The Timothy syndrome mutation differentially affects
voltage- and calcium-dependent inactivation of
CaV1.2 L-type calcium channels
Curtis F. Barrett* and Richard W. Tsien†
Department of Molecular and Cellular Physiology, Stanford University School of Medicine, Stanford, CA 94305-5345
Contributed by Richard W. Tsien, November 7, 2007 (sent for review September 6, 2007)
Calcium entry into excitable cells is an important physiological signal,
supported by and highly sensitive to the activity of voltage-gated
Ca2?channels. After membrane depolarization, Ca2?channels first
open but then undergo various forms of negative feedback regula-
tion including voltage- and calcium-dependent inactivation (VDI and
a rare yet devastating disorder known as Timothy syndrome (TS),
whose features include autism or autism spectrum disorder along
with severe cardiac arrhythmia and developmental abnormalities.
Most cases of TS arise from a sporadic single nucleotide change that
generates a mutation (G406R) in the pore-forming subunit of the
L-type Ca2?channel CaV1.2. We found that the TS mutation power-
fully and selectively slows VDI while sparing or possibly speeding the
was further substantiated by measurements of Ca2?channel gating
currents and by analysis of another channel mutation (I1624A) that
hastens VDI, acting upstream of the step involving Gly406. As high-
an absorbing inactivation process. Thus, the TS mutation offers a
unique perspective on mechanisms of inactivation as well as a
promising starting point for exploring the underlying pathophysiol-
ogy of autism.
autism ? autism spectrum disorder ? channelopathy ? mutation ?
disorders, typified by impaired social interaction and commu-
nication skills and restricted and repetitive behavior. Despite
great interest in ASD, their etiology remains largely unknown.
However, genetic evidence supports the notion that the roots of
the pathology will ultimately be uncovered at the level of cellular
and developmental neurobiology and that insights into funda-
mental mechanisms may emerge from studies of rare forms of
the disease with simple genetic origin. Accordingly, increasing
attention has been directed toward Timothy syndrome (TS), a
rare childhood disorder whose manifestations include a very
strong association with autism or ASD (P ? 1.2 ? 10?8) along
with abnormally prolonged cardiac action potentials and a
wide-ranging set of developmental abnormalities. TS was iden-
tified in 1992 (1–3), but its likely importance as an exemplar only
came into sharp focus a dozen years later, when Splawski et al.
(4) showed that the diverse symptoms of TS could be traced in
most cases to a single amino acid defect in a single protein
molecule. The mutation (a Gly-to-Arg missense mutation at
position 406) was identified in a well known signaling molecule,
the pore-forming subunit of the class C (CaV1.2) L-type Ca2?
channel. Recently, a mouse model of TS bearing the Gly-to-Arg
mutation was reported to exhibit behavioral characteristics
reminiscent of ASD.‡That TS arises from a mutation in CaV1.2
is particularly instructive because this L-type Ca2?channel is
utism and autism spectrum disorders (ASD) are a contin-
uum of debilitating and mysterious neurodevelopmental
critically important for electrical activity, development, and
What is the functional impact of the mutation at the cellular
level? When introduced into rabbit recombinant CaV1.2 chan-
nels, the TS mutation produced no obvious changes in either the
voltage dependence of channel activation or the level of channel
expression (4). However, the mutation greatly impaired the
ability of the channels to stop conducting during depolarization,
a process known generically as inactivation. These results raised
a series of questions about the basic mechanism by which the
G406R mutation affects channel function. First, because inac-
tivation of voltage-gated Ca2?channels is strongly affected by
the identity of the auxiliary ? subunit, is the effect of the TS
mutation on L-type channels equally prominent regardless of
whether the ? subunit is typical of those found in either heart or
brain? Second, how strong are the effects of the TS mutation
when studied in combination with other amino acid changes that
themselves affect inactivation? Third, L-type channels display
multiple forms of negative feedback, including voltage-
dependent inactivation (VDI), which can be studied with Ba2?
(CDI), observed with Ca2?as the permeant ion (5–7). Given
proposals that VDI and CDI share the same final common
pathway (8, 9), do G406R and other mutations affect both forms
of inactivation in the same general way? Answers to these
questions would help clarify mechanisms of inactivation and also
provide useful clues for future explorations of the higher-order
effects of the TS mutation in its primary organ targets, including
the autistic brain.
