Available via license: CC BY-NC-SA 4.0
Content may be subject to copyright.
The Journal of Neuroscience, February 1, 1996, 76(3):1066-i 071
The Possible Role of Hydrogen Sulfide as an
Endogenous Neuromodulator
Kazuho Abe and Hideo Kimura
The Salk Institute for Biological Studies, San Diego, California 92 138
Hydrogen sulfide (H,S), which is well known as a toxic gas, is
produced endogenously from L-cysteine in mammalian tissues.
H,S is present at relatively high levels in the brain, suggesting
that it has a physiological function. Two other gases, nitric
oxide and carbon monoxide, are also endogenously produced
and have been proposed as neuronal messengers in the brain.
In this work we show the following: (1) an H,S-producing
enzyme, cystathionine p-synthase (CBS), is highly expressed in
the hippocampus; (2) CBS inhibitors hydroxylamine and amino-
oxyacetate suppress the production of brain H,S; and (3) a
CBS activator, S-adenosyl-L-methionine, enhances H,S pro-
duction, indicating that CBS contributes to the production of
endogenous H,S. We also show that physiological concentra-
tions of H,S selectively enhance NMDA receptor-mediated
responses and facilitate the induction of hippocampal long-
term potentiation. These observations suggest that endoge-
nous H,S functions as a neuromodulator in the brain.
Key words: hydrogen sulfide; cystathionine P-synthase; long-
term potentiation; NMDA receptors; hippocampus
Endogenous hydrogen sulfide (H2S) can be formed from cysteine
by pyridoxal-5’-phosphate-dependent enzymes, including cystathi-
onine /3-synthase (CBS) and cystathionine y-lyase (CSE)
(Stipanuk and Beck, 1982; Griffith, 1987; Erickson et al., 1990;
Swaroop et al., 1992). Both CBS and CSE have been intensively
studied for their activities in the liver and kidney (Stipanuk and
Beck, 1982; Erickson et al., 1990; Swaroop et al., 1992) but little
is known about them in the brain. Recently, endogenous levels of
H,S in the brain have been measured in the rat, human, and
bovine (Goodwin et al., 1989; Warenycia et al., 1989a; Savage and
Gould, 1990). The relatively high concentration of endogenous
H2S in the brain (50-160
PM)
suggests that it has a physiological
function.
Although it has not been possible to determine which form of
H,S (H,S, HS, or S’-) is active, the term “hydrogen sulfide” has
been used in toxicity studies. The term “hydrogen sulfide” is also
used here. Most studies about H,S have been devoted to its toxic
effects (Reiffenstein et al., 1992) with little attention paid to its
physiological function. Two other gases, nitric oxide (NO) and
carbon monoxide (CO), are also produced endogenously by en-
zymes localized in the brain. NO is produced by NO synthase via
the metabolism of arginine (Palmer et al., 1988; Bredt and Snyder,
1992) and CO is produced by heme oxygenase via the metabolism
of heme to biliverdin (Maines, 1988). Both NO and CO have been
proposed as retrograde messengers in hippocampal long-term
potentiation (LTP) (O’Dell et al., 1991; Schuman and Madison,
1991; Haley et al., 1992; Stevens and Wang, 1993; Zhuo et al.,
1993) a synaptic model of learning and memory (Bliss and Col-
lingridge, 1993).
The H,S-producing enzyme CBS was found to be highly expressed
Recrwed Aug. X. 1995; revised Nov. 29, IYYS; accepted Dec. 4, IYYS.
This work was supported by grants from the National Institutes of Health (NIH)
to Dr. David Schubert (ROINSU9658) and from the NIH (R29NS31202) and the
Alzheimeis Association to H.K. K.A. was supported by NIH Grant ROINS09658.
WC thank Dr. J. P. Kraus for a CBS cDNA plasmid and Dr. P. F. Erickson for a CSE
cDNA plasmid. We thank Drs. D. Schubert, J. P. Kraus. and P. F. Erickson for
discussions. We also thank Drs. D. Schubert, T. Saitoh, Y. Goda, Y. Sagara, and P.
Maher for readine this manuscriot.
