DOI: 10.1126/science.
1197598
, 599 (2011); 331 Science
et al. Takuya Sasaki
Action-Potential Modulation During Axonal Conduction
This copy is for your personal, non-commercial use only.
clicking here. colleagues, clients, or customers by
, you can order high-quality copies for your If you wish to distribute this article to others
here. following the guidelines
can be obtained by Permission to republish or repurpose articles or portions of articles
): May 5, 2014 www.sciencemag.org (this information is current as of
The following resources related to this article are available online at
http://www.sciencemag.org/content/331/6017/599.full.html
version of this article at:
including high-resolution figures, can be found in the online Updated information and services,
http://www.sciencemag.org/content/suppl/2011/02/02/331.6017.599.DC1.html
can be found at: Supporting Online Material
http://www.sciencemag.org/content/331/6017/599.full.html#related
found at:
can be related to this article A list of selected additional articles on the Science Web sites
http://www.sciencemag.org/content/331/6017/599.full.html#ref-list-1
, 5 of which can be accessed free: cites 15 articles This article
http://www.sciencemag.org/content/331/6017/599.full.html#related-urls
16 articles hosted by HighWire Press; see: cited by This article has been
http://www.sciencemag.org/cgi/collection/neuroscience
Neuroscience
subject collections: This article appears in the following
registered trademark of AAAS.
is a Science 2011 by the American Association for the Advancement of Science; all rights reserved. The title
Copyright American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the Science
o
n
M
a
y
5
,
2
0
1
4
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d
f
r
o
m
o
n
M
a
y
5
,
2
0
1
4
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d
f
r
o
m
o
n
M
a
y
5
,
2
0
1
4
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d
f
r
o
m
o
n
M
a
y
5
,
2
0
1
4
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d
f
r
o
m
Action-Potential Modulation During
Axonal Conduction
Takuya Sasaki,
1
Norio Matsuki,
1
Yuji Ikegaya
1,2
*
Once initiated near the soma, an action potential (AP) is thought to propagate autoregeneratively
and distribute uniformly over axonal arbors. We challenge this classic view by showing that APs
are subject to waveform modulation while they travel down axons. Using fluorescent patch-clamp
pipettes, we recorded APs from axon branches of hippocampal CA3 pyramidal neurons ex vivo.
The waveforms of axonal APs increased in width in response to the local application of glutamate
and an adenosine A
1
receptor antagonist to the axon shafts, but not to other unrelated axon branches.
Uncaging of calcium in periaxonal astrocytes caused AP broadening through ionotropic glutamate
receptor activation. The broadened APs triggered larger calcium elevations in presynaptic boutons
and facilitated synaptic transmission to postsynaptic neurons. This local AP modification may enable
axonal computation through the geometry of axon wiring.
C
ontrary to the prevailing notion of the
digital-like uniformity of action poten-
tials (APs), recent evidence has shown
that APs (or axons) are capable of conveying
information in a graded analog manner (14). It
remains unknown, however, whether already-
generated APs can be modulated during axonal
conduction. Because axons express various types
of transmitter receptors and ion channels, in par-
ticular on their presynaptic terminals (5, 6), local
alterations in the ion conductance of aligned
synapses along an axon may modify the wave-
forms of APs traveling down its length.
We recorded CA3 pyramidal cells (PCs) in
hippocampal slice cultures, unless otherwise spec-
ified (7). Alexa Fluorloaded patch pipettes
were used to visualize their axons (Fig. 1A). An
axonal branch that was 150 to 700 mm away from
the axon hillock was targeted for cell-attached
recording with a fluorophore-coated pipette un-
der spinning-disk confocal visualization (8). APs
were evoked by current injection into the soma
and were extracellularly captured at the axon
as sharp sink potentials. Extracellularly recorded
APs (eAPs) were likely to be a mixture of the
inverse of intracellularly recorded APs (iAPs)
and their derivatives (fig. S1).
