliminary results, we are intrigued that an interaction of GIRK and integrin may occur and that this interaction might regulate the amount or timing of GIRK expression or localization in the oocytes. We must, however, perform more direct experiments to confirm whether this integrin-GIRK interaction exists and to resolve how this interaction could regulate GIRK currents. Such experiments are the focus of our future research.
and ß subunits of
G proteinsWolfgang Schreibmayer‡, Carmen W. Dessauer §, Dimitry Vorobiov†, Alfred G. Gilman§, Nathan Dascal†*.
† Department of Physiology and Pharmacology, Sackler School of Medicine, Tel Aviv University, Ramat Aviv 69978, Israel; ‡Institut für Medizinische Physik und Biophysik, Universität Graz, A-8010 Graz, Austria; § Department of Pharmacology, University of Texas, Southwestern Medical School, Dallas, TX 75235-9041, USA
Cholinergic muscarinic, serotonergic, opioid, and several other G protein-coupled
neurotransmitter receptors activate inwardly rectifying K+ channels
of the GIRK family, slowing the heartbeat and decreasing excitability of neuronal
cells. Previously both G and
Gß subunits
of heterotrimeric G proteins have been implicated in causing channel opening,
but recent studies attribute this role primarily to the Gß dimer
that activates GIRKs in a membrane-delimited fashion, probably by direct binding
to the channel protein. We show that free GTP S-activated
G i1,
but not G i2 or
G i3,
potently inhibits Gß1 2-induced
GIRK activity in excised membrane patches of Xenopus oocytes expressing
GIRK1. High-affinity but partial inhibition is produced by G s-GTP S. G i1-GTP S also inhibits
Gß1 2-activated
GIRK in atrial myocytes. Antagonistic interactions between G and
Gß may
be among the mechanisms determining specificity of G protein coupling to GIRKs.
Inhibitory modulation of GIRKs by G protein-coupled receptors may have important
implications in cardiac and brain physiology.
Markus U. Ehrengruber, Mark C. Jasek, Youfeng Xu, S. Jennifer Stary, Craig A. Doupnik, Justine Garvey, Norman Davidson
Heteromeric G protein-activated inwardly rectifying K+ channels (GIRKs) are important in modulating cellular excitability. GIRK1 was originally cloned from rat atrium where it forms a heteromultimer with GIRK4. The family members GIRK1, GIRK2, GIRK3, and GIRK4 are expressed in the brain. To study the role of GIRKs in neuronal synaptic transmission, we plan to use recombinant adenovirus for in vitro and in vivo ectopic overexpression of GIRKs. We have inserted GIRK1 and GIRK2 cDNA under the control of a CMV promoter into adenovirus. Adenoviral vectors with a Shaker K+ channel cDNA and the serotonin receptor 1A cDNA were also constructed. We infected primary cultures of rat atria and ventricle cells, and secondary cultures of pancreatic ßTC3, HEK 293, CHO, and HeLa cells with these adenoviral recombinants. Shaker peak currents of several nA were recorded in all cell types 1-2 d after infection. To test the GIRK1 recombinant we cotransfected CHO cells with plasmids carrying cDNA for GIRK4 and m2 acetylcholine (ACh) receptor, and infected them with GIRK1. ACh-activated GIRK currents of up to 1 nA (at -120 mV and 25 mM [K+]o) were found, proving that our GIRK1 adenovirus is useful for ectopic overexpression. The GIRK2 adenovirus, however, induced no GIRK currents in combination with GIRK1, and no GIRK2 protein was detected by immunoblots. By contrast, total RNA from infected cells expressed GIRK2 currents in Xenopus oocytes. Significant GIRK currents were detected in cells cotransfected with GIRK1 and GIRK4, but not with GIRK1 and GIRK2, nor with GIRK2 and GIRK4. Thus an unknown mechanism interferes with expression of our GIRK2 adenoviral construct in mammalian cells. We are now constructing an adenovirus carrying GIRK4 cDNA to allow ectopic expression of GIRK currents by coinfection with GIRK1 and GIRK4.