Zinc Potentiates GluK3 Glutamate Receptor Function
by Stabilizing the Ligand Binding Domain
Julien Veran,1,2Janesh Kumar,3Paulo S. Pinheiro,1,2,5Axel Athane ´,1,2Mark L. Mayer,3David Perrais,1,2,4
and Christophe Mulle1,2,4,*
1University of Bordeaux
Interdisciplinary Institute for Neuroscience, UMR 5297, 33000 Bordeaux, France
3Laboratory of Cellular and Molecular Neurophysiology, Porter Neuroscience Research Center, National Institute of Child Health and Human
Development, National Institutes of Health, Department of Health and Human Services, Bethesda, MD 20892, USA
4These authors contributed equally to this work
5Present address: Department of Neuroscience and Pharmacology, Faculty of Health Sciences, University of Copenhagen,
2200 Copenhagen N, Denmark
Kainate receptors (KARs) play a key role in the regu-
lation of synaptic networks. Here, we show that zinc,
a cation released at a subset of glutamatergic
synapses, potentiates glutamate currents mediated
by homomeric and heteromeric KARs containing
GluK3 at 10–100 mM concentrations, whereas it
inhibits other KAR subtypes. Potentiation of GluK3
currents is mainly due to reduced desensitization,
as shown by kinetic analysis and desensitization
mutants. Crystallographic and mutation analyses re-
vealed that a specific zinc binding site is formed at
the base of the ligand binding domain (LBD) dimer
interface by a GluK3-specific aspartate (Asp759),
together with two conserved residues, His762 and
Asp730, the latter located on the partner subunit. In
addition, we propose that tetrameric GluK2/GluK3
receptors are likely assembled as pairs of heterodi-
meric LBDs. Therefore, zinc binding stabilizes the
labile GluK3 dimer interface, slows desensitization,
and potentiates currents, providing a mechanism
for KAR potentiation at glutamatergic synapses.
Glutamate released at excitatory synapses acts on ligand-gated
ionotropic receptors, which fall into three classes, named
after their preferred or selective agonist: a-amino-3-hydroxy-5-
methyl-4-isoxazolepropionic acid (AMPA), N-methyl D-aspar-
tate (NMDA), and kainate. Like other neurotransmitter receptors,
ionotropic glutamate receptors harbor binding sites for small
molecules that act as allosteric regulators of receptor function.
Allosteric modulators are likely to play a key role in the regulation
of synaptic transmission and moreover represent potential phar-
of AMPA, kainate, and NMDA receptors (NMDARs) are thought
to be promising therapeutic agents in the treatment of cognitive
dysfunctions (Traynelis et al., 2010). It is thus important to
increase our understanding of the molecular mechanisms by
in the regulation of synaptic transmission.
Ionotropic glutamate receptors are mainly localized at post-
synaptic sites where they are directly involved in the transfer
of information across synapses. There is however increasing
evidence that ionotropic glutamate receptors are also present
at presynaptic sites, where they regulate the release of neuro-
transmitters and participate in presynaptic forms of plasticity
(Pinheiro and Mulle, 2008). In the brain, kainate receptors
(KARs) support a variety of functions contributing to the re-
gulation of the activity of synaptic networks (Contractor et al.,
2011). KARs are tetramers composed of a combination of the
five subunits, GluK1–GluK5, previously GluR5–GluR7,KA1–KA2
(Contractor et al., 2011). KARs share a similar architecture with
other ionotropic glutamate receptors; the subunits have a large
extracellular domain composed of an amino-terminal domain
(ATD) and a ligand binding domain (LBD), a membrane region
composed of three membrane a helices and a reentrant loop,
and an intracellular carboxy-terminal region (Mayer, 2011). The
various roles of KARs at pre- or postsynaptic sites arise in part
from the diversity of functional properties of the different KAR
subtypes (Perrais et al., 2010). At hippocampal mossy fiber
synapses onto CA3 pyramidal cells, KARs are present at both
pre- and postsynaptic levels (Contractor et al., 2011). Postsyn-
aptic KARs are composed of the GluK2, GluK4, and GluK5
subunits (Contractor et al., 2003; Fernandes et al., 2009; Mulle
et al., 1998), whereas presynaptic KARs are thought to comprise
the GluK2 and GluK3 subunits (Contractor et al., 2001; Pinheiro
et al., 2007). The functional properties of GluK3 (and GluK2/
GluK3) receptors set it apart from the other ionotropic gluta-
mate receptors (Perrais et al., 2009a; Schiffer et al., 1997). In
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 565
particular, its sensitivity to glutamate is the lowest of all known
ionotropic glutamate receptors, due in large part to fast desensi-
tization of receptors with only one or two bound glutamate mole-
cules (Perrais et al., 2009a). The low agonist sensitivity of this
receptor raises questions about its relevance for synaptic
function (Perrais et al., 2010). Therefore, it is possible that en-
dogenous modulators may potentiate its responsiveness to
Among potential endogenous modulators of KAR function,
we chose to address the role of zinc, known to be present in
large amounts in hippocampal mossy fiber terminals (Paoletti
et al., 2009). Zinc is accumulated into synaptic vesicles and
thought to be coreleased with glutamate in the extracellular
milieu during neuronal activity (Paoletti et al., 2009). The best-
characterized synaptic zinc targets are NMDARs (Westbrook
and Mayer, 1987). Zinc inhibits NMDAR function with affinities
ranging from low nanomolar for GluN1/GluN2A receptors to
low micromolar for GluN1/GluN2B subunits (Paoletti et al.,
1997). The binding site accounting for the high-affinity binding
of zinc to GluN2A and GluN2B has been mapped to the large
ATD of GluN2 subunits (Choi and Lipton, 1999; Karakas et al.,
2009; Paoletti et al., 2000; Rachline et al., 2005). Zinc binding
to the ATD has been suggested to inhibit NMDAR channel gating
through destabilization of the dimer interface of the LBD (Erreger
et al., 2005; Gielen et al., 2008), by mechanisms that resemble
desensitization of AMPA and KARs (Armstrong et al., 2006;
Weston et al., 2006). Zinc has also been reported to inhibit native
dent, with KARs containing GluK4 or GluK5 subunits being more
sensitive, IC50?1–2 mM, than GluK1-GluK2, IC50?70 mM (Mott
et al., 2008). Despite the proposed presynaptic function of
GluK3-containing KARs at hippocampal mossy fiber synapses,
which are highly enriched in vesicular zinc, modulation of these
receptors by zinc has not yet been addressed.
In this study, we show that zinc at micromolar concentrations
potentiates recombinant GluK3 receptor currents evoked by
glutamate. Zinc markedly slows receptor desensitization and
increases apparent affinity for glutamate. By analysis of chi-
meric GluK2/GluK3 KARs and of GluK3 bearing selected point
mutations, we mapped the zinc binding domain to the S2
segment of the LBD, in a region forming the interface between
two GluK3 subunits in an LBD dimer assembly. Crystallographic
studies for GluK3 LBD complexes with both glutamate and
kainate revealed that zinc ions bind at multiple sites formed
by aspartate, histidine, and glutamate residues, which are
present in both the upper and lower lobes of the LBD. Based
on these crystal structures, a GluK3 LBD dimer model was
generated by superposition of GluK3 monomers on previously
solved KAR LBD dimers. This identified D730 as the dimer
partner component of the binding site underlying zinc poten-
tiation, together with D759 and H762 from the adjacent sub-
unit. Based on these structure-function studies and on modeling
of KAR activity, we show that zinc plays a very distinct role in
GluK3-KARs by stabilizing the LBD dimer assembly, thereby
reducing desensitization. Given the proposed presynaptic
localization of GluK3 close to zinc-containing synaptic vesicles,
zinc may be an endogenous allosteric modulator for native
Zinc Reversibly Facilitates Recombinant GluK3
tion of glutamate on lifted HEK293 cells transfected with GluK3
cDNA. Currents evoked by sustained applications (100 ms) of
10 mM glutamate, a concentration close to the EC50value for
GluK3 (Perrais et al., 2009a; Schiffer et al., 1997), were markedly
enhanced with preapplication of 100 mM zinc (Figure 1A; 193% ±
38% of control amplitude, n = 17), and this potentiation was
rapidly reversible upon removal of zinc. In contrast to GluK3
potentiation, and as previously reported in Xenopus oocytes
(Mott et al., 2008), zinc reversibly inhibited GluK2 currents at all
concentrations tested (Figures 1A and 1D), with an IC50 of
102 ± 11 mM and a Hill coefficient (nH) of 1.1 ± 0.1 (n = 4–9).