The TS Mutation Slows Inactivation of CaV1.2 Irrespective of the
Coexpressed ? Subunit. We began by examining the effect of the
TS mutation in the context of various calcium channel accessory
subunits. The ? subunit subtype varies widely among tissues
affected by TS and profoundly influences Ca2?channel inacti-
vation (8, 10). For example, ?2(predominant in heart) confers
much slower inactivation than ?1(prominent in brain). Whereas
Splawski et al. (4) studied ?2b, we chose ?2a, also found in heart,
Author contributions: C.F.B. and R.W.T. designed research; C.F.B. performed research;
C.F.B. analyzed data; and C.F.B. and R.W.T. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
*Present address: Departments of Neurology and Human Genetics, Leiden University
Medical Center, Leiden, The Netherlands.
Beckman Center, Room B105, Stanford, CA 94305-5345. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
‡Wersinger SR, Hesse RA, Badura MA, Bett GCL, Rasmussen RL, 2007 Society for Neuro-
science Annual Meeting, November 3–7, 2000, San Diego, CA, abstr. 62.6.
© 2008 by The National Academy of Sciences of the USA
February 12, 2008 ?
vol. 105 ?
no. 6 ?
because it confers the slowest inactivation of all known ?
subunits and thus provides the most extreme case for probing the
effect of the TS mutation. Slowed inactivation by ?2ais by virtue
of a palmitoylation site (unique among the ? subunits) that
supports anchoring to the plasma membrane (11, 12). Accord-
ingly, we tested whether the TS mutation might affect inactiva-
tion even in the context of the ?2asubunit. Whole-cell currents
were recorded from CaV1.2 channels coexpressed with ?2aand
?2?, using prolonged step depolarizations (Fig. 1B), with Ba2?as
the charge carrier to examine VDI; fractional current remaining
at 3 s (r3,000) was used as a metric of channel inactivation (13).
Wild-type channels exhibited only modest inactivation after 3 s
at 10 mV, as expected with the ?2asubunit (8). In the context of
the ?2asubunit, the TS mutation powerfully slowed VDI; indeed,
virtually all inactivation was abolished (Fig. 1 B and C). This
incremental effect of the TS mutation is consistent with the
cardiac manifestations of the disorder, which include a consid-
erable prolongation of the ventricular action potential as re-
flected by the long Q–T interval of the electrocardiogram (14).
We turned next to testing the effect of the TS mutation in the
context of the neuronal ?1subunit. Given that G406R hampered
VDI in a setting where inactivation was already weak (with ?2a),
it was of interest to see whether the mutation was comparably
effective when basal inactivation was relatively strong. A ?1
isotype was chosen as an exemplar of a ? subunit common in the
brain. Relative to ?2a, significantly more VDI was observed with
the ?1c subunit, yet the TS mutation once again powerfully
slowed VDI (Fig. 1 B and C). In fact, ?1crevealed the true power
reduced VDI by ?45% compared with only a ?27% reduction
with ?2a. The overall conclusion is that the TS mutation is able
to influence the development of VDI regardless of whether the
resident ? subunit confers fast or slow VDI. In subsequent
experiments, we examined inactivation in channels containing
?1c, thus providing ample dynamic range for mutations to have
The C terminus of CaV1.2 contains an IQ motif (Ile1624–Gln1625)
implicated in mediating CDI (15, 16). As shown using Xenopus
accelerated VDI (Fig. 2A). Given that Ile1624is located in the C
terminus of the channel, ?1,000 residues away in primary
we wondered whether the functional contributions of these
structural elements would be exerted at the same or different
stages in the signaling pathway leading up to inactivation.
Accordingly, we asked whether mutating these critical amino
acids in combination would lead to a summation of the effects
of the individual mutations or instead would reveal a dominant
influence of one over the other, reflecting an action further
along the signaling pathway. When examined in the context of
I1624A, the G406R mutation dominated, as the GR/IA double
mutant channels exhibited VDI and CDI profiles that were
indistinguishable from channels bearing G406R alone (Fig. 2),
which suggests that the slowing effect of the TS mutation lies
downstream of where I1624A acts to accelerate VDI. One
possibility is that the IQ motif acts as a latch to brake the
initiation of inactivation (17, 18), and even when the I1624A
mutation releases the latch, VDI remains intrinsically slowed by
the G406R mutation.