Glutamate or g-aminobutyric acid (GABA)
was locally puff-applied to the axon midway be-
tween two patch-clamp pipettes. The effective
radius of this local application was less than
100 mm, as confirmed by diffusion of coapplied
Alexa Fluor dye. Application of 10 mM glutamate
induced a rapid and reversible increase in the
half-maximal duration of eAPs that were recorded
within 200 mm of the site of glutamate admin-
istration (Fig. 1B, t
10
= 5.44, P = 0.0003 paired
t test). This low concentration of glutamate did
not change the iAP width recorded at the soma
(Fig. 1B, 0.61 T 1.6%, t
10
= 0.28, P = 0.78) or
the somatic resting membrane potential (0.4 T
0.8 mVof change, t
10
= 0.53, P = 0.60). It also
did not affect the spike activity of nearby neu-
rons (fig. S2). Unlike glutamate, 1 mM GABA
had no significant effect on the eAP width (Fig.
1C, t
5
= 1.03, P = 0.34).
Glutamate did not broaden eAPs in the pres-
ence of 10 mM 6-cyano-7-nitroquinoxaline-2,3-
dione (CNQX), a nonN-methyl-D-aspartate
(NMDA) receptor antagonist (Fig. 1C, t
6
= 1.31,
P = 0.23), but it did in the presence of 50 mM
L,D-2-amino-5-phosphonopentanoic acid (AP5),
an NMDA receptor antagonist (MCPG) (Fig.
1C, t
5
= 3.47, P = 0.02), and 500 mM (S )-a-
methyl-4-carboxyphenylglycine, a group I/II metabo-
tropic glutamate receptor antagonist (Fig. 1C, t
5
=
7.42, P = 0.0006). The eAP broadening was not af-
fected by bath application of a glutamate transporter
inhibitor [100 mM DL-threo-b-benzyloxyaspartic
acid (TBOA)] (Fig. 1C, t
3
= 4.00, P = 0.028), pre-
incubation with a vesicle fusion inhibitor [10 nM
tetanus toxin (TeNT)] (Fig. 1C, t
4
= 4.06, P =
0.0097), or intracellular injection of a Ca
2+
chelator
[20 mM1,2-biso-aminophenoxy)ethane-N,N,N',N'-
tetraacetic acid (BAPTA)] into the somata (Fig. 1C,
t
3
= 4.19, P = 0.024). Thus, glutamate appears to
1
Laboratory of Chemical Pharmacology, Graduate School of
Pharmaceutical Sciences, University of Tokyo, Tokyo 113-0033,
Japan.
2
Precursory Research for Embryonic Science and Tech-
nology, Japan Science and Technology Agency, 4-1-8 Honcho
Kawaguchi, Saitama 332-0012, Japan.
*To whom correspondence should be addressed. E-mail:
ikegaya@mol.f.u-tokyo.ac.jp
Fig. 1. Glutamate applied to axons induces AP broadening. (A) Confocal
image of dual patch-clamping from the soma and axon of a CA3 PC.
Drugs were locally puffed onto the path of the axon (red). Traces show
somatic iAP and axonal eAP 300 mm from the axon hillock. (B) The half-
maximal width of axonal eAPs (middle), but not of somatic iAPs (bottom),
increased during local application of glutamate. Top: eAP traces at time
points 1 and 2. (C) Effects of pharmacological reagents on eAP width. n = 4
to 7 slices, *P < 0.05, paired t test. (D) Glutamate-induced inward current in
whole-cell recorded axon blebs. (E) Phase-space analysis of an eAP wave-
form. (Left) An eAP was plotted in the space of V versus dV/dt, where V
represents the eAP voltage at a given time. The phase q (blue arrow) was
determined to maximize the difference between the orbits of eAPs before
(black) and during drug application (red). Right: The q values were sim-
ilar for eAPs modulated by glutamate, high-K
+
depolarization, or 4-AP,
but not by TTX. n = 4 or 5 slices.
www.sciencemag.org SCIENCE VOL 331 4 FEBRUARY 2011 599
REPORTS
modulate APs by depolarizing axons through
AMPAreceptor activation rather than by trigger-
ing transmitter release or Ca
2+
-dependent cellular
signals. Consistent with this, 20 mM K
+
applied
to axons induced eAP broadening (Fig. 1C, t
3
=
4.59, P = 0.0037).
In CA3 PCs of acute hippocampal slices, the
cut ends (blebs) of the axons were whole-cell
patch-clamped (2). Puffing glutamate to the blebs
elicited an inward current of 38.2 T 5.6 pA at
a holding potential of 70 mV (Fig. 1D, n = 3
blebs).