Correspondence should he aidrrased to Hideo Kimura, The Salk Institute for
Biological Studies, P.0. Box X.5800, San Diego, CA 92138.
Copyright 0 lYY6 Society for Neuroscience 0270.6474/96/161066-06$05.00/0
in the hippocampus and cerebellum in the present study. Brain
homogenates produce H,S in the presence of cysteine and pyridoxal-
5’-phosphate. The production of H,S is inhibited by CBS inhibitors
hydroxylamine and amino-oxyacetate and is increased by an activator
of CBS, S-adenosyl-L-methionine (AdoMet), indicating that CBS
contributes to the production of endogenous H,S. Although high
concentrations of H,S inhibit synaptic responses, physiological con-
centrations of H,S facilitate the induction of LTP in the hippocam-
pus. These observations suggest that endogenous H,S functions as a
neuromodulator in the brain.
MATERIALS AND METHODS
Northern blot
analysis. Total RNAs (10 pg) were electrophorcsed
in a 0.66
M formaldehyde denaturing gel and blotted on Hybond-N nylon mem-
brane (Amersham, Arlington Heights, IL). Hybridization was performed
in a solution of 50% formamide, 0.65 M NaCI, 0.2%’ SDS, and 100 pg/ml
salmon sperm DNA at 42°C for 16 hr, and the blot was then washed twice
with 0.1X SSC, 0.2% SDS for 30 min at 65°C.
Measurement of H,S production. Enzymatic capacity for HzS produc-
tion in brain homogenates was measured according to the method by
Stipanuk and Beck (1982). Briefly, the whole brain was isolated from
adult rats and homogenized in ice-cold 50 mst potassium phosphate
buffer, pH 6.8, with a Polytron homogenizer (KINEMATICA, Lucerne,
Switzerland). One milliliter of an assay reaction mixture contained (in
mM): 10 L-cysteine, 2 pyridoxal 5’.phosphate, 100 potassium phosphate
buffer, pH 7.4, and 12% (w/v) b rain homogenate. o,L-Propargylglycinc (2
mM) and S-adenosyl-L-methionine (2 mM) were incubated with tissue
homogenates before the enzyme reaction at 37°C for 5 and IS min,
respectively. All the other inhibitors were added to a final reaction
mixture at concentrations shown in Figure 2, and the enzyme activities
were measured. Incubations for the enzyme reactions were performed in
25 ml Erlenmeyer flasks fitted with septum stoppers and plastic center
wells (Kontes, Vineland, NJ). Center wells were filled with 0.5 ml of 1%
(w/v) zinc acetate and a filter paper for trapping evolved H,S as zinc
sulfide. Each flask was flushed with Nz for 20 set and then sealed. The
reactions were initiated by transferring the flasks from an ice bath to a
37°C shaking water bath. After 90 min at 37°C reactions were stopped by
injecting 0.5 ml of 50% (w/v) trichloroacetic acid. Flasks were incubated
in the shaking water bath at 37°C for an additional hour to complete
trapping of H,S. The center wells and contents were transfcrrcd to test
tubes and mixed with 3.5 ml of distilled water. To each tube, 0.4 ml of 20
mM N,N-dimethyl-p-phenylenediamine sulfate in 7.2 M HCI was added,
immediately followed by the addition of 0.4 ml of 30 mM FeCI, in 1.2 M
HCI. After 20 min of incubation at room temperature, the absorbance of
the resulting solution at 670 nm was measured with a spectrophotometer.
All assays were done in duplicate. The calibration curve of absorbance
versus sulfide concentration was made by using defined concentrations of
Abe and Kimura
l
Hydrogen Sulfide as an Endogenous Neuromodulator J. Neurosci., February 1, 1996, 76(3):1066-1071 1067
sodium hydrosulfide (NaHS) solution. A stock solution of NaHS (100
mM)
was prepared by dissolving NaHS compensated with the ratio of
NaHS/H,O immediately before use.