Is CDI Spared by Mutations that Slow or Accelerate VDI? A closer
examination of the inactivation data (Figs. 2 and 3A) revealed
that whether VDI is slowed by G406R or speeded by I1624A,
Ca2?still exerts a significant effect on whole-cell currents,
increasing decay in the cases of both mutations. This finding led
us to consider the possibility that CDI might proceed indepen-
dently from VDI, developing with its own kinetics regardless of
whether VDI is intact, slowed or accelerated. The approach
taken to address this possibility (Fig. 3A, Lower) is based on
experiments shown in Fig. 3B.
1 s1 s
I II IIIIV
subunit. (A) Secondary structure of the CaV1.2 ?1Csubunit. The approximate
locations of mutations G406R and I1624A are indicated. (B) Exemplar normal-
ized Ba2?currents recorded from WT and G406R channels expressed with
depolarization to ?10 mV from a holding potential of ?90 mV. (C) Summary
of fraction remaining at 3 s.*, P ? 0.001 versus WT; #, P ? 0.001 versus ?2a.
The TS (G406R) mutation slows VDI irrespective of the coexpressed ?
or Ca2?(B). Step depolarizations (400-ms) were applied to ?10 mV from a
holding potential of ?90 mV. Shown are average traces ? SEM. (Right)
G406R slows current decay for both WT and I1624A channels regard-
www.pnas.org?cgi?doi?10.1073?pnas.0710501105Barrett and Tsien
VDI but Not CDI Causes Gating Charge Immobilization. To disentan-
gle VDI and CDI and to test whether these processes may be
partially or largely independent, we turned to measurements of
gating charge movement. Conventional VDI of ion channels
involves a transition into an absorptive state in which the gating
charge becomes immobilized. A classic example of this mecha-
nism is the ball-and-chain model of Shaker potassium channel
inactivation, in which an N-terminal particle (the ‘‘ball’’) binds
to the pore of the channel (19–23). A similar mechanism, only
using the intracellular loop connecting the first and second
domains of the channel (the I–II loop), has been proposed to
mediate inactivation of L-type Ca2?channels (8, 24, 25). If CDI
of L-type Ca2?channels were mediated by the same effector
mechanism as VDI, we would expect both processes to cause
gating charge immobilization. However, using a photolabile
chelator, Hadley and Lederer (26; but see also 27) found that
uncaging of intracellular Ca2?caused CDI but not gating charge
We reexamined the question of whether CDI causes charge
immobilization by using an approach geared to CDI produced
physiologically, by Ca2?influx through the channel itself. To
dissociate gating charge movement from ionic current, gating
currents were isolated by a sudden depolarization to the reversal
potential (Erev), where ionic current is zero. This method was
validated by bath application of the pore blocker La3?(28–30).
In the continued presence of 10 mM Ca2?, 5.2 mM La3?reduced
the peak ionic current by ?90% but was without effect on gating
currents elicited at Erev[supporting information (SI) Fig. 6].
To look for a possible effect of CDI on gating charge
immobilization, we used a multipulse protocol (Fig. 3B). First, a
gating current was elicited by stepping to Erevfor 10 ms (Ig1),
after which the membrane was stepped to ?10 mV for 100 ms
to permit ion flux. After a 20-ms step back to ?90 mV (of
sufficient length to close the channels yet brief enough to allow
only minimal recovery from inactivation), a second gating
current was elicited at Erev(Ig2). The ratio of the second gating
current to the first (Ig2/Ig1) provided a measure of gating charge
recorded in Ba2?to express VDI, then with the Ba2?replaced
by Ca2?to drive CDI. As illustrated by representative current
traces (Fig. 3B) and pooled data (Fig. 3C), the extent of the
consistent with CDI; the disparity in the fraction of peak current
remaining after 100 ms (r100) was ?2-fold (Fig. 3C). Notably,
however, the ratio of gating charge movements was not signif-
icantly affected by prior entry of Ca2?. Thus, Ca2?influx
the immobilization of gating charge.
VDI and CDI Treated as Proceeding with Independent Probabilities.