Because a glutamate concentration of 10 mM is saturating for
GluK2 (Perrais et al., 2009a), this could mask a potentiating
effect of zinc. However, currents evoked by 500 mM glutamate,
a concentration below the EC50for GluK2, were also inhibited
by 100 mM zinc (48% ± 10%, n = 10; data not shown). It has
been proposed that presynaptic GluK2/GluK3 heteromeric
receptors regulate glutamate release at hippocampal mossy
fiber synapses (Pinheiro et al., 2007). Therefore, it was important
to determine the effect of zinc on heteromeric GluK2/GluK3
receptors. To test the specific effects of zinc on GluK2/GluK3
heteromers in cells cotransfected with GluK2 and GluK3, we
reduced the likeliness of activating homomeric GluK2 or GluK3
subunits as described previously (Perrais et al., 2009b). First,
the GluK2b(Q) splice variant was used because of its reduced
expression at the cell surface as a homomer (Jaskolski et al.,
2004). Second, GluK3 homomeric receptors were specifically
blocked with 1 mM UBP310 (Perrais et al., 2009b). In cells
cotransfected with GluK2b(Q) and GluK3, application of 1 mM
UBP310 inhibited glutamate-activated currents by 55% (n = 6;
p < 0.05). The fraction of current resistant to UBP310 was
enhanced by zinc (100 mM) to a similar extent (157% ± 7%, n =
18) as for homomeric GluK3 receptors (p = 0.65; Figures 1B
and 1C). The small fraction of homomeric GluK2 receptors at
the cell surface would, if anything, lead to an underestimation
of the potentiation of GluK2/GluK3 receptors by zinc. Therefore,
these results clearly demonstrate that heteromeric GluK2/GluK3
receptors are, like GluK3 receptors, potentiated by zinc.
The modulation of GluK3 by zinc showed a dose-depen-
dent biphasic effect: increasing the concentration of zinc up
to 100 mM potentiated currents (half-maximal effect around
20 mM), and higher concentrations progressively inhibited
currents (Figure 1D). In order to fit the dose-response curve
with combined potentiation/inhibition Hill equations, we hypoth-
esized that the inhibition of GluK3 by higher concentrations of
zinc was similar to that of GluK2 (a notion supported by the
effects of point mutations described in Figure 6). This attempt
to separate potentiation and inhibition in the GluK3 dose-
response curves yielded an EC50 value of 46 ± 17 mM, nH
1.82 ± 0.95, and a maximal potentiation of 475% ± 47%,
although the moderate quality of the combined fit suggests
that potentiation and inhibition might not be independent
processes. Surprisingly, zinc potentiated currents mediated by
Zinc Potentiates GluK3 Receptors
566 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
GluK2/GluK3 at all concentrations tested (Figure 1D), with an
EC50of 477 ± 1638 mM, nH 0.6 ± 0.4, consistent with a reduced
potentiation of 286% ± 195% of control, and by contrast to
homomeric GluK3 receptors, there was no inhibition for zinc
concentrations up to 1 mM.
Zinc Changes the Kinetic Properties of GluK3 and Its
affinity for Glutamate
Zinc could affect GluK3-mediated currents in several ways:
it could increase single-channel conductance, increase open
probability, allow activation of ‘‘silent’’ receptors, or slow down
receptor desensitization. It was shown previously that the low
glutamate sensitivity of GluK3 receptors was due to fast transi-
tions of glutamate bound receptors to desensitized states
(Perrais et al., 2009a); hence, a change in desensitization may
translate into an increased apparent affinity for glutamate. In
order to analyze the mechanisms of zinc potentiation, we char-
acterized key parameters of GluK3 receptor activation using
outside-out patches (Figure 2). The current decay (tdes) evoked
by 100 ms applications was well fitted with single exponential
functions, and zinc increased tdesin a dose-dependent manner
(tdes= 5.0 ±0.2 ms; versus 9.8 ± 0.4 ms, before and after 100 mM
line), whereas it had no effect on GluK2 desensitization (tdes=
3.4 ± 0.1 ms versus 3.7 ± 0.1 ms before and after 100 mM zinc,
Figure 1. Zinc Reversibly Facilitates GluK3-
Containing KAR Currents in a Dose-Depen-
(A) Effect of a sustained application of 100 mM zinc
(Zn) on GluK3 (top traces) and GluK2 (bottom
traces) currents, evoked by 10 mM glutamate (Glu)
for 100 ms and recorded in the whole-cell mode, is
(B) Effect of 100 mM zinc on GluK2b/K3 receptors
in presence of 1 mM UBP310 to block homomeric
GluK3 is illustrated.
(C) Histogram presents the effect of 100 mM zinc
on GluK2, GluK3, and GluK2/GluK3 receptors.
***p < 0.001, **p < 0.01, *p < 0.05, ns, p > 0.05.
norm. amplitude, normal amplitude.
(D) Dose-response curves for the effects of zinc on
GluK3 (black circles), GluK2 (white circles), and
GluK2/GluK3 (black squares) mediated currents
elicited by 10 mM glutamate (n = 4–18 for each
point) are illustrated. Solid lines show Hill equa-
477 mM, nH = 0.6 and maximum 286% for GluK2b/
K3). The dotted line shows the fit for a combined
Hill equation, with an EC50of 46 mM for potentia-
tion of GluK3, and an IC50of 100 mM, calculated
from inhibition for GluK2 (solid line).
n = 5; p = 0.20; Figure 2C) despite its
strong effect on GluK2 current ampli-
tudes. Therefore, zinc markedly affects
GluK3 receptor desensitization, in con-
trast to its lack of effect on GluK2
kinetics, suggesting a different mecha-
nism of action. Moreover, zinc increased currents evoked by
1 ms pulses of 10 mM glutamate (191% ± 22% of control ampli-
tude, n = 4; Figure 2D) and slowed down their deactivation
kinetics (from 1.5 ± 0.05 ms to 2.3 ± 0.1 ms, n = 4; p = 0.002;
Next, we measured the EC50 for glutamate in outside-out
patches in the absence or presence of zinc (100 mM, Figures
2F and 2G). Zinc increased the sensitivity of GluK3 receptors
for glutamate from an EC50of 10.1 ± 1 mM (nH = 1.6 ± 0.1) in
control condition to 4.8 ± 1.1 mM (nH = 1.1 ± 0.2) with 100 mM
zinc (n = 4). Consequently, zinc is markedly more potent at low
glutamate concentrations (1–3 mM) than at 30 mM. In addition,
we found that zinc increases the time constants for desensitiza-
tion at all glutamate concentrations (Figure 2H). Finally, zinc
speeds up recovery from desensitization (time for half-recovery:
902 ± 1.1 ms and 460 ± 1.2 ms in absence and presence of zinc
suggest that zinc, by affecting the fast desensitization properties
of GluK3 receptors, enhances GluK3 currents and increases its
sensitivity to glutamate.
Reducing Receptor Desensitization Blocks Potentiation
Indeed, previous experimental and kinetic modeling data (Per-
rais et al., 2009a) have shown that fast desensitization of partially
liganded GluK3 receptors limits their activation. Therefore, the
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 567
potentiating effect of zinc should be reduced in GluK3 mutants
in which desensitization is slowed or abolished. We directly
tested this prediction with two GluK3 mutant receptors that
show reduced desensitization (Figure 3). We constructed a
mutant GluK3 receptor by analogy with a mutant GluK2 receptor
with four substitutions (K525E, K696R, I780L, and Q784K) that
greatly slowed down desensitization, termed GluK2(ERLK)
(Chaudhry et al., 2009; Zhang et al., 2006). Indeed, GluK3(ERLK)
desensitization (15.3 ± 1.9 ms, n = 7; p < 0.0001; Figures 3A and
3D) was about 3-fold slower than that of wild-type (WT) GluK3,
albeit the changes were not as dramatic as with GluK2(ERLK),
for which a 50-fold increase is observed (Chaudhry et al.,
2009; Zhang et al., 2006). As predicted, GluK3(ERLK)-mediated
currents were inhibited rather than potentiated by zinc (Figures
3A and 3C) with no change in desensitization rate (Figure 3D).
Another mutant,GluK3(H492C,L753C), forwhichdesensitization
2006), was also inhibited by zinc (100 mM) to a similar extent
(Figures 3B and 3C). Overall, the potentiating effect of zinc on
GluK3 is absent in two variants where GluK3 desensitization is
Effect of pH on GluK3 Function and Modulation by Zinc
An interaction between zinc modulation and pH has been docu-
mented for many zinc binding sites, in particular for NMDA (Choi
and Lipton, 1999; Low et al., 2000) and KARs (Mott et al., 2008).
This could reflect the protonation of the zinc binding site or other
allosteric mechanisms. Studying the interaction between pH and
zinc may provide information on the nature of the site involved in
GluK3 potentiation. We have observed a strong effect of pH on
GluK3 function: the current amplitude was much smaller at pH
8.3 and slightly higher at pH 6.8 than at pH 7.4. At pH 6.8, in
the absence of zinc, there was a slight decrease in rate of desen-
sitization of GluK3 currents (tdes4.7 ± 0.3 ms, n = 11 at pH 7.4, to
6.0 ± 0.5 ms, n = 8 at pH 6.8; p = 0.014). Interestingly, at pH 8.3,
we observed a much lower current amplitude and accelerated
Figure 2. Zinc Changes Desensitization and
Sensitivity to Glutamate of GluK3 Currents
Recorded in Outside-Out Patches
(A) Effect of various concentrations of zinc on
currents evoked by glutamate (10 mM) on an
outside-out patch pulled from a GluK3-expressing
cell is presented.
(B) Normalization of traces in (A) reveals the
dose-dependent effect of zinc on desensitization
(C) Values of tdesfor GluK2 and GluK3 mediated
currents with increasing zinc concentration (n = 5).
A total of 30 and 100 mM zinc induces a signifi-
cant increase of tdes relative to control (Ctrl).