To investigate the mechanisms underlying the
AP broadening, we analyzed the eAP waveform
in the phase space of V versus dV/dt (Fig. 1E).
The orbits of control and glutamate-broadened
eAPs separated maximally in a late phase during
their temporal evolution. The late-phase modu-
lation was replicated by high-K
+
depolariza-
tion and pharmacological blockade of the Kv1
family of voltage-activated K
+
channels by 10 mM
4-aminopyridine (4-AP), whereas the earlier phase
was sensitive topartial blockade of voltage-activated
Na
+
channels by 10 nMtetrodotoxin (TTX). Thus,
depolarization-induced K
+
channel inactivation
(1, 4) is likely to underlie the AP broadening.
To determine the extent of glutamates in-
fluence, we patched axons at various distances
downstream of the site of glutamate administra-
tion. eAP broadening was more evident at axons
nearer the site of glutamate administration, and
the decay constant was found to be 223 mm
(Fig. 2A, n = 17 axons). Likewise, the eAP-
broadening effect was not observed at distances
greater than 250 mmupstreamfromthe glutamate
spot (t
4
= 0.51, P = 0.60) or at axon branches
other than those that received glutamate puffs
(Fig. 2B, t
5
= 0.66, P = 0.54).
Adenosine A
1
receptors are expressed in pre-
synaptic axons (9). Local application of 1 mM
adenosine did not modulate eAPs (Fig. 1C, t
5
=
0.33, P = 0.75), but 100 mM 8-cyclopentyl-1,3-
dipropylxanthine (DPCPX), an A
1
receptor antag-
onist, increased the eAP width (Fig. 1C, t
5
= 7.61,
P = 0.0006), suggesting tonic AP suppression by
endogenous adenosine.
We next sought to determine whether AP mod-
ulation affects presynaptic Ca
2+
dynamics (fig. S3).
Oregon Green BAPTA-1 (OGB1), a Ca
2+
indi-
cator, was intracellularly injected into PCs. In
some cases, its AM form was bolus-injected into
the stratum oriens. Intra-axonally diffused dye was
confocally imaged from individual presynaptic
boutons. APs induced a transient rise in Ca
2+
fluorescence in the boutons. The increases in Ca
2+
were further enhanced by glutamate applied to
axons. This enhancement did not occur in the
presence of CNQX and AP5. DPCPX, but not
adenosine, also enhanced the Ca
2+
response.
To investigate whether AP broadening mod-
ulates synaptic efficacy, we recorded synaptically
connected pairs of CA3 PCs. In each experi-
ment, we identified the axon branch that was
responsible for the synaptic connection by se-
quentially puffing 1 mM TTX onto individual
Alexa-visualized axon collaterals (Fig. 3A). Focal
activation of the identified branch by glutamate
application increased the mean amplitude of uni-
tary excitatory postsynaptic currents (uEPSCs)
in the downstream postsynaptic neurons (Fig. 3,
B and C, t
11
= 3.49, P = 0.006) and decreased
the failure of synaptic transmission (fig. S4) and
the paired-pulse response ratio (PPR), defined as
the second uEPSC amplitude divided by the
first evoked by two 50-ms-interval stimuli (Fig.
3D, t
4
= 4.31, P = 0.01). DPCPX (Fig. 3D, t
5
=
5.70, P = 0.002), but not adenosine (Fig. 3D,
t
5
=1.38, P = 0.23), increased the uEPSC am-
plitude. 4-AP also increased synaptic efficacy
(Fig. 3D, t
5
= 5.46, P = 0.003) and decreased
PPR (t
5
= 7.81, P = 0.0006).
Most local axons of CA3 PCs are unmyeli-
nated, and their shafts and presynaptic varicos-
ities contact astrocytes (10). These periaxonal
astrocytes may modulate APs through the re-
lease of gliotransmitters. Astrocytes residing near
the axons of CA3 PCs (150 to 400 mm from axon
Fig. 2. The AP-broadening effect
reaches beyond the glutamate-
activated region. (A) A glutamate-
induced increase in axonal eAP
width is plotted against the path
distance between the glutamate
puff and the cell-attached record-
ing positions. Red indicates the
puffed area (<100 mm f). The line
was best fit to a single exponential
decay (R
2
= 0.50, P = 0.01, n =
17 axon recordings). The decay con-
stant l (223 mm) was smaller than
the axon length constant reported
previously (~500 mm) (24). Extra-
cellular recording of APs may underestimate the effect of AP broadening. (B) Glutamate applied to axon
branches in different locations than the recording site had no effect. n = 6 axons.