Our gating current measurements suggested that CDI might be
mediated by an effector mechanism distinct from VDI. It was
therefore appropriate, at least approximately, to treat them as
separate processes, each with its own probability of allowing the
channel to conduct (Pconduct). Moreover, the probability that the
channel is not inactivated (and hence conducting) at any given
of hVand hCa(which are the probabilities that the channel has
not undergone VDI or CDI, respectively). Thus, during the
decaying phase of whole-cell Ca2?current,
whereas with Ba2?as the charge carrier, hCa? 1, and therefore,
To a good approximation, hCathen can be determined by taking
the ratio of ICa to IBa, measured within the same cell, thus
providing a useful way to measure CDI in isolation. It is
important to note that CDI is frequently expressed as the
difference between VDI and CDI (16, 31–34). However, such an
approach can potentially fail to account for changes in VDI. For
example, when VDI is accelerated (e.g., by the I1624A muta-
tion), simply subtracting CDI from VDI can underestimate the
effects of Ca2?on inactivation. To avoid this complication, we
the course of a 400-ms depolarization (Fig. 3A).Expressed in this
way, the ICa/IBaratio (hCa) of wild-type channels decays by ?50%
over the course of a 400-ms test depolarization. Using this
metric, we found that the extent of CDI was similar in channels
bearing either the G406R or I1624A mutation.
Importantly, the ratio for G406R channels decayed signifi-
cantly faster than wild type (WT): when the ICa/IBaratio plots
were fitted with single exponentials, we found that the time
constant averaged 55 ? 8 ms for G406R compared with 142 ?
18 ms for WT (P ? 0.005). Thus, CDI was not only intact in
G406R channels but proceeded more rapidly than in WT
channels. CDI measured for I1624A channels decayed at a rate
similar to WT channels (105 ? 14 ms; P ? 0.4 vs. WT).
The G406R Mutation Reveals That CDI Cannot Be an Absorptive Form
of Inactivation. The large extent to which the G406R mutation
slowed VDI provided a fresh opportunity to study CDI with
relatively little interference from VDI (that is, with hV? 1). Our
gating current measurements indicate that CDI is not associated
0 200 400
Time (msec)Time (msec)
0 200 400
Exemplar normalized Ba2?and Ca2?currents recorded at ?10 mV from WT,
Shown are average plots ? SEM; the data were not significantly different as
determined with either ANOVA or unpaired Student’s t test. (B) Exemplar
Summary of inactivation and gating charge immobilization with either Ba2?
or Ca2?. For every cell recorded, Erevwas determined empirically; with Ba2?or
Ca2?as the charge carrier, Erev was 59.4 ? 1.4 mV and 81.9 ? 2.0 mV,
respectively.*, P ? 0.001 versus Ba2?. N.S., not significant.
CDI is spared by mutations that slow or accelerate VDI. (A) (Upper)
Barrett and TsienPNAS ?
February 12, 2008 ?
vol. 105 ?
no. 6 ?
fundamental respects. To test whether CDI is an absorbing event,
proceeding unidirectionally at inactivating membrane voltages, we
again turned to prolonged (3-s) depolarizations and measured
currents by using either Ba2?or Ca2?as the charge carrier. As
already shown in Fig. 1, the TS mutation dramatically slowed
charge carrier, G406R currents exhibited a rapid decay to ?50%
amplitude, and this decay was followed by a very slowly decaying
plateau (Fig. 4A, Right). Based on the gating current-based logic as
described earlier, we analyzed the overall current decay as the
product of two terms, one for CDI and the other for VDI (see
Materials and Methods). For simplicity, each of the factors was
described as a partially decaying exponential, as in the classical
description of sodium channel inactivation (35).
The decaying phase of the Ca2?currents in Fig. 4A were well
described by such an equation, as illustrated by exemplar fits
(Fig. 4B); the fitting parameters are summarized in Fig. 4C. The
TS mutation greatly slowed the slow time constant (?slow), as
expected for slowed VDI, whereas the fast time constant (?fast)
was not slowed at all, but in fact became ?2-fold faster (124 ?
14 ms vs. 54 ? 4.7 ms). When analyzed in this way, the extent of
CDI was generally unaffected by the TS mutation. The fraction
of G406R channels remaining noninactivated after CDI (hCa,??
1 ? Afast) was ?50%, which underscores the idea that CDI
cannot be regarded as an absorptive process, proceeding toward
a steady-state value of zero, but reflects instead a reversible
balance between two conditions of different open probability
(Fig. 4D, and see Discussion).