***p < 0.001.
(D and E) Effect of zinc on currents evoked by 1 ms
pulses of glutamate (10 mM) is illustrated. In (E),
the two traces of (D) are normalized to maximum
(F–H) Effect of zinc on glutamate sensitivity is
demonstrated. (F) The amplitude of currents rela-
tive to their maximum amplitude (30 mM gluta-
mate) is increased in the presence of 100 mM zinc
(red traces). (G) Glutamate dose-response curves
in the absence (black circles) or presence (red
squares) of 100 mM zinc are shown. The lines are
fits with the Hill equation, with EC50 values of
10.1 ± 1.1 and 4.8 ± 1.1 mM, respectively.
(H) Relationship between current amplitude and
decay time with (red squares) or without (black
circles) 100 mM zinc (n = 4–9) is shown.
(I) Paired applications of glutamate (10 mM;
100 ms) separated by intervals ranging from 0.2 to
5 s on an outside-out patch pulled from a cell-ex-
pressing GluK3, in the absence (top) or presence
(bottom) of 100 mM zinc, are illustrated.
(J) Plot of the amplitude ratio of two successive
responses versus time interval for GluK3, with (red
squares) or without (black circles) 100 mM zinc, is
presented. The lines are fits with an exponential
equation, with time constant values of 432 ± 1 and
933 ± 1 ms, respectively (n = 5). fract. amplitude,
Zinc Potentiates GluK3 Receptors
568 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
desensitization (tdes2.7 ± 0.3 ms, n = 9; p < 0.0001; Figures 4A–
4C). Application of zinc (100 mM) inhibited currents at pH 6.8 but
potentiated currents at pH 8.3 (Figures 4D–4F). This suggests
that amino acid protonation at pH 6.8, most likely a histidine,
might be responsible for the loss of potentiation at low pH.
Facilitatory Effect of Zinc on GluK3 Receptors Is
Mediated by the S2 Extracellular Segment of the LBD
In AMPA receptors (AMPARs) and KARs, several studies have
shown that residues lining the interface between the LBDs of
two adjacent subunits are a key component of dimer stability
and regulate desensitization kinetics (Armstrong et al., 2006;
Chaudhry et al., 2009; Horning and Mayer, 2004; Nayeem
et al., 2009; Sun et al., 2002; Weston et al., 2006). To identify
the zinc binding sites responsible for the facilitatory effect on
GluK3 currents, we constructed chimeric receptors of GluK2
and GluK3. Receptors composed of the extracellular domain
of GluK3 and the transmembrane and intracellular segments of
GluK2 were potentiated by zinc to similar levels as GluK3
(175% ± 9% of control amplitude with 100 mM zinc, n = 5; Fig-
ure 5A, left, and Figure 5D). By contrast, zinc inhibited currents
mediated by chimeric receptors that contained the transmem-
brane and intracellular segments of GluK3 and the extracellular
domain of GluK2 (40% ± 8%, n = 4; p = 0.0077; Figure 5A, right,
and Figure 5D). In the GluN2A and GluN2B subunits of NMDARs,
the ATD harbors a discrete zinc binding site (Choi and Lipton,
1999; Karakas et al., 2009; Paoletti et al., 2000; Rachline et al.,
2005). GluK3 subunits deleted of their ATD form functional
receptors, which fully preserve potentiation by zinc (186% ±
13%, n = 5; p = 0.023; Figures 5B, left and 5D). Conversely, inhi-
bition is conserved in GluK2 receptors where the ATD was
excised (44% ± 11%, n = 4; p = 0.035; Figures 5B, right, and
5D). Thus, this first set of experiments seems to rule out a role
of the intracellular region, ion channel, and the ATD of GluK3
and points to the LBD asa potential zinc binding domain respon-
sible for the facilitatory effect of zinc. The LBD is formed by two
extracellular segments referred to as S1 and S2 (Stern-Bach
et al., 1994), which form a clamshell-like structure where S1
forms most of the upper half of the clamshell, and S2 forms
most of the lower half. We next tested which of these segments
is involved in the facilitatory effect of zinc on GluK3 receptors by
placed by the other and vice versa. Interestingly, currents medi-
ated by GluK3/GluK2 chimeric receptors that contain S2 and the
intracellular part of the GluK2 subunit (GluK3/K2S2) were in-
hibited (54% ± 6%; n = 5; p = 0.002; Figure 5C, left, and Fig-
ure 5D), whereas currents mediated by GluK3/GluK2 chimeric
receptors containing S2 and the intracellular part of the GluK3
subunit (GluK2/K3S2) were facilitated by 100 mM zinc (197% ±
9%, n = 5; p = 0.034; Figure 5C, right, and Figure 5D). Moreover,
whereas desensitization kinetics in control conditions was not
affected for most of the constructs tested (Figure 5E), the
kinetics of GluK3/K2S2 and GluK2/K3S2 was considerably
changed (from 5.0 ± 0.2 ms, n = 8 for WT GluK3 to 14.5 ±
0.8 ms, n = 4, p < 0.0001 for GluK3/K2S2; and from 3.4 ±
0.1 ms, n = 5 for WT GluK2 to 2 ± 0.1 ms for GluK2/K3S2,
n = 4, p < 0.0001). Slowed desensitization kinetics could explain
why GluK3 with the S2 segment substituted for GluK2 is func-
tional as assessed by slow glutamate application on Xenopus
oocytes (Strutz et al., 2001). These experiments clearly point to
the S2 segment of GluK3 as a target for zinc binding.
Identification of Key Residues at the Dimer Interface
Mediating the Facilitatory Effect of Zinc
To further characterize the zinc binding site, we hypothesized
that it might stabilize the interface between LBDs by binding to
a unique site generated by amino acids found only in GluK3.
Amongresiduesthatusuallybind zinc(histidine, cysteine,aspar-
tate, and glutamate), a single residue in S2 differs between
GluK3 and the other KARs: An aspartate in GluK3 (D759) is re-
placed by a glycine in GluK1 and GluK2 and by an asparagine
in GluK4 and GluK5 (Figure 6A). We tested the effects of
zinc (100 mM) on the reciprocal mutants GluK3(D759G) and
GluK2(G758D). Glutamate-activated currents were potentiated
by zinc in GluK2(G758D) receptors to the same extent as
6B and 6F). Conversely, GluK3(D759G) currents were inhibited
by zinc (32% ± 9%, n = 5; Figures 6C and 6F). These results
clearly indicate that the replacement of G758 in GluK2 by
an aspartate is sufficient to confer zinc potentiation in GluK2.
Figure 3. Reducing Desensitization Abolishes Potentiation by Zinc
(A and B) Inhibition by 100 mM zinc of GluK3 mutants attenuating desensiti-
zation is shown. GluK3(ERLK) (left traces) and GluK3(CC) (right traces)
currents, evoked by 10 mM glutamate for 100 ms, are illustrated.
(C) Histogram represents the effect of 100 mM zinc on GluK3(ERLK) and
GluK3(CC) currents, evoked by sustained (100 ms) applications of 10 mM
glutamate on outside-out patches.
(D) Histogram presents the desensitization kinetics of GluK2, GluK3, and
GluK3(ERLK), recorded inthesameconditionsas in(C).***p<0.001, *p< 0.05,
ns, p > 0.05.
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 569
Moreover, desensitization was markedly slower in GluK3
(D759G) (tdes= 18.4 ± 1.8 ms; n = 8; p < 0.0001) and greatly
accelerated in GluK2(G758D) (tdes= 1.3 ± 0.1 ms; n = 6; p <
0.0001; Figure 6G) compared to WT GluK3 and GluK2, respec-
tively. We observed the same effect when the whole S2 segment
was substituted in GluK3/K2S2 and GluK2/K3S2 chimeric re-
on GluK3, D759 in GluK3 appears as a key residue to explain the
specific desensitization properties of GluK3 as compared to
GluK1 and GluK2.
We searched for additional residues involved in the binding
of zinc in the vicinity of D759. This residue is localized in the
turn between helices J and K of the GluK3 LBD (Venskutonyt_ e
et al., 2011), three residues downstream of a conserved histidine
near the N terminus of helix K. This conserved histidine, together
with D759, is a candidate for residues forming part of the zinc
binding site for GluK3. To test this hypothesis, GluK3 H762
was replaced by an alanine. As expected, the facilitatory effect
of zinc was turned to an inhibitory effect in GluK3(H762A)
(peak amplitude 45% ± 6%, n = 5; p = 0.014; Figures 6D and
6F), strongly suggesting that H762 participates in the zinc
binding site of GluK3 receptors. In addition, the desensitization
kinetics of GluK3(H762A) was faster (tdes= 3.9 ± 0.1 ms, n = 5;
p = 0.0009; Figure 6G) than for WT GluK3, ruling out an indirect
effect of reduced desensitization on the effect of zinc in this
Structure-Based Analysis of the Zinc Binding Site in
To obtain further insight into the zinc binding site on GluK3, three
independent crystal structures were solved for the GluK3 LBD in
protomers in the glutamate complexes was similar to that re-
ported recently by Venskutonyt_ e et al. (2011) but with small vari-
ations in the extent of domain closure, from 25.3?to 23.4?, and
larger cavity volumes (299 ± 6 A˚3, mean ± SD, n = 6) than the
value of 274 ± 4 A˚3reported previously, indicating that the LBD
of GluK3 is more similar to GluK1 than GluK2, with cavity
volumes of 305 and 255 A˚3, respectively (Mayer, 2005). For the
two protomers in the GluK3 kainate complex (Figure S1), domain
closure was 90% and 70% of that induced by glutamate, indi-
cating that the GluK3 LBD can adopt both more closed, and
also more open, conformations than observed previously for
the GluK3 kainate complex for which a value of 81% was
reported by Venskutonyt_ e et al. (2012). In the crystal forms
reported here, the GluK3 protomers assemble as two different
dimers, both of which diverge from the canonical arrangement
Figure 4. Effect of pH on GluK3 Function and Modulation by Zinc
(A) Whole-cell GluK3 currents evoked by glutamate (10 mM) at the indicated pH are demonstrated.