Elecrrophysiology. Hippocampal slices (400-450 pm) were prepared
from Sprague-Dawley rats (3-6 weeks old for LTP experiments and
12-15 d old for whole-cell recordings) and maintained in a chamber (1.5
ml) at 34”C, where they were continuously perfused with artificial CSF
(ACSF) consisting of (in
mM):
124 NaCl, 4 KCI, 2.4 CaCl,, 1.3 MgSO,,
1.24 NaH,POd, 26 NaHCO,, and 10 glucose, bubbled with 95% O,/
5% co,. - _
A bipolar stimulating electrode was placed in the stratum radiatum in
the CAl/CA2 border region, and the evoked EPSP and the population
snike were extracellularly recorded from the stratum radiatum and the
pyramidal cell layer in the CA1 region, respectively, with a glass capillary
microelectrode (3-5 Ma) filled with 0.9% NaCI. A single test stimulation
(0.1 msec duration) was applied at intervals of 20 sec. The stimulus
intensity was adjusted in the range of 35-55 /.LA to evoke 0.8-1.0 mV of
the field EPSPs, and in the range of 50-100 PA which induce 50% of the
maximum amplitude for the population spikes. Changes in field potential
were recorded in current clamp mode with an Axopatch 200A amplifier
(Axon Instruments. Foster Citv. CA) and digitized with a DigiData 1200
r&log-to-digital converter (Axon Instruments). Nine consecutive records
were averaged, and the data were stored on a computer (Date1 486, San
Diego, CA).
To induce potentiation of evoked field potentials, a tetanic stimulation
was applied at the same intensity with the test stimulation for the
population spikes and at twice as much intensity for field EPSPs. Nine
consecutive records were averaged and the data were collected at inter-
vals of 3 min. After all LTP experiments, a strong tetanic stimulation (100
pulses at 100 Hz, twice at an interval of 20 set) was applied to determine
whether slices were able to induce LTP.
To record membrane currents induced by NMDA or AMPA, whole-
cell patch recording was performed in CA1 pyramidal neurons in hip-
pocampal slices. Patch electrodes (4-4.5 MR) were filled with an internal
solution consisting of (in mM): 113 CsF, 7 KCI, 1 MgCl,, 1 CaCI,, 10
EGTA, 2 Mg-ATP, and 10 HEPES, pH 7.3. After a cell-attached gigaohm
seal was made, the whole-cell recording was achieved by applying addi-
tional suction to rupture the membrane patch. Membrane potential was
clamped at -60 mV, and membrane currents were monitored using an
Axopatch 200A amplifier. The cells were allowed to equilibrate for lo-15
min until the baseline membrane current became stable. NMDA (20
pM)
or AMPA (10
p,~)
was added to ACSF and applied by perfusion (1.5
A
1234
B
Figure
1. Expression of CBS mRNA in the brain. A, Northern blot
analysis of total RNA extracted from the cerebral cortex (lane I), hip-
pocampus (lane 2), brainstem (lane 3), and cerebellum (lane 4) of rats. The
blot was hybridized with an EcoRI fragment of CBS cDNA obtained from
Dr. J. P. Kraus. B, The membrane was stained with methylene blue before
hybridization. Methylene blue stains RNA and allows the comparison of
the amount of RNA loaded into each lane.
a
T
r.
E
g 200
‘ii
E
.E 100
. . . . . . . . . . . . .
ii
s
e
Q
tn 0
.I 1 5
.Ol .l 1 5 .Ol .l 1 5
PGly
**----.. __----. .---- A&Met
NH20H Amlnooxy-
(mW
acetate (mM)
Figure
2. H,S production in the brain. H,S produced from cysteine in
brain homogenates was measured. Brain homogenates produced 22.6 -C
1.6 nmol H,S/min per G-protein (n = 7) in the presence of 10 mM
L-cysteine and 2
IIIM
pyridoxal 5’-phosphate. CSE inhibitors
D,L-
propargylglycine (PGZy) and /3-cyano+alanine (PCNA) did not suppress
the production of H,S. On the other hand, CBS inhibitors hvdroxvlamine
(NH,OH)
and amino-oxyacetate
(Aminoo&zcetate)
suppressed I-&S pro-
duction, and a CBS activator, S-adenosvl+methionine
(Adokfetl
ooten-
\ II .
tiated the H,S production. The values’in drug-treated groups were ex-
pressed as a percentage of those in the control. All data are represented
as the mean + SEM of five experiments.
ml/mm). A stock solution of NaHS (100 mivt) was prepared by dissolving
NaHS immediately before use.