The TS mutation G406R is an exemplar of a disease-causing
missense mutation (channelopathy) that not only provides in-
sight into the workings of a key signaling molecule but also offers
a toehold in understanding a complex human disorder (36). This
recurrent de novo mutation in CaV1.2 L-type Ca2?channels
produces an unusually wide range of phenotypic effects on
multiple organs (4). In two of these systems, heart and brain,
excitable cells rely on L-type Ca2?channels for important
aspects of development and physiology (37, 38). Our experi-
ments focused on altered properties of L-type channel currents
and also support a bottom-up approach to higher-order dysfunc-
tion, including arrhythmia and autism, that arise from the single
Building on the work of Splawski et al. (4), who described
effects of G406R on VDI, we used G406R and other structural
modifications to probe the basis of VDI and CDI. How does the
TS mutation match up with other structural modifications that
affect inactivation of L-type channels? Do VDI and CDI share
a common final step, as often suggested? We approached these
questions by recording both Ba2?and Ca2?currents in the same
cells, under patch-clamp recording conditions where ionic and
gating currents could be isolated and VDI and CDI could be
appropriately dissected. We found that G406R exerted powerful
effects on inactivation of expressed CaV1.2 channels: the TS
mutation greatly slowed VDI while speeding the kinetics of CDI.
This slowing of VDI was observed irrespective of whether the
auxiliary ? subunit was typical of heart or brain. Furthermore,
VDI caused immobilization of gating charge, whereas CDI did
not, even when evoked in a physiological manner, by Ca2?entry
through the L-type channel itself. Our data support the idea that
CDI does not reflect an absorptive transition to an inactivated
state but arises instead from a Ca2?-dependent transition to a
gating mode with decreased Po, for example, by switching from
mode 1 to mode calcium gating (39), or from mode 2 to mode
1 gating (40).
Our results provide a fresh perspective on earlier studies of
L-type channel inactivation approached with other structural
interventions. VDI is slowed in chimeric L-type Ca2?channels
in which the I–II loop of host CaV1.2 subunits is replaced by the
corresponding region of CaV1.1 channels that lack CDI (41). In
contrast, CDI (at least as we have defined it here) was hardly
affected in those chimeric channels, suggesting that the region
mediating VDI (the I–II loop) may not play a role in mediating
CDI. Likewise, the recordings of Cens et al. (8) for Cav1.2
expressed with different ? subunits showed a dramatic effect of
? subunit identity on the time course of inactivation with Ca2?
as the charge carrier. However, if one allows for the idea that
VDI can proceed unchecked even while Ca2?permeation takes
place, the extra degree of inactivation that is specifically Ca2?-
dependent (CDI) appears unchanged in their data.§Thus, we
contend that both switching of ? subunits and swapping of the
I–II loop (a region of known importance for ? subunit interac-
tions) provide support for a dissociation between VDI and CDI,
consistent with the effects of G406R presented here. Taken
together, all of these structurally based interventions reinforce
§The authors arrived at a different conclusion because they used ‘‘CDI’’ to denote inacti-
vation (decay of whole-cell current amplitude) in the presence of Ca2?rather than the
extra decay that is specifically Ca2?-dependent.
Ba2?or Ca2?currents were elicited at ?10 mV and normalized to the peak
inward current. Shown are average currents ? SEM. (B) Normalized Ca2?
and the solid lines are fits to the product of two exponentials (see Materials
and Methods). (C) Summary of the fitting parameters for WT (solid bars) and
G406R (hatched) Ca2?currents, as in B.*, P ? 0.01 vs. WT. (D) Gating scheme
showing transitions through the various proposed states. C, closed; O, open;
I, inactivated. In Ba2?, channels predominantly display mode 1 gating fol-
lowed by transitioning to the inactivated state (VDI). With Ca2?as the charge
carrier, open channels rapidly transition to mode Ca (39, 40), displaying
The G406R mutation slows VDI and accelerates CDI. (A) Whole-cell
www.pnas.org?cgi?doi?10.1073?pnas.0710501105Barrett and Tsien
the earlier notion that VDI and CDI are mediated by distinct
effector mechanisms (5, 26).
Further analysis was performed on the effects of a previously
described mutation, I1624A, in a region of the CaV1.2 ?1subunit,
the C-terminal IQ motif, where partial structural information is
available. Like the TS mutation, I1624A altered VDI, but in this
case by acceleration rather than slowing. Once again, CDI was
dissected as a multiplicative factor and was found to be essentially
spared. For purposes of epistasis analysis, the effects of I1624A
were also studied in combination with G406R. Our results indicate
that the influence on VDI by I 3 A occurs at least one step
upstream to that affected by G 3 R. Our findings on the effects of
? subunit switching, I1624A and G406R, on VDI are summarized
in diagram form in Fig. 5 (see legend for details).