(B) Current amplitude at different pHs, relative to value at pH 7.4 (n = 5), is presented.
(C) Values of tdesat the three pHs tested are illustrated. Inset is of normalized currents shown in (A), with tdesof 5.7, 4.3, and 3.9 ms at pH 6.8, 7.4, and 8.3,
(D) Effect of zinc (100 mM) on GluK3 at different pHs is presented: at pH 8.3, currents are potentiated, but not at pH 6.8.
(E and F) Effect of zinc on current amplitude and tdesat pH 6.8 (n = 6), 7.4 (n = 11), and 8.3 (n = 8) is shown. Open bars represent conditions with zinc. ***p < 0.001,
*p < 0.05, ns, p > 0.05.
Zinc Potentiates GluK3 Receptors
570 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
found in full-length GluA2 (Sobolevsky et al., 2009). In the P2221
glutamate and P2221kainate complexes, two GluK3 protomers
are arranged head to tail such that helix D of subunit A packs
against the N terminus of helix K in subunit B (Figure 7A). Crys-
tallographic symmetry operations generate a second dimer,
arranged in a head-to-head assembly but with a >20 A˚lateral
displacement of the two protomers such that helix J is packed
against helix J of its symmetry mate (Figure S1C). In the
P21212 glutamate complex, which has four protomers in the
asymmetric unit, both dimer forms are generated by noncrystal-
Calculation of anomalous difference Fourier electron density
molecular contacts between protomers in both crystal forms
(Figures 7A and S1C). It is likely that these zinc-mediated con-
Figure 5. Facilitatory Effect of Zinc on
GluK3 Receptors Is Mediated by the S2
Segment of the GluK3 LBD
(A–C) Chimeric GluK2/GluK3 receptors were
designed to isolate the region involved in the
facilitatory effect of zinc. Currents are evoked by
glutamate (10 mM, 100 ms) with or without 100 mM
zinc. (A) Left view shows chimeric receptors
composed of the extracellular domains of the
GluK3 subunit (in blue) with the transmembrane
and intracellular domains of the GluK2 subunit
(in red). Right view illustrates chimeric receptors
composed of the extracellular domains of the
GluK2 subunit (in red) with the transmembrane
and intracellular domains of the GluK3 subunit (in
blue). (B) Left view presents GluK3 receptors with
ATD deletion. Right view demonstrates GluK2
receptors with ATD deletion. (C) Left view shows
chimeric GluK3 receptors that contain the S2
segment, the last transmembrane a helix, and the
C-terminal domain of the GluK2 subunit. Right
view illustrates chimeric GluK2 receptors that
contain the S2 segment, the last transmembrane
a helix, and the C-terminal domain of the GluK3
(D) Histogram presents the effect of 100 mM
zinc on GluK2, GluK3, and chimeric receptor
current amplitude in whole cells. The asterisks
indicate significant difference withcontrol currents
(E) Histogram presents the desensitization kinetics
of GluK2, GluK3, and chimeric receptor currents
recorded in the same conditions as in (D). The
asterisks indicate a significant difference with
GluK3 or GluK2 currents.
***p < 0.001, **p < 0.01, *p < 0.05.
in the crystal structures, which could
explain why GluK3 failed to pack in the
canonical arrangement found in many
crystal structures for other iGluR LBDs.
However, GluK3 LBD glutamate and
kainate complexes, which were crystal-
lized in the absence of zinc ions (Ven-
skutonyt_ e et al., 2011, 2012), were also
packed in a noncanonical dimer configuration in the P41space
of nonbiological assemblies, as observed previously for GluK1
and GluK2 (Mayer, 2005; Naur et al., 2005). Relevant to the
zinc potentiation of GluK3, zinc ions were bound at a site labeled
Zn1, which was created by D759 in all three crystal forms (Fig-
ure 7B); the zinc ion at this site was also coordinated by H762,
mutation of which abolished potentiation by zinc, and by H492
and E495 located in helix D of the dimer partner subunit. Addi-
tional zinc binding sites, which stabilize the alternative dimer
assembly, were created by E757 at the base of helix J together
with E757 of its symmetry mate (Zn2), by H444 at the N terminus
of helix B (Zn3), by E713 in helix I together with H762 and
E766 from a symmetry related molecule (Zn4), by H479 at the
C terminus of helix C (Zn5), by the main-chain carbonyl oxygen
of Glu495 and the side-chain carboxylate of D499 just after the
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 571
C terminus of helix D (Zn6), by H479 with its symmetry mate
(Zn7), and by E441 with H444 of a symmetry-related molecule
(Zn8). For most of these sites, the binding of zinc was character-
ized by short bond lengths, on the order of 1.9–2.0 A˚(Figure 7A).
Although the GluK3 LBD dimer arrangements observed here
differ from the canonical arrangement of full-length GluA2
receptors (Sobolevsky et al., 2009), we tested functionally the
involvement of H492, which participates in the Zn1 site (Fig-
ure 7B), but which is absent in other iGluR subunits. We mutated
the histidine into a tyrosine, the equivalent residue in GluK2, and
also into an alanine. For both GluK3(H492Y) and GluK3(H492A),
zinc still potentiated currents (Figures 7D and 7F). Conversely,
GluK2(Y490H) was, like WT GluK2, inhibited by zinc (Figures
zinc binding site for GluK3 potentiation. This result is consistent
with the fact that substitution of the S1 region of GluK2 to GluK3
does not abolish zinc potentiation (Figures 5C and 5D).
The essentially identical conformation of GluK2 and GluK3
LBD glutamate complexes in crystal forms obtained both in the
presence and absence of zinc (Mayer, 2005; Venskutonyt_ e
et al., 2011), combined with 87% amino acid identity, and 94%
amino acid similarity of the GluK2 and GluK3 LBDs, provides
a basis for modeling a biological dimer assembly for GluK3,
based on GluK2 LBD dimer crystal structures. This approach is
further validated by the similar LBD dimer assemblies found in
the full-length GluA2 structure (Sobolevsky et al., 2009). The
rmsd for superposition of a protomer from the GluK3 P2221
glutamate complex on each of the two subunits in a GluK2
LDB dimer assembly (Protein Data Bank ID Code [PDB] 3G3F)
was 0.42 and 0.40 A˚for 242 Ca atoms, indicating that the
structures of the GluK2 and GluK3 LBDs are nearly identical.
Following this superposition, inspection of the GluK3 dimer
model revealed that selection of new rotamers for D730, D759,
and H762 would allow formation of intersubunit contacts with
appropriate bonding distances for zinc coordination; likewise,
binding sites for Na+and Cl?like those found in GluK1 and
GluK2 LBD dimers (Plested et al., 2008; Chaudhry et al., 2009)
could be created by adjusting side-chain torsion angles for
E495 and R745. The resulting GluK3 dimer model shows the
location and stoichiometry of three discrete binding sites for
allosteric ions: with a single Cl?ion on the 2-fold axis of dimer
symmetry, two Na+ions binding near the upper surface of
domain 1, and two zinc ions binding at the base of domain 1
(Figure 8A). This model identified D730 as the residue that
completes the coordination shell for zinc, together with the
main-chain carbonyl oxygen atom of Q756 and the side chains
of D759 and H762 from the adjacent subunit, together with one
or two water molecules that were not included in the model (Fig-
ure 8B). The resulting structure reveals two key features. First,
zinc acts as an intermolecular bridge between the pair of
subunits in an LBD dimer assembly. Second, in the absence of
zinc, the side chains of D730 and D759, which are separated
by only 2.9–3.8 A˚, would likely repel each other, destabilizing
the dimer assembly and accelerating desensitization. In sup-
port of this, neutralizing these charges by mutating D759 into
a glycine strongly reduces desensitization. Conversely, introdu-
cing a negatively charged aspartate at the equivalent position in
GluK2(G758D) markedly accelerates desensitization (Figures
6B, 6C, and 6G). We suggest that the bound zinc ions act as
a countercharge that reduces this repulsive interaction. We
tested the prediction that D730 participates in the zinc binding
site by constructing the GluK3(D730A) mutant. This receptor
was no longer potentiated but rather inhibited by zinc (33% ±
2% of control amplitude, n = 6; p = 0.02; Figures 6E–6G),
whereas the GluK3(D730N) mutant retained zinc potentiation
(Figure 6F). Therefore, the GluK3 zinc binding site is formed by
residues located on two adjacent LBDs.