RESULTS
Expression of H&producing enzyme in the brain
To determine whether CBS and CSE are present in rat brain, we
tested the expression of their mRNAs by Northern blot analysis.
CBS was highly expressed in the hippocampus and cerebellum
compared with the cerebral cortex and brainstem (Fig. 1). Al-
though a small amount of CSE mRNA was detected in the brain
by PCR (Erickson et al., 1990), it was not detectable by Northern
blot analysis (data not shown).
H,S production in the brain
Because CBS is expressed in the brain, H,S production in this
tissue was measured according to the method by Stipanuk and
Beck (1982). Brain homogenates produced 22.6 r: 1.6 nmol H,S/
min per G-protein (n = 7) in the presence of 10 mM L-cysteine and
2 mM pyridoxal5’-phosphate. The H,S production was suppressed
by potent CBS inhibitors hydroxylamine (ICs, = 10e4
M)
and
amino-oxyacetate (IC,, =
10e4
M)
(Braunstein et al., 1971) in a
concentration-dependent manner (Fig. 2). Pretreatment with
AdoMet, a specific activator of CBS (EC,, = 10m4
M)
(Finkelstein
et al., 1975; Stipanuk and Beck, 1982), increased the H,S produc-
tion by 125% (Fig. 2). In contrast, D,L-propargylglycine, an irre-
versible inhibitor of CSE (IC,, = 10m4
M)
that has no effect on
CBS (Uren et al., 1978; Stipanuk and Beck, 1982), and p-cyano-
L-alanine, a competitive inhibitor of CSE (IC,, = 10e5
M),
which
also does not block CBS (Rfeffer and Ressler, 1967; Uren et al.,
1978), only weakly suppressed the H,S production (13 and 19%,
respectively).
High concentrations of H2S inhibit synaptic
transmission in the hippocampus
Because CBS is expressed and produces H,S in the brain, and
because the endogenous concentration of H,S in the brain is
relatively high (50-160
PM)
(Goodwin et al., 1989; Warenycia
et al., 1989a; Savage and Gould, 1990), H,S may play a role in
1066 J. Neurosci., February 1, 1996, 76(3):1066-1071
Abe and Kimura
l
Hydrogen Sulfide as an Endogenous Neuromodulator
A
NaHS
130 PM
320 PM 640 pM
;
I
0.5 ,_ .--ye-
l . 0
. ..J! . . . . . . - . . . . . . . . .
1 I ” ” ’ “1
0
30 60 90
Time (min)
Control
640
pM NaHS
Superimposed
Figure
3. High concentrations of H,S inhibit synaptic transmission in the
hippocampus. A, NaHS (320 and 640
FM)
decreased the slopes of the field
EPSP in a concentration-dependent manner. NaHS was applied by per-
fusion during the times indicated by
while bars.
B, Sample records of field
EPSPs and population spikes suppressed by 640
pM
NaHS. The signals
denoted by as/erisk.s represent the action potentials generated by direct
stimulation of presynaptic fibers, which are completely abolished by 1
pM
tetrodotoxin, a Na’ channel blocker.
synaptic transmission. The effect of H,S on synaptic transmis-
sion was investigated by recording the field EPSPs and popu-
lation spikes evoked by the electrical stimulation of the Schaf-
fer collaterals in the CA1 region of rat hippocampal slices.
NaHS was used as a source of H,S for the following reasons.
(1) NaHS dissociates to Nap’ and HS- in solution, then HS-
associates with Ht and produces H,S. It does not matter
whether the H,S solution is prepared by bubbling H,S gas or by
dissolving NaHS. In physiological saline, approximately one-
third of the H,S exists as the undissociated form (H,S), and the
remaining two-thirds exists as HS at equilibrium with H,S
(Beauchamp et al., 1984; Reiffenstein et al., 1992). (2) The use
of NaHS enables us to define the concentrations of H,S in
solution more accurately and reproducibly than bubbling H,S
gas. (3) The influence of 51 mM sodium ion qn the electro-
physiological experiments is negligible, because the perfusing
medium (ACSF) contains 150 mM sodium ion. (4) NaHS at
concentrations used in the present study does not change the
pH of buffered ACSF. For these reasons, NaHS has been
widely used for studies of H,S (Beauchamp et al., 1984; Ware-
nycia et al., 1989a,b; Kombian et al., 1993).