Our data in HEK293 cells revealed several key differences
from previous findings in Xenopus oocytes (15, 17), although the
same constructs for WT and mutant CaV1.2 channels were used.
In the oocyte recordings, I1624A Ca2?or Ba2?currents decayed
with a similar time course, and thus the channels appeared to
lack CDI altogether. In contrast, when expressed in human
embryonic kidney (HEK) cells, we found that I1624A channels
for this discrepancy is that in oocytes, CDI was obscured by a
robust Ca2?-dependent facilitation (CDF) that occurred even
during a single depolarizing pulse,¶which may have been less
problematic in the mammalian cells studied here; I1624A chan-
nels showed considerably less CDF in HEK cells than in oocytes
when evoked by pulse trains (data not shown).
Our findings with I1624A mutant channels help provide
insight into the role of the IQ motif in mediating VDI and CDI.
Consistent with previous results (15, 17), we found that the
I1624A mutation accelerates VDI. However, unlike previous
results and as described above, in our hands this accelerated VDI
fails to occlude CDI, consistent with biochemical (17) and
structural (42, 43) data, suggesting that the interaction of the
channel with calmodulin, a Ca2?-dependent associated protein
required for CDI (15), is unaffected by the mutation.
In summary, our evidence supports the notion that VDI and
CDI occur with independent probabilities, likely through inde-
pendent mechanisms. We propose that a shift to low-Pogating,
and not an absorptive inactivated state, underlies CDI. This
alternate way of interpreting CDI can be helpful in trying to
understand channelopathies that affect inactivation of Ca2?
channels. Such mutations include congenital stationary night
blindness type 2 in CaV1.4 (44) as well as TS in CaV1.2 (4, 45).
Splawski et al. (4) presented compelling evidence for how the
classical TS mutation, and an additional mutation that can
generate a related disorder known as TS2 (45), can lead to
prolonged cardiac action potentials, excessive Ca2?entry, and
arrhythmias involving Ca2?overload. This scenario is supported
by the finding that the L-type Ca2?channel antagonist verapamil
decreased ventricular arrhythmia in a patient with TS (46) and
a recent report that the L-type agonist Bay K 8644 recapitulates
many of the cardiac features of TS (47). Compared with ar-
rhythmias, understanding autism from the ground up will likely
be even more challenging. The use of agents like Bay K 8644 may
serve as pharmacological mimics of the TS mutation by pro-
moting an increase in L-type Ca2?entry and possibly a shift in
gating mode as well (48), a possibility not excluded by our
findings on VDI. It would be interesting to test the effects of such
pharmacological interventions on brain circuits of possible sig-
nificance to ASD.
Materials and Methods
in mammalian cells. The TS mutation G406R was introduced by mutating the
codon GGA to AGA by using the QuikChange II XL site-directed mutagenesis kit
terminus of the I1624A cDNA into the G406R construct.
Cell Culture. HEK293 cells stably expressing the calcium channel auxiliary ?1c
and ?2?-1 subunits were cultured and transfected as described in ref. 30 by
using either Lipofectamine 2000 (Invitrogen) or calcium phosphate precipita-
tion (Clontech). Cells were transfected with either WT or mutant ?1C,77con-
structs together with pEGFP-N3, and currents were recorded 24–48 h after
transfection. For the data in Fig. 1, tsA201 cells were transfected with WT or
and either ?2aor ?1c(kind gifts from Gerald Zamponi, University of Calgary,
3 B and C were generated from a stable HEK293 cell line expressing WT ?1C,77
with an IRES-EGFP, ?1cand ?2?-1.
using an Axopatch 200B patch-clamp amplifier (Molecular Devices). Borosili-
cate glass capillaries were pulled in a model P-87 or P-97 puller (Sutter
Instruments) and heat-polished before use. Pipette resistance was ?2–3 M?
when filled with an internal solution consisting of 122 mM Cs-Asp, 10 mM
¶Although CDF is most often seen as a progressive increase in Ca2?currents during trains
This was most obvious in CaV1.2 subunits bearing the IQ/AA double mutation, where CDF
in Ca2?than in Ba2?), consistent with the principle that CDF can obscure CDI.