The contribution of two LBD subunits to the zinc binding site
explains the potentiation of heteromeric GluK2/GluK3 receptors
by zinc (Figures 1B–1D). In principle, for receptors composed of
Figure 6. D759,H762, andD730Are Involved intheFacilitatoryEffect
of Zinc on GluK3 Receptors
(A) Sequence alignment of a segment of the extracellular S2 region for the five
KAR subunits and GluA2 reveals an amino acid that is unique to GluK3,
between a helices J and K. Residues known to form zinc binding sites in other
proteins are shown in red boxes; D represents mutated residues in (B)–(F).
(B–E) Traces illustrate the effects of 100 mM zinc on currents evoked by 10 mM
glutamate for GluK3 and GluK2 point mutants.
(F) Histogram presents the effect of 100 mM zinc on the currents evoked by
glutamate for GluK2, GluK3, and point mutants recorded in outside-out
patches. The asterisks indicate significant difference with control currents
(G) Histogram presents the desensitization kinetics of GluK2, GluK3, and point
mutants recorded in the same conditions as in (E). The asterisks indicate
significant difference with corresponding WT receptors. ***p < 0.001, **p <
0.01, *p < 0.05.
Zinc Potentiates GluK3 Receptors
572 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
two GluK2 and two GluK3 subunits, two arrangements are
possible: (1) pairs of LBD homodimers, one composed of
GluK3 containing two zinc binding sites, with no zinc binding
each containing one zinc binding site formed by residues
Q756, D759, and H762 in GluK3 and D729 in GluK2 (Figure 8C).
To distinguish between these two possibilities, we measured
the effect of zinc on receptors composed of GluK2b and
GluK3(D730A) in the presence of 1 mM UBP310 to record
primarily the activity of heteromeric receptors. Application of
zinc (100 mM) led to potentiation of currents (Figures 8D and
8E), similar to WT receptors, whereas for GluK3(D730A) mutant
homomeric dimers, zinc potentiation was abolished (Figure 6E).
This result is consistent with hypothesis (2). Moreover, in cells
transfected with GluK2b(D729A) and GluK3, zinc did not poten-
tiate currents (Figures 8D and 8E), strongly suggesting that the
zinc binding site is lost in these heteromeric receptors. Again,
this is consistent with hypothesis (2), namely that heteromeric
GluK2/GluK3 contains at least an LBD heterodimer and, if
composed of two GluK2 and two GluK3 subunits, is arranged
as a pair of heterodimers at the level of the LBDs.
Our results identify zinc as a positive allosteric modulator of
KARs containing the GluK3 subunit and provide a molecular
and mechanistic basis for this allosteric modulation. We identify
critical amino acids at the interface between the LBD of two
partner subunits that form a pocket for zinc binding. Zinc stabi-
lizes the interface by cross-bridging the two partner LBDs in
the dimer. By its action as a counter ion that reduces repul-
sion between opposed aspartate side chains, hence strongly
reducing desensitization, zinc binding translates into potentia-
tion of the GluK3 response. Our data also provide a mechanistic
and structural explanation for the specific properties of the
GluK3 subunit of KARs and reveal important information about
KAR architecture. In particular, our study provides a structural
explanation for the functional differences between the two
closely related KAR subunits GluK2 and GluK3 and about the
probable arrangement of subunits in a heteromeric GluK2/
GluK3 receptor, the only native GluK3-containing receptor iden-
tified so far (Pinheiro et al., 2007).
Zinc Acts as a Positive Allosteric Modulator of KARs
The positive allosteric modulation of KARs by zinc appears as
a specific feature of GluK3. Homomeric GluK1 and GluK2, as
well as GluK2/GluK4 and GluK2/GluK5, are inhibited by zinc in
the concentration range that potentiates GluK3 (this study and
Mott et al., 2008). The properties of GluK3, especially the fast
desensitization and low agonist sensitivity, set it apart from the
other KARs (Perrais et al., 2010; Schiffer et al., 1997). We previ-
ously showed that the properties of GluK3 are dominant over
those of GluK2 when expressed in heteromeric combinations
(Perrais et al., 2009a; Pinheiro et al., 2007). Interestingly, zinc
potentiates GluK2/GluK3 receptors as well as GluK3 receptors,
erties on the heteromer. Zinc potentiation appears to act by
Figure 7. Multiple Zinc Binding Sites in
GluK3 LBD Crystal Structures
(A) Ribbon diagram shows two protomers in the
Zinc ions are drawn as gray spheres; anomalous
difference Fourier maps contoured at 4 s are
shown for eight zinc ions labeled as Zn1–Zn8; the
side chains that coordinate zinc ions and the
kainate ligand are drawn as sticks.
(B)Enlargedview of theZn1 bindingsitecreatedby
the side chains of D759 and H762 in helix K of
protomer A, together with H492 and E495 in helix D
of protomer B, is illustrated; dashed lines indicate
intermolecular contacts for the zinc ion with dis-
tances shown in angstroms.
(C) Sequence alignment of a section of the extra-
cellular S1 segment for GluK1, GluK2, GluK3, and
GluA2 reveals an amino acid that is unique to
GluK3 in a helix D. Residues known to be involved
in the binding of zinc are shown in red boxes; D
indicates the mutated residue in a helix D.
(D) Traces illustrate the effects of 100 mM zinc
on currents evoked by 10 mM glutamate for
(E) Traces illustrate the effects of 100 mM zinc
on currents evoked by 10 mM glutamate for
(F) Histogram presents the effect of 100 mM zinc on the currents evoked by glutamate for GluK2, GluK3, and point mutants. The asterisks indicate significant
difference with control currents without zinc.
(G) Histogram presents the desensitization kinetics of GluK2, GluK3, and point mutants recorded in the same conditions as in (E). The asterisks indicate
a significant difference with corresponding WT GluK3 or GluK2 receptors. ***p < 0.001, **p < 0.01, *p < 0.05.
See also Figure S1.
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 573
reducing the desensitization rate of GluK3. To confirm this link
between reduced desensitization and potentiation of the GluK3
response, we showed that the effects of zinc on GluK3 are
abrogated in GluK3 variants with reduced desensitization rates.
Inadditionto itspotentiating effect, zincatconcentrations above
300 mM for WT GluK3 receptors and for many mutant GluK3
receptors (Table S1) clearly inhibits currents. Similar to GluK2,
this inhibition is not accompanied by a change in desensitiza-
tion kinetics, suggesting that these two effects of zinc rely
on different mechanisms and, hence, different binding sites.
Interestingly, we did not see inhibition of heteromeric GluK2/
GluK3 receptors; moreover, when potentiation is abolished in
the GluK2(D729A)/K3 heteromeric receptor, zinc also does not
induce any inhibition. This suggests that inhibitory zinc binding
sites are screened or absent in heteromeric receptors.
There is a growing number of positive allosteric modulators
related to aniracetam and cyclothiazide that bind to the LBD
dimer assembly of structurally related AMPARs and that poten-
tiate activity through modification of the deactivation and/or
Figure 8. A Model for GluK3 Zinc Binding
Sites Based on GluK2 LBD Crystal Struc-
(A) Ribbon diagram and molecular surface of
a GluK3 model dimer in which the pair of subunits
was arranged by least-squares superposition
on the crystal structure of a GluK2 LBD dimer
assembly, illustrating the location of the binding
sites for two Zn2+(D730, Q756, D759, and H762),
two Na+(E495, I498, and D499), and one Cl?
(K502, R745, T749).
(B) Enlarged view of one of the pair of zinc binding
sites, showing coordination of the zinc ion by the
main chain of Q756, and the side chains of D730,
D759, and H762.
(C) Possible arrangements of heteromeric LBD
dimers on a KAR composed of two GluK2 and two
GluK3 subunits are demonstrated: homodimers in
a, and heterodimers in b. Purple spheres show the
predicted location of zinc binding sites.
(D) Effect of zinc on currents evoked by gluta-
mate (10 mM) in a cell expressing GluK2b and
GluK3(D730A) (top) and in a cell expressing
GluK2b(D729A) and GluK3 (bottom). In the former,
the potentiation by zinc is preserved, whereas in
the latter, the potentiation is lost.
(E) Summary graph of the effect of zinc on het-
eromeric mutant receptors is presented. ***p <
0.001, **p < 0.01, ns, p > 0.05.
desensitization time course (Traynelis
for KARs, for which only concanavalin
A (ConA) and a few other plant lectins
have been identified as positive allosteric
modulators, although not of GluK3 (Per-
rais et al., 2009a; Schiffer et al., 1997).
ConA appears to reduce desensitization
of KARs and increase the apparent
agonist affinity (Partin et al., 1993; Bowie
et al., 2003). There are however clear
differences with the mode of action of zinc, including the fact
that lectins stabilize different KAR open states (Bowie et al.,
2003) and bind to carbohydrate chains in the ATD (Everts
et al., 1999). Moreover, although monovalent ions, such as Na+
and Cl?, also regulate the gating of KARs, but not AMPA or
NMDARs (Bowie, 2010), they act as necessary cofactors for
KAR function (Figure 8A; Plested and Mayer, 2007; Plested
et al., 2008). Finally, protons typically inhibit the function of
KARs (Mott et al., 2003; 2008). Here, we show that for GluK3
receptors, desensitization is instead increased at pH 8.3, sug-
gesting that protonation of dimer interface residues stabilizes
the LBD dimer assembly. In addition, zinc potentiation is lost at
pH 6.8, which suggests that the zinc binding site itself can be
protonated, which is likely the case for the key histidine H762.