Fig. 3. Glutamate applied to axons facilitates downstream synaptic ef-
ficacy. (A) Identification of the axon branch (no. 4) connecting a patched
neuron pair by sequential application of TTX to spots 1 to 7. Red, axons;
blue, somatodendrites. Only the main axon fibers are illustrated. (B) In-
dividual uEPSC traces (gray traces) and their averages (thick traces) before and during
glutamate application. (C) uEPSCs increased in amplitude during glutamate applica-
tion. (D) Effects of local drug application on synaptic responses evoked by single-pulse (left) and paired-pulse (right) stimulation. n = 5 to 12 slices, *P <
0.05, paired t test.
4 FEBRUARY 2011 VOL 331 SCIENCE www.sciencemag.org 600
REPORTS
hillocks) were bolus-loaded with OGB1-AM
and O-nitrophenyl-ethylene glycol tetraacetic acid
AM (NP-EGTA-AM), a membrane-permeating
caged-Ca
2+
compound (Fig. 4A). Astrocytes were
morphologically and immunohistochemically
identified (fig. S5) and preferentially loaded with
NP-EGTA-AM (11). Photolysis of caged Ca
2+
by
an ultraviolet (UV) pulse (120 mm in diameter)
elicited oscillatory Ca
2+
fluctuations in the illu-
minated astrocytes, which persisted for several
minutes (fig. S6D). The activity did not propagate
to neighboring, nonilluminated astrocytes. UV
illumination did not activate astrocytes that were
not loaded with NP-EGTA (fig. S6E, UValone).
Ca
2+
uncaging in astrocytes increased the
duration of eAPs recorded at downstream axons
within 200 mm of the UV spot (Fig. 4, B and C,
t
6
= 5.90, P = 0.001). The same effect was ob-
served in axons of CA3 PCs of acute hippocam-
pal slices (a 4.5 T 1.0% increase in eAP width,
t
3
= 4.83, P = 0.02). UV pulses did not induce
eAP broadening in the presence of CNQX and
AP5 (Fig. 4C, t
4
= 1.40, P = 0.23) or in slices
without NP-EGTA-AM loading (Fig. 4C, t
4
=
0.38, P = 0.73). We rule out the possibility that
UV-induced Ca
2+
elevation in NP-EGTAloaded
axons caused the eAP broadening, because UV
illumination of axons of presynaptic neurons in-
jected with 200 mM NP-EGTA failed to broaden
eAPs (Fig. 4C, t
4
= 0.02, P = 0.98). Astrocyte ac-
tivation increased the amplitude of AP-triggered
Ca
2+
transients in presynaptic boutons, an effect
that was not observed in the presence of CNQX
and AP5 or in slices without NP-EGTA loading
(fig. S7).
To examine whether the Ca
2+
activity of peri-
axonal astrocytes facilitated downstream synap-
tic efficacy, astrocytes near the presynaptic axon
branches connecting patched PC pairs were loaded
with NP-EGTA-AM (Fig. 4D). UV activation of
these astrocytes amplified uEPSCs in the post-
synaptic neurons (Fig. 4E, t
18
= 2.27, P= 0.03) and
reduced PPR (t
15
= 2.54, P = 0.02). Both effects
were blocked by coapplication of CNQX and AP5
to the UV-illuminated area (Fig. 4E, t
8
= 0.24, P =
0.81). Neither UVillumination alone (Fig. 4E, t
9
=
0.97, P = 0.36) nor UVactivation of astrocytes on
irrelevant axon branches (t
5
= 0.82, P = 0.45)
facilitated synaptic transmission. To examine the
impact of single astrocyte activation, we injected
NP-EGTA into a periaxonal astrocyte through a
whole-cell pipette (fig. S8A). In 5 of 13 exper-
iments, photostimulation led to a modest but
significant increase in uEPSC efficacy as well as
a decrease in PPRs. When the data for the entire
sample were analyzed statistically, these effects
were still significant (fig. S8B). Ca
2+
uncaging in
the axons of NP-EGTAloaded presynaptic neu-
rons failed to modulate synaptic transmission.