NaHS at concentrations of 5130
PM
did not affect the field
EPSPs (Fig. 3A) or the population spikes, whereas higher con-
centrations of NaHS (320 and 640
PM)
suppressed both field
EPSPs and the population spikes (Fig. 3AJ).
Physiological concentrations of H,S facilitate
hippocampal LTP
The abnormal expression of CBS activity causes several diseases,
including mental retardation (Mudd et al., 1989). For example,
, Control
NaHS
B
Control
NaHS ,
I
I I I I
-20 -10 0 10 20 30 40 50
Time (min)
$
L I
Control
IO
60 130
NaHS (PM)
Figure
4. Physiological concentrations of H,S facilitate the induction of
hippocampal LTP. A, B, The effects of 130
pM
NaHS on a weak tetanic
stimulation-induced potentiation of the field EPSPs
(A) and
population
spikes (B). A weak tetanic stimulation (1.5 pulses at 100 Hz; arrows), which
alone did not induce LTP (open
circles),
produced LTP in the presence of
NaHS
(filled circles).
NaHS was applied
by
perfusion during the times
indicated by
black bars.
The field EPSP slopes and the population spike
amplitudes were expressed as the percentage of baseline values before the
tetanic stimulation. The mean field EPSP slope (123. I ? 6.6%, II = 5) and
the population spike amplitude (139.5 + 8.5%, n = 6) 30-48 min after the
tetanic stimulation in the presence of NaHS were significantly different 0,
< 0.05, Student’s f test) from those in control (EPSP: 99.2 ? 1.20/o, II = 6;
spike: 110.3 5 4.0%, n = 5). Representative records at the times denoted
by the numbers are shown as
insets.
C, Concentration dependency of the
LTP-facilitating effect of NaHS. The mean field EPSP slope 30-48 min
after the weak tetanic stimulation was measured. All data are represented
as the mean t SEM.
the CBS gene is located on the chromosome 21, and trisomy 21 in
Down syndrome may contribute to the pathophysiology of this
disease (Kraus, 1990). We therefore examined the effect of phys-
iological concentrations of H,S on LTP. To test whether H,S has
an effect on the induction of LTP, we first assayed the effect of
NaHS at concentrations 5130
PM
with a weak tetanic stimulation
(1.5 pulses at 100 Hz), which alone did not induce LTP. In the
presence of 130
pM
NaHS, a weak tetanic stimulation induced
LTPs of both the field EPSPs and the population spikes (Fig.
4,4,B). This effect of H,S was concentration-dependent in the
range of lo-130
PM
(Fig. 4C). To test whether H2S is required at
Abe and Kimura . Hydrogen Sulfide as an Endogenous Neuromodulator J. Neurosci., February 1, 1996, 76(3):1066-1071 1069
40 -io b 10 20 30 40 50
Time (min)
- NaHS
2 I
IL
I
-10 0 10 20 30 40 50 60
Time (min)
Figure 5. Simultaneous application of H,S with a tetanic stimulation is
required for the induction of LTP. A, B, When 130 pM NaHS was applied
10 min before (A) or 10 min after (B) the application of a weak tetanic
stimulation (15 pulses at 100 Hz), LTP was not induced (n = 5).
the same time as the tetanic stimulation, we added NaHS 10 min
before or after the tetanic stimulation. The perfusion of 130
PM
NaHS either before or after a weak tetanic stimulation did not
facilitate the induction of LTP (Fig. 5). These results indicate that
the physiological concentrations of H,S facilitate the induction of
LTP only when it is simultaneously applied with a weak tetanic
stimulation.