Zinc Binds GluK3 at the Dimer Interface and Stabilizes
We took advantage of the opposite modulation of GluK2 and
GluK3 by zinc to narrow down the region responsible for zinc
Zinc Potentiates GluK3 Receptors
574 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
binding. By studying chimeric GluK2/GluK3 receptors and point
mutants of the two subunits, we identified key residues in the S2
segment of GluK3 that are responsible for the potentiating
effects of zinc. Strikingly, the single mutation D759G in GluK3
reverts zinc potentiation into an inhibition, and the converse
mutation in GluK2 imparts potentiation by zinc. In addition to
D759, which is unique to GluK3, the binding site for zinc is
composed of a carbonyl oxygen from the main chain and two
conserved residues: H762 in the same subunit as D759, and
D730 in the dimer partner. Prior analysis of the effects of muta-
tions at the LBD dimer interface has confirmed that there
is a common mechanism for desensitization in AMPA/KARs,
dependent on the stability of the LBD dimer interface (Weston
bilization of the dimer interface by electrostatic repulsion (Fig-
ure S1D), generating fast desensitization properties. The binding
ofzinctothedimerinterface cancels thisrepulsionandstabilizes
the LBD dimer. Consistent with this, GluK3(D759G) desensitizes
much more slowly, whereas the converse mutant GluK2(G758D)
desensitizes very rapidly. However, mutation of the other aspar-
tate in the zinc binding site, D730, did not yield receptors with
reduced desensitization: for GluK3(D730A), desensitization is
similar to WT, and for GluK3(D730N), it is even faster than WT
(Table S1). This unexpected effect could be due, for example,
between LBDs. The presence of Asp759 in the D730A mutant
would cancel this effect. Alternatively, structural changes in the
mutant receptors could complicate the interpretation. Similar
results have been reported for some GluK2 LBD dimer interface
mutants, for which the GluK2(E757Q) mutant, which swaps
a GluA2 for GluK2 residue, increases desensitization (Chaudhry
et al., 2009), most likely by subtly perturbing the structure of
Because D730 is conserved between GluK3 and GluK2,
it provides an explanation why zinc potentiates heteromeric
GluK2/GluK3 receptors, with the zinc binding site partitioning
between the two subunits in the dimer. Our structural model
suggests that there is only one zinc binding site in a heteromeric
LBD dimer (see Figure 8C). Consistent with this, GluK2/GluK3
receptors have a higher EC50and lower nH for zinc than homo-
meric receptors. Moreover, the analysis of mutant heteromeric
receptors shows that zinc binding requires Asp729 on the
GluK2 subunit. Consequently, the zinc binding site is most
probably shared by GluK2 and GluK3 in LBD heterodimers
and therefore, we propose that other combinations of hetero-
meric receptors containing GluK3 could all comprise a zinc
binding site leading to potentiation. Moreover, the GluK3 speci-
ficity of potentiation by zinc provides structural insights into the
specific gating and desensitization properties of GluK3. Notably,
in the LBD of 14 out of 18 iGluR subunits, with the exception of
GluK3, GluK4, GluK5, and GluN2D, the turn between helices J
and K contains a highly conserved glycine residue. In GluK3,
this glycine residue is replaced by D759, producing unique rapid
desensitization, whereas in GluK4 and GluK5, the asparagine
substitution at this position would likely not destabilize the LBD
dimer interface, imparting different gating properties to these
Zinc can modulate excitatory synaptic transmission through
multiple mechanisms, which are not all well described (Paoletti
et al., 2009). Although the inhibitory regulation of postsynaptic
glutamate receptors, principally NMDARs, appears as a primary
function of synaptic zinc, other potential roles in the regulation of
synaptic transmission have also been proposed by Paoletti et al.
(2009). Technical limitations have yet precluded a direct
measurement of zinc in the synaptic cleft. However, the peak
concentration was initially estimated to be in the order of
100 mM (Vogt et al., 2000). This value is well within the range
of efficacy for the allosteric potentiation of GluK3 by zinc but
may be overestimated and may depend on experimental condi-
tions (Paoletti et al., 2009). Moreover, simultaneous application
of zinc and glutamate does not potentiate GluK3-mediated
currents (data not shown), which likely excludes an effect of
zinc during low-frequency stimulation. However, high-frequency
trains of synaptic stimulation are thought to trigger a substan-
tial increase in extracellular zinc, and the accumulated zinc
could potentiate presynaptic GluK2/GluK3 receptors present at
hippocampal mossy fiber synapses (Pinheiro et al., 2007). Inter-
estingly, it has been shown that vesicular zinc is required for
presynaptic LTP at hippocampal mossy fiber synapses by a
yet undisclosed mechanism (Pan et al., 2011). This result can
be correlated with the fact that mossy fiber LTP is absent in
GluK3?/?mice (Pinheiro et al., 2007). The positive allosteric
modulation of these presynaptic GluK2/GluK3 receptors may
impart increased sensitivity to glutamate and prolonged channel
opening, inducing a possible increase in presynaptic Ca2+influx.
Hence, allosteric modulation of presynaptic GluK3 receptors
may be one of the mechanisms by which zinc promotes presyn-
aptic long-term potentiation.
In conclusion, we have identified zinc as a positive allosteric
modulator of presynaptic KARs, with a potential role in synaptic
the notion that the stability of the LBD dimer interface is essential
for dictating the desensitization properties of KARs. Our data
help explain the fast desensitization properties of GluK3 as
compared to GluK2 and pinpoint a single amino acid residue in
the upper lobe of the clamshell of the GluK3 LBD, D759, as
responsible for the specific properties of GluK3. Finally, the
structural data about the site for positive allosteric modulation
synaptic plasticity and cognitive processes by acting on the
Clones corresponding to rat GluK2a(Q), GluK2b(Q), and GluK3a were
corresponding to GluK2e/3i and GluK3e/2i were described in Perrais et al.
(2009a). Site-directed mutagenesis was performed using QuikChange XL
Site-Directed Mutagenesis Kit (Stratagene) with specific oligonucleotides cor-
responding to each mutant (oligonucleotide forward sequences, primed on
their respective WT subunits: GluK3(D759G), 50-GCTGCAGGAGGAGGGCAA
GCTTCACATCATGAAGGAG-30; GluK3(H762A), 50-GCTGCAGGAGGAGGAC
AAGCTGGCCATCATGAAGGAGAAGTGGTGGC-30; GluK3(D730A), 50-CGG
CGGCCTCATCGCTTCCAAGGGCTACGG-30; GluK3(D730N), 50-AACTGCAA
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 575
CGCCCATGG-30; GluK3(H492Y), 50-GGCTCCCCTGACCATCACATATGTCC
GAGAGAAGGCC-30; GluK3(H492A), 50-CCCCTGACCATCACCGCGGTCCGA
GAGAAGGCC-30; GluK2(Y490H), 50-GCTCCACTGGCTATAACCCACGTGCG
TGAGAAGGTCATCG-30; GluK2(G758D), 50-CAGCTGCAGGAGGAAGACAA
GCTTCAC ATGATGAAGGAG-30; GluK2b(D729A), 50-ATTGGCGGCCTTATA
N-terminal deletion versions of GluK2a(Q) and GluK3a(Q) were generated
using restriction site addition by PCR before the amino acid sequence GRI
at positions 344 and 347, respectively: GluK2 DATD: forward 50-GAATTCCGG
CAGAATAACATTT AACAAAACCAATGG-30, reverse 50-CACCAAATGC CTC
CCACTATC-30primed on GluK2a ; GluK3 DATD: forward: 50-GAATTCAG
GACGGATTGTTTTCAACAAAACCAGTGGC-30, reverse 50-GGCTTAGAGAA
GTCAATGGCCTCCTCTCGGACATGGG-30primed on GluK3a.
The GluK2/3S2 and GluK3/2S2 chimeras were constructed by exchange of
the C-terminal domain of one GluK subunit to the other one using the EcoRV
restriction site already present on GluK2 in position 532 on the first transmem-
brane domain and one created by mutagenesis into the same amino acid
sequence in position 534 on GluK3 with a forward primer of 50-CCCCC
TGTCCCCAGATATCTGGATGTACGTGC-30. All DNAs were verified by restric-
tion analysis and sequencing.
Cell Culture and Transfection
HEK293 cells (European Collection of Cell Cultures) were grown and trans-
fected as previously described by Coussen et al. (2005). Cells were cotrans-
fected using FuGENE 6 (Roche) with green fluorescent protein (GFP) and
GluK2a(Q), GluK3a, or mutated forms of these receptors at a cDNA ratio of
1:3. For experiments on heteromeric GluK2/GluK3 receptors, cells were trans-
fected withGFP, GluK2b(Q),and GluK3a ataDNA ratio of 1:1.5:1.5.Cells were
used 1–3 days after transfection and replated the day before recording for
lifted cells, or 1–2 days before recording for outside-out patches.