We demonstrated that activation of AMPA
receptors directly or indirectly causes a depolar-
izing current in axons and thereby broadens APs
during axonal conduction. The endogenous agonist
glutamate appears to be provided by periaxonal
astrocytes [but see (12)]. Astrocytes are known to
regulate local synaptic transmission (1316), but
this work reveals that they are far more influential
than expected, because the length constant of the
axon cable (24) exceeds their cell diameter. Our
findings derived from experimentally designed ex
vivo systems using artificial stimulation should be
extrapolated to other systems with caution. Eluci-
dating their physiological relevance requires further
investigation.
References and Notes
1. Y. Shu, Y. Yu, J. Yang, D. A. McCormick, Proc. Natl. Acad.
Sci. U.S.A. 104, 11453 (2007).
2. Y. Shu, A. Hasenstaub, A. Duque, Y. Yu, D. A. McCormick,
Nature 441, 761 (2006).
3. H. Alle, J. R. Geiger, Science 311, 1290 (2006).
4. M. H. Kole, J. J. Letzkus, G. J. Stuart, Neuron 55, 633
(2007).
5. K. W. Schicker, M. M. Dorostkar, S. Boehm, Curr. Mol.
Pharmacol. 1, 106 (2008).
6. H. S. Engelman, A. B. MacDermott, Nat. Rev. Neurosci. 5,
135 (2004).
7. Materials and methods are available as supporting
material on Science Online.
8. D. Ishikawa et al., Neural Netw. 23, 669 (2010).
9. R. A. Cunha, Neurochem. Int. 38, 107 (2001).
10. R. Ventura, K. M. Harris, J. Neurosci. 19, 6897
(1999).
11. Q. S. Liu, Q. Xu, G. Arcuino, J. Kang, M. Nedergaard,
Proc. Natl. Acad. Sci. U.S.A. 101, 3172 (2004).
12. N. B. Hamilton, D. Attwell, Nat. Rev. Neurosci. 11, 227
(2010).
13. J. Kang, L. Jiang, S. A. Goldman, M. Nedergaard,
Nat. Neurosci. 1, 683 (1998).
14. P. Jourdain et al., Nat. Neurosci. 10, 331 (2007).
15. G. Perea, A. Araque, Science 317, 1083 (2007).
16. C. Henneberger, T. Papouin, S. H. Oliet, D. A. Rusakov,
Nature 463, 232 (2010).
17. This work was supported in part by a Grant-in-Aid for
Science Research (nos. 22115003, 22650080, and
22680025) from the Ministry of Education, Culture,
Sports, Science, and Technology of Japan. T.S. collected
experimental data and carried out the data analysis.
Y.I. conceived the project and carried out the data
analysis. T.S. and Y.I. wrote the manuscript. N.M.
supervised the project and provided feedback on the
manuscript.
Supporting Online Material
www.sciencemag.org/cgi/content/full/331/6017/599/DC1
Materials and Methods
Figs. S1 to S8
References
9 September 2010; accepted 29 December 2010
10.1126/science.1197598
Fig. 4. Ca
2+
uncaging
in periaxonal astrocytes
broadens APs and facili-
tates downstream synaptic
transmission. (A) Astro-
cytes (magenta) bolus-
loaded with NP-EGTA-AM
near the axon whose so-
ma and downstream axon
(green) were patch-clamp
recorded. UV illuminated
the circled area. (B) Pho-
tolysis of NP-EGTA induced
axonal eAP broadening.
Inset at top: eAPs at time
points 1 and 2. (C) Sum-
marized data, including
slices that were not loaded
with NP-EGTA-AM (UV
alone) and slices in which
the presynaptic neuron
was intracellularly loaded
with NP-EGTA (pre-neuron).
(D) Astrocytes near the
axon connecting two patch-clamped neurons were bolus-loaded
with NP-EGTA-AM. (E) CNQX and AP5 were locally applied to the UV-illuminated
area. UV was also applied to unloaded slices (UV alone) and slices in which NP-EGTA-AM was bolus-loaded
into the astrocytes around irrelevant axon branches (other branch). n = 6 to 19 slices, *P < 0.05, paired t test.
www.sciencemag.org SCIENCE VOL 331 4 FEBRUARY 2011 601
REPORTS