If the potentiation induced by a weak tetanic stimulation in the
presence of H,S shares common mechanisms with LTP induced
by a strong tetanic stimulation, they should occlude each other. To
test this possibility, occlusion experiments (Zhuo et al., 1993;
Kang and Schuman, 1995) were performed. It was tested whether
H,S-induced potentiation occludes the induction of LTP by a
strong tetanic stimulation. After potentiation induced by H,S with
a weak tetanic stimulation reached a plateau, a strong tetanic
stimulation (100
pulses
at 100 Hz, twice at an interval of 20 set)
was applied. There was no significant difference between LTP
induced by a strong tetanic stimulation after the application of
H,S and that of control (Fig. 6A). It was also tested whether the
induction of LTP by a strong tetanic stimulation occludes H,S-
induced potentiation. After LTP had been induced by a strong
tetanic stimulation, H,S with a weak tetanic stimulation produced
no further potentiation (Fig. 6B). These results indicate that the
H,S-induced LTP shares common mechanisms with LTP induced
by a strong tetanic stimulation.
The observation that NO and CO induce LTP even under the
blockade of NMDA receptors (Zhuo et al., 1993) supports the
idea that NO and CO act as retrograde messengers at synapses
(O’Dell et al., 1991; Schuman and Madison, 1991; Stevens and
Wang, 1993). To determine whether the facilitation of LTP by
H,S requires NMDA receptor activation, the effect of NaHS on
LTP induction in the presence of 2-amino-S-phosphonovalerate
(APV), an NMDA receptor antagonist, was examined. NaHS (130
PM)
with a weak tetanic stimulation did not induce LTP in the
presence of 50
PM
APV (mean field EPSP slope 30 min after
Ii I ,
-10 0 10 20 30 40 50 60 70
Time (min)
Strong
tetanus
P
Reset Weak tetanus
v 4
t %PPnnnn
- NaHS
LL
I
I
-10 0 10 20 30 40 50 60 70
Time (min)
Figure 6. Potentiation induced by H2S is not additive with LTP induced
by a strong tetanic stimulation. A, In control slices (open circles), a weak
tetanic stimulation (single arrow) was first applied in the absence of NaHS
and then LTP was induced by a strong tetanic stimulation (100 pulses at
100 Hz, twice at an interval of 20 set; double urrow). In tested slices (filled
circles), after LTP had been induced by 130 yM NaHS (bluck bar) paired
with a weak tetanic stimulation (sing/e anaw), a strong tetanic stimulation
was applied (double UI*OW). Strong tetanic stimulation-induced LTP in the
slices previously potentiated by H,S (mean EPSP slope, 149.4 2 5.5%. II
= 5) was not significantly different from that in control slices (147.3 +
4.2%, n = 5). E, After LTP had been induced by a strong tetanic
stimulation (double arrow), the potentiated response was reset by reducing
stimulus intensity (open arrowhead), and the effect of 130
PM
NaHS with
a weak tetanic stimulation (single UYTOW arId bar) was examined. LTP
induced by a strong tetanic stimulation completely occluded H&induced
potentiation.
tetanus, 99.6 t 0.4%, II = 4), suggesting that the induction of LTP
by H,S requires the activation of NMDA receptors.
H,S enhances NMDA receptor-mediated responses
Hippocampal LTP induced by a tetanic stimulation requires the
activation of NMDA receptors (Collingridge et al., 1983; Harris et
al., 1984). To determine whether H,S modifies NMDA receptors,
we examined the effect of physiological concentrations of H,S on
NMDA-induced currents by whole-cell patch recording with hip-
pocampal slices. Bath application of NMDA (20
PM,
90 set)
induced an inward current of 524.7 i 51.4 pA (n = 7) at a holding
potential of -60 mV (Fig. 7A1,C). NaHS (130
PM)
alone did not
induce any apparent currents but significantly increased the
NMDA-induced inward current (Fig. 7A2,C). This effect of NaHS
disappeared after it was washed out (Fig. 7A3,C). The NMDA-
induced current was completely blocked by APV, confirming that
the response was mediated by NMDA receptors (Fig. 7A4). The
enhancing effect of H,S on NMDA response was concentration-
dependent in the range of lo-130
PM
(Fig. 70), consistent with its
LTP-facilitating effect (Fig. 4C). Although the effect of H,S was
already saturated at 130
PM,
up to 200
PM
H,S potentiated
NMDA responses. We could not perform patch-clamp analysis at
higher concentrations of H,S because the membrane potential
was unstable, which was probably attributable to the general
toxicity of high concentrations of H,S. In contrast, even 130
PM
NaHS had no effect on the currents induced by a non-NMDA
1070 J. Neurosci., February 1, 1996, 76(3):1066-1071
Abe and Kimura
l
Hydrogen Sulfide as an Endogenous Neuromodulator
-J
200 pA
5
ml”
C D
1000 E
E?