Brightly fluorescent, isolated cells were selected for recording. Cells were
bathed in a solution containing the following: 145 mM NaCl, 2 mM KCl,
2 mM MgCl2, 2 mM CaCl2, 10 mM glucose, and 10 mM HEPES, adjusted to
320 mOsm/l and pH 7.4 with NaOH, at room temperature. Recording pipettes
(resistance 3–5 MU) were filled with a solution containing the following:
130 mM CsCH3SO3, 2 mM NaCl, 2 mM MgCl2, 10 mM EGTA, 10 mM HEPES,
4 mM Na2ATP, and 0.1 mM spermine, adjusted to 310 mOsm/l, and pH 7.2
with CsOH. Cells were recorded at room temperature (22?C–25?C), in
whole-cell or outside-out patch mode, held at ?80mV to ?40mV, and placed
under the flow of a theta tube pulled to a final opening of ?100 mm mounted on
a piezoelectric translator (P-245.50 and E-470 amplifier; Polytec PI). Currents
were evoked by long (100 ms) or short (1 ms) applications of glutamate
every 20 s and were filtered at 2.9 kHz and recorded at a sampling frequency
of 20 kHz by an EPC10 amplifier (HEKA). Up to four different glutamate
concentrations, or three zinc concentrations, were applied to a cell or an
outside-out patch with a manual valve. Zinc (ZnCl2) was added in both control
and glutamate solutions. The exchange between two different concentrations
was completed within 2 min. All chemicals were from Sigma-Aldrich. UBP310
was from Tocris.
All electrophysiological recordings were analyzed with IGOR Pro 5
(WaveMetrics). Current amplitudes were measured with built-in tools, and
tdeswas measured with exponential fit using a least-squares algorithm. For
each condition, we averaged five sweeps and corrected amplitude changes
for run down. Statistical analysis was performed using GraphPad (Prism). In
the text and figures, data are presented as mean ± SEM; Student’s t test was
used for assessment of difference, with the following coding: ***p < 0.001,
The GluK3 LBD construct was created by PCR from the full-length cDNA and
contained residues N402-K515 in S1 connected via a GT linker to P638-E776
in S2, with an N-terminal His8 affinity tag and LVPRGS thrombin site. The
GluK3 LBD was expressed as a soluble protein in Origami B(DE3) E. coli and
purified as described previously for other KAR LBD constructs (Mayer,
2005). Crystals were grown using hanging drops in the presence of either
5 mM glutamate or 4 mM kainate, together with 2–10 mM zinc acetate added
to the protein solution at 5–10 mg/ml in a buffer containing 200 mM NaCl,
complexes contained 6%–14% PEG 3350, 100 mM Bis-Tris propane (pH 8.4–
8.5), and 0.2 M Na2SO4. For the kainate complexes, the reservoir contained
4%–6% PEG 8K and 0.1 M Tris (pH 7.8). Crystals were cryoprotected by serial
tion data were collected from single crystals at 100 K at APS 22-ID. Data sets
were indexed, scaled, and merged using DENZO and SCALEPACK from the
HKL2000 suite (Otwinowski and Minor, 1997). None of the data sets showed
twinning as analyzed by PHENIX xtriage (Adams et al., 2010). For the gluta-
mate complex, two different crystal forms were obtained; the kainate complex
was isomorphous with one of the glutamate complexes (Table S2). The struc-
tures were solved by molecular replacement using Phaser (McCoy et al., 2007)
with a GluR6 LBD monomer (PDB 1S50) as a search probe. The solution found
two protomers with high rotation and translation Z scores for the glutamate
P2221(RFZ1 = 15.5, TFZ1 = 17.4; RFZ2 = 17.4 and TFZ2 = 52.4) and kainate
P2221(RFZ1 = 12.1, TFZ1 = 20.5; RFZ2 = 14.8, TFZ2 = 40.8) complexes. For
the second crystal form of the glutamate complex in the P21212 space group,
the molecular replacement solution located four protomers, also with high Z
scores (RFZ1 = 13.8, TFZ1 = 17.6; RFZ2 = 18.6 and TFZ2 = 31.4; RFZ3 =
using ARP/wARP (Morris et al., 2003) and then refined by alternate cycles of
crystallographic refinement with PHENIX (Adams et al., 2010) coupled with
rebuilding and real-space refinement with Coot (Emsley and Cowtan, 2004)
using TLS groups determined by motion determination analysis (Painter and
Merritt, 2006). The final models (Table S2) were validated with MolProbity (Da-
vis et al., 2004). Figures were prepared using PyMOL (Schro ¨dinger).
The coordinates and structure factors the GluK3 kainate, GluK3 glutamate
P2221, and GluK3 glutamate P21212 complexes have been deposited into
the Protein Data Bank under ID codes 3U92, 3U93, and 3U94, respectively.
with this article online at http://dx.doi.org/10.1016/j.neuron.2012.08.027.
This work was supported by the Centre National de la Recherche Scientifique,
the Fondation pour la Recherche Medicale, the Conseil Re ´gional d’Aquitaine,
the Agence Nationale de la Recherche (contract SynapticZinc), and the intra-
mural research program of NICHD, NIH. Synchrotron diffraction data were
collected at SER-CAT beamline 22 ID. Use of the Advanced Photon Source
was supported by the U.S. Department of Energy, Office of Science, Office
of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. We
thank Remi Sterling for cell culture maintenance, and Franc ¸oise Coussen, Se ´-
verine Desforges, and Carla Glasser for help with molecular biology. Pierre
Paoletti provided insightful suggestions along the course of this study. We
are also grateful for members of the C.M. laboratory for helpful discussions.
Accepted: August 7, 2012
Published: November 7, 2012
Adams, P.D., Afonine, P.V., Bunko ´czi, G., Chen, V.B., Davis, I.W., Echols, N.,
Headd, J.J., Hung, L.W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010).
PHENIX: a comprehensive Python-based system for macromolecular struc-
ture solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.
Zinc Potentiates GluK3 Receptors
576 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.
Armstrong, N., Jasti, J., Beich-Frandsen, M., and Gouaux, E. (2006).
Measurement of conformational changes accompanying desensitization in
an ionotropic glutamate receptor. Cell 127, 85–97.
Bowie, D. (2010). Ion-dependent gating of kainate receptors. J. Physiol. 588,
Bowie, D., Garcia, E.P., Marshall, J., Traynelis, S.F., and Lange, G.D. (2003).
Allosteric regulation and spatial distribution of kainate receptors bound to
ancillary proteins. J. Physiol. 547, 373–385.
Chaudhry, C., Weston, M.C., Schuck, P., Rosenmund, C., and Mayer, M.L.
(2009). Stability of ligand-binding domain dimer assembly controls kainate
receptor desensitization. EMBO J. 28, 1518–1530.
Choi, Y.B., and Lipton, S.A. (1999). Identification and mechanism of action of
two histidine residues underlying high-affinity Zn2+ inhibition of the NMDA
receptor. Neuron 23, 171–180.
Contractor, A., Swanson, G., and Heinemann, S.F. (2001). Kainate receptors
are involved in short- and long-term plasticity at mossy fiber synapses in the
hippocampus. Neuron 29, 209–216.
Contractor, A., Sailer, A.W., Darstein, M., Maron, C., Xu, J., Swanson, G.T.,
and Heinemann, S.F. (2003). Loss of kainate receptor-mediated heterosynap-
tic facilitation of mossy-fiber synapses in KA2-/- mice. J. Neurosci. 23,
Contractor, A.,Mulle,C., and Swanson, G.T. (2011). Kainate receptorscoming
of age: milestones of two decades of research. Trends Neurosci. 34, 154–163.
Coussen, F., Perrais, D., Jaskolski, F., Sachidhanandam, S., Normand, E.,
Bockaert, J., Marin, P., and Mulle, C. (2005). Co-assembly of two GluR6 kai-
nate receptor splice variants within a functional protein complex. Neuron 47,
Davis, I.W., Murray, L.W., Richardson, J.S., and Richardson, D.C. (2004).
MOLPROBITY: structure validation and all-atom contact analysis for nucleic
acids and their complexes. Nucleic Acids Res. 32(Web Server issue),
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
Erreger, K., Dravid, S.M., Banke, T.G., Wyllie, D.J., and Traynelis, S.F. (2005).
Subunit-specific gating controls rat NR1/NR2A and NR1/NR2B NMDA
channel kinetics and synaptic signalling profiles. J. Physiol. 563, 345–358.
Everts, I., Petroski, R., Kizelsztein, P., Teichberg, V.I., Heinemann, S.F., and
Hollmann, M. (1999). Lectin-induced inhibition of desensitization of the kainate
receptor GluR6 depends on the activation state and can be mediated by
a single native or ectopic N-linked carbohydrate side chain. J. Neurosci. 19,
Fernandes, H.B., Catches, J.S., Petralia, R.S., Copits, B.A., Xu, J., Russell,
T.A., Swanson, G.T., and Contractor, A. (2009). High-affinity kainate receptor
subunits are necessary for ionotropic but not metabotropic signaling. Neuron
Gielen, M., Le Goff, A., Stroebel, D., Johnson, J.W., Neyton, J., and Paoletti, P.
(2008). Structural rearrangements of NR1/NR2A NMDA receptors during allo-
steric inhibition. Neuron 57, 80–93.