WT + β1c
IA + β1c
GR/IA + β1c
WT + β2a
GR + β2a
GR + β1c
C-terminal tail of the ?1subunit are depicted; other regions have also been
reported to contribute to VDI (52–54). Approximate positions of Gly406and
Ile1624are shown by green and red circles, respectively. Shown below are
idealized whole-cell Ba2?currents. VDI of CaV1.2 L-type calcium channels is
(Middle) Replacing ?2awith the brain ?1subunit speeds up VDI. (Bottom) The
IQ motif appears to serve as a Ca2?-independent brake against VDI, possibly
by association with the hinged lid of the I–II loop. The I1624A mutation may
disrupt this interaction, thereby accelerating VDI. Under all conditions, the
effect of the TS mutation (G406R, dashed box) is to slow VDI powerfully.
Proposed model to account for our findings with VDI. For simplicity,
Barrett and TsienPNAS ?
February 12, 2008 ?
vol. 105 ?
no. 6 ?
Hepes, 10 mM EGTA, 5 mM MgCl2, 4 mM ATP, 0.4 mM GTP, pH 7.5. The bath
solution contained 155 mM NMDG-Asp, 0.1 mM EGTA, 10 mM Hepes, 10 mM
BaCl2or CaCl2, pH 7.4. Series resistance (8.95 ? 1.1 M?) was compensated
electronically by ?90%, and membrane capacitance (19.6 ? 0.9 pF) was
corrected online; where applicable, residual linear capacitive and leak cur-
rents were subtracted by the ?P/4 method.
EGFP-positive cells were visualized by epifluorescence and selected for
recording. Cells were voltage-clamped at ?90 mV, and pulse depolarizations
Bessel filter at 1–10 kHz, digitized at 5–100 kHz with a Digidata 1320A
(Molecular Devices), and stored on a personal computer.
Data Analysis. Data were acquired and analyzed with pClamp 8.2 (Molecular
Devices). Summary data are presented as mean ? SEM, and n ? 4–12 cells per
condition. Statistical significance was tested by using a two-tailed Student’s
unpaired t test, except where indicated.
The fits in Fig. 4 were calculated by using the equation:
I ? ??1 ? Afast? ? Afaste?t/?fast? ??1 ? Aslow? ? Aslowe ?t/?slow?
where I is normalized current amplitude, t is time in ms, Afastand Asloware the
fast and slow amplitudes, respectively, and ?fastand ?sloware the fast and slow
time constants, respectively.
ACKNOWLEDGEMENTS. We thank Roger Zu ¨hlke and Harald Reuter for pro-
viding the wild-type and I1624A cDNAs, Gerald Zamponi for providing acces-
sory subunit cDNAs, and Harald Reuter and Damian Wheeler for helpful
discussions. This work was supported by National Heart, Lung, and Blood
Training Grant 5T32HL007708-14 (to C.F.B.) and National Institutes of Health
Grants 5R01NS024067-22 and 5R01GM058234-08 (to R.W.T.).
1. Reichenbach H, Meister EM, Theile H (1992) Kinderarztl Prax 60:54–56.
2. Marks ML, Trippel DL, Keating MT (1995) Am J Cardiol 76:744–745.
3. Marks ML, Whisler SL, Clericuzio C, Keating M (1995) J Am Coll Cardiol 25:59–64.
4. Splawski I, Timothy KW, Sharpe LM, Decher N, Kumar P, Bloise R, Napolitano C,
Schwartz PJ, Joseph RM, Condouris K, et al. (2004) Cell 119:19–31.