Horning, M.S., and Mayer, M.L. (2004). Regulation of AMPA receptor gating by
ligand binding core dimers. Neuron 41, 379–388.
Jaskolski, F., Coussen, F., Nagarajan, N., Normand, E., Rosenmund, C., and
Mulle, C. (2004). Subunit composition and alternative splicing regulate mem-
brane delivery of kainate receptors. J. Neurosci. 24, 2506–2515.
Jaskolski, F., Normand, E., Mulle, C., and Coussen, F. (2005). Differential traf-
ficking of GluR7 kainate receptor subunit splice variants. J. Biol. Chem. 280,
Karakas, E., Simorowski, N., and Furukawa, H. (2009). Structure of the zinc-
bound amino-terminal domain of the NMDA receptor NR2B subunit. EMBO
J. 28, 3910–3920.
Low, C.M., Zheng, F., Lyuboslavsky, P., and Traynelis, S.F. (2000). Molecular
determinants of coordinated proton and zinc inhibition of N-methyl-D-aspar-
tate NR1/NR2A receptors. Proc. Natl. Acad. Sci. USA 97, 11062–11067.
Mayer, M.L. (2005). Crystal structures of the GluR5 and GluR6 ligand binding
cores: molecular mechanisms underlying kainate receptor selectivity. Neuron
Mayer, M.L. (2011). Structure and mechanism of glutamate receptor ion
channel assembly, activation and modulation. Curr. Opin. Neurobiol. 21,
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,
and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40,
Morris, R.J., Perrakis, A., and Lamzin, V.S. (2003). ARP/wARP and automatic
interpretation of protein electron density maps. Methods Enzymol. 374,
Mott, D.D., Washburn, M.S., Zhang, S., and Dingledine, R.J. (2003). Subunit-
dependent modulation of kainate receptors by extracellular protons and poly-
amines. J. Neurosci. 23, 1179–1188.
Mott, D.D., Benveniste, M., and Dingledine, R.J. (2008). pH-dependent inhibi-
tion of kainate receptors by zinc. J. Neurosci. 28, 1659–1671.
Mulle, C., Sailer, A., Pe ´rez-Otan ˜o, I., Dickinson-Anson, H., Castillo, P.E.,
Bureau, I., Maron, C., Gage, F.H., Mann, J.R., Bettler, B., and Heinemann,
S.F. (1998). Altered synaptic physiology and reduced susceptibility to
kainate-induced seizures in GluR6-deficient mice. Nature 392, 601–605.
Naur, P., Vestergaard, B., Skov, L.K., Egebjerg, J., Gajhede, M., and Kastrup,
J.S. (2005).Crystal structure of the kainate receptor GluR5ligand-binding core
in complex with (S)-glutamate. FEBS Lett. 579, 1154–1160.
Nayeem, N., Zhang, Y., Schweppe, D.K., Madden, D.R., and Green, T. (2009).
A nondesensitizing kainate receptor point mutant. Mol. Pharmacol. 76,
Otwinowski, Z., and Minor, W. (1997).  Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol. 276, 307–326.
terms of multiple groups undergoing TLS motion. Acta Crystallogr. D Biol.
Crystallogr. 62, 439–450.
Pan, E., Zhang, X.A., Huang, Z., Krezel, A., Zhao, M., Tinberg, C.E., Lippard,
S.J., and McNamara, J.O. (2011). Vesicular zinc promotes presynaptic and
inhibits postsynaptic long-term potentiation of mossy fiber-CA3 synapse.
Neuron 71, 1116–1126.
Paoletti, P., Ascher, P., and Neyton, J. (1997). High-affinity zinc inhibition of
NMDA NR1-NR2A receptors. J. Neurosci. 17, 5711–5725.
Paoletti, P., Perin-Dureau, F., Fayyazuddin, A., Le Goff, A., Callebaut, I., and
Neyton, J. (2000). Molecular organization of a zinc binding n-terminal modula-
tory domain in a NMDA receptor subunit. Neuron 28, 911–925.
Paoletti, P., Vergnano, A.M., Barbour, B., and Casado, M. (2009). Zinc at glu-
tamatergic synapses. Neuroscience 158, 126–136.
Partin, K.M., Patneau, D.K., Winters, C.A., Mayer, M.L., and Buonanno, A.
(1993). Selective modulation of desensitization at AMPA versus kainate recep-
tors by cyclothiazide and concanavalin A. Neuron 11, 1069–1082.
Perrais, D., Coussen, F., and Mulle, C. (2009a). Atypical functional properties
of GluK3-containing kainate receptors. J. Neurosci. 29, 15499–15510.
Perrais, D., Pinheiro, P.S., Jane, D.E., and Mulle, C. (2009b). Antagonism
of recombinant and native GluK3-containing kainate receptors. Neuro-
pharmacology 56, 131–140.
Perrais, D., Veran, J., and Mulle, C. (2010). Gating and permeation of kainate
receptors: differences unveiled. Trends Pharmacol. Sci. 31, 516–522.
Pinheiro, P.S., and Mulle, C. (2008). Presynaptic glutamate receptors: physio-
logical functions and mechanisms of action. Nat. Rev. Neurosci. 9, 423–436.
Pinheiro, P.S., Perrais, D., Coussen, F., Barhanin, J., Bettler, B., Mann, J.R.,
Malva, J.O., Heinemann, S.F., and Mulle, C. (2007). GluR7 is an essential
subunit of presynaptic kainate autoreceptors at hippocampal mossy fiber
synapses. Proc. Natl. Acad. Sci. USA 104, 12181–12186.
Plested, A.J., and Mayer, M.L. (2007). Structure and mechanism of kainate
receptor modulation by anions. Neuron 53, 829–841.
Zinc Potentiates GluK3 Receptors
Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc. 577
Plested,A.J.,Vijayan,R.,Biggin,P.C.,andMayer,M.L.(2008).Molecularbasis Download full-text
of kainate receptor modulation by sodium. Neuron 58, 720–735.
Rachline, J., Perin-Dureau, F., Le Goff, A., Neyton, J., and Paoletti, P. (2005).
The micromolar zinc-binding domain on the NMDA receptor subunit NR2B.
J. Neurosci. 25, 308–317.
Schiffer, H.H., Swanson, G.T., and Heinemann, S.F. (1997). Rat GluR7 and
a carboxy-terminal splice variant, GluR7b, are functional kainate receptor
subunits with a low sensitivity to glutamate. Neuron 19, 1141–1146.
Sobolevsky, A.I., Rosconi, M.P., and Gouaux, E. (2009). X-ray structure,
symmetry and mechanism of an AMPA-subtype glutamate receptor. Nature
Stern-Bach, Y., Bettler, B., Hartley, M., Sheppard, P.O., O’Hara, P.J., and
Heinemann, S.F. (1994). Agonist selectivity of glutamate receptors is specified
by two domains structurally related to bacterial amino acid-binding proteins.
Neuron 13, 1345–1357.
Strutz, N., Villmann, C., Thalhammer, A., Kizelsztein, P., Eisenstein, M.,
Teichberg, V.I., and Hollmann, M. (2001). Identification of domains and amino
acids involved in GLuR7 ion channel function. J. Neurosci. 21, 401–411.
Sun, Y., Olson, R., Horning, M., Armstrong, N., Mayer, M., and Gouaux, E.
(2002). Mechanism of glutamate receptor desensitization. Nature 417,
Traynelis, S.F., Wollmuth, L.P., McBain, C.J., Menniti, F.S., Vance, K.M.,
Ogden, K.K., Hansen, K.B., Yuan, H., Myers, S.J., and Dingledine, R. (2010).
Glutamate receptor ion channels: structure, regulation, and function.
Pharmacol. Rev. 62, 405–496.
Venskutonyt_ e, R., Frydenvang, K., Gajhede, M., Bunch, L., Pickering, D.S.,
and Kastrup, J.S. (2011). Binding site and interlobe interactions of the iono-
tropic glutamate receptor GluK3 ligand binding domain revealed by high reso-
lution crystal structure in complex with (S)-glutamate. J. Struct. Biol. 176,
Venskutonyt_ e, R., Frydenvang, K., Hald, H., Rabassa, A.C., Gajhede, M.,
Ahring, P.K., and Kastrup, J.S. (2012). Kainate induces various domain
closures in AMPA and kainate receptors. Neurochem. Int. 61, 536–545.
Vogt, K., Mellor, J., Tong, G., and Nicoll, R. (2000). The actions of synaptically
released zinc at hippocampal mossy fiber synapses. Neuron 26, 187–196.
Westbrook, G.L., and Mayer, M.L. (1987). Micromolar concentrations of Zn2+
antagonize NMDA and GABA responsesof hippocampal neurons. Nature 328,
Weston, M.C., Schuck, P., Ghosal, A., Rosenmund, C., and Mayer, M.L.
(2006). Conformational restriction blocks glutamate receptor desensitization.
Nat. Struct. Mol. Biol. 13, 1120–1127.
Zhang, Y., Nayeem, N., Nanao, M.H., and Green, T. (2006). Interface interac-
tions modulating desensitization of the kainate-selective ionotropic glutamate
receptor subunit GluR6. J. Neurosci. 26, 10033–10042.
Zinc Potentiates GluK3 Receptors
578 Neuron 76, 565–578, November 8, 2012 ª2012 Elsevier Inc.