5. Lee KS, Marban E, Tsien RW (1985) J Physiol (London) 364:395–411.
6. Cens T, Rousset M, Leyris JP, Fesquet P, Charnet P (2006) Prog Biophys Mol Biol
7. Budde T, Meuth S, Pape HC (2002) Nat Rev Neurosci 3:873–883.
8. Cens T, Restituito S, Galas S, Charnet P (1999) J Biol Chem 274:5483–5490.
9. Findlay I (2004) J Physiol (London) 554:275–283.
10. SatherWA,TanabeT,ZhangJF,MoriY,AdamsME,TsienRW(1993) Neuron11:291–303.
11. Chien AJ, Zhao X, Shirokov RE, Puri TS, Chang CF, Sun D, Rios E, Hosey MM (1995) J Biol
13. Barrett CF, Rittenhouse AR (2000) J Gen Physiol 115:277–286.
14. Rosen MR (2002) Circulation 106:2173–2179.
15. Zu ¨hlke RD, Pitt GS, Deisseroth K, Tsien RW, Reuter H (1999) Nature 399:159–162.
16. Peterson BZ, DeMaria CD, Adelman JP, Yue DT (1999) Neuron 22:549–558.
17. Zu ¨hlke RD, Pitt GS, Tsien RW, Reuter H (2000) J Biol Chem 275:21121–21129.
18. Kim J, Ghosh S, Nunziato DA, Pitt GS (2004) Neuron 41:745–754.
19. Armstrong CM, Bezanilla F (1977) J Gen Physiol 70:567–590.
20. Zhou M, Morais-Cabral JH, Mann S, MacKinnon R (2001) Nature 411:657–661.
21. Bezanilla F, Perozo E, Papazian DM, Stefani E (1991) Science 254:679–683.
22. Hoshi T, Zagotta WN, Aldrich RW (1990) Science 250:533–538.
23. Zagotta WN, Hoshi T, Aldrich RW (1990) Science 250:568–571.
24. Bernatchez G, Talwar D, Parent L (1998) Biophys J 75:1727–1739.
26. Hadley RW, Lederer WJ (1991) J Physiol (London) 444:257–268.
27. Isaev D, Solt K, Gurtovaya O, Reeves JP, Shirokov R (2004) J Gen Physiol 123:555–571.
28. Bean BP, Rios E (1989) J Gen Physiol 94:65–93.
29. Jones LP, DeMaria CD, Yue DT (1999) Biophys J 76:2530–2552.
30. Barrett CF, Cao YQ, Tsien RW (2005) J Biol Chem 280:24064–24071.
31. Peterson BZ, Lee JS, Mulle JG, Wang Y, de Leon M, Yue DT (2000) Biophys J 78:1906–
32. Zu ¨hlke RD, Reuter H (1998) Proc Natl Acad Sci USA 95:3287–3294.
33. Liang H, DeMaria CD, Erickson MG, Mori MX, Alseikhan BA, Yue DT (2003) Neuron
34. Mori MX, Erickson MG, Yue DT (2004) Science 304:432–435.
35. Hodgkin AL, Huxley AF (1952) J Physiol (London) 116:497–506.
36. Ptacek LJ (1997) Neuromusc Disord 7:250–255.
37. Catterall WA (2000) Annu Rev Cell Dev Biol 16:521–555.
38. Piedras-Renterı ´a ES, Barrett CF, Cao YQ, Tsien RW (2007) in Calcium: A Matter of Life
or Death, eds Krebs J, Michalak M (Elsevier, New York), pp 127–166.
39. Imredy JP, Yue DT (1994) Neuron 12:1301–1318.
40. Erxleben C, Gomez-Alegria C, Darden T, Mori Y, Birnbaumer L, Armstrong DL (2003)
Proc Natl Acad Sci USA 100:2929–2934.
41. Adams B, Tanabe T (1997) J Gen Physiol 110:379–389.
42. Fallon JL, Halling DB, Hamilton SL, Quiocho FA (2005) Structure (London) 13:1881–
43. Van Petegem F, Chatelain FC, Minor DL, Jr (2005) Nat Struct Mol Biol 12:1108–1115.
Nat Neurosci 9:1108–1116.
45. Splawski I, Timothy KW, Decher N, Kumar P, Sachse FB, Beggs AH, Sanguinetti MC,
Keating MT (2005) Proc Natl Acad Sci USA 102:8089–8098.
46. Jacobs A, Knight BP, McDonald KT, Burke MC (2006) Heart Rhythm 3:967–970.
47. Sicouri S, Timothy KW, Zygmunt AC, Glass A, Goodrow RJ, Belardinelli L, Antzelevitch
C (2007) Heart Rhythm 4:638–647.
48. Erxleben C, Liao Y, Gentile S, Chin D, Gomez-Alegria C, Mori Y, Birnbaumer L, Arm-
strong DL (2006) Proc Natl Acad Sci USA 103:3932–3937.
49. Soldatov NM, Bouron A, Reuter H (1995) J Biol Chem 270:10540–10543.
50. Zu ¨hlke RD, Bouron A, Soldatov NM, Reuter H (1998) FEBS Lett 427:220–224.
51. Soldatov NM (1992) Proc Natl Acad Sci USA 89:4628–4632.
52. Livneh A, Cohen R, Atlas D (2006) Neuroscience 139:1275–1287.
53. Bernatchez G, Berrou L, Benakezouh Z, Ducay J, Parent L (2001) Biochim Biophys Acta
www.pnas.org?cgi?doi?10.1073?pnas.0710501105Barrett and Tsien