a
El-
u= 150
a
Pt
E
gl 500
Y8
-g.Lo
2 “8
s-
z
0
100
v 1
__1200pA
5
mm
NMDA
AMPA
10 60 130
NaHS (PM)
Figure
7. H,S selectively enhances
NMDA receptor-mediated currents.
A, B, Representative records of inward currents induced by bath applica-
tion of
NMDA (20 pM, 90 set; A)
or
AMPA
(10
pM,
60 see; B) at a holding
potential of -60 mV. I, control; 2, in the presence of 130
pM
NaHS; 3,20
min after washing out NaHS; 4, in the presence of 50
pM
APV
(A) or 30
PM
CNQX (B). C, Collected data of the effects of 130
pM
NaHS on inward
currents induced by
NMDA
(n = 7) or
AMPA
(n = 5). The responses to
each agonist were evaluated by measuring peak current amplitude. (*p <
0.05 vs control; paired
t
test).
D,
Concentration dependency of the en-
hancement of NMDA response by NaHS. NMDA-induced currents in the
presence of NaHS were normalized by taking each control value as 100%.
The numbers of observations are shown in
parentheses.
receptor agonist AMPA (10
pM, 60
set) (Fig. 7BI-3,C). The
AMPA-induced current was completely blocked by 6-cyano-7-
nitroquinoxaline-2,3-dione (CNQX), a non-NMDA receptor
antagonist (Fig. 7B4). The failure of H,S to enhance AMPA
response was not attributable to saturation of the AMPA
response because application of a higher concentration (30
PM)
of AMPA induced larger currents (963.6 rf: 53.5 pA, n = 5).
These results indicate that physiological concentrations of H,S
selectively enhance the function of NMDA receptors.
Disulfide bonds play a role in modulating the function of many
proteins, including NMDA receptors (Aizenman et al., 1989;
Tang and Aizenman, 1993). It is therefore possible that H,S
interacts with disulfide bonds or free thiols in NMDA receptors.
To test this possibility, we determined the effect of the irreversible
thiol-protecting agent dithiothreitol (DTT) on the LTP-
facilitating action of NaHS. DTT (1
KIM)
with a weak tetanic
stimulation, which by itself does not cause LTP weakly, but
significantly, facilitated the induction of LTP (Fig. 8). NaHS with
a weak tetanic stimulation, however, still induced LTP even after
treatment with D’IT (Fig. 8) demonstrating that DTT does not
occlude the effect of H,S. These observations suggest that the
thiol redox sites contribute little, if any, to the potentiating effect
of H,S on the induction of LTP.
DISCUSSION
Endogenous H,S is formed primarily by CBS and CSE (Stipanuk
and Beck, 1982; Griffith, 1987; Erickson et al., 1990; Swaroop et
al., 1992). CBS is expressed in the brain (Fig. 1). The production
of H,S in the brain is suppressed efficiently by CBS inhibitors
hydroxylamine and amino-oxyacetate and is strongly potentiated
by a CBS activator, AdoMet (Fig. 2). In contrast, the expression in
the brain of another H,S-producing enzyme, CSE, is under de-
tectable levels by Northern blot analysis. CSE inhibitors
D,L-
propargylglycine and P-cyano-L-alanine do not suppress the pro-
duction of H,S in the brain (Fig. 2) although these inhibitors
suppress H,S production effectively in the liver and kidney
(Stipanuk and Beck, 1982). It is unlikely that the other pyridoxal-
5’-phosphate-dependent enzyme, cysteine aminotransferase, con-
tributes to the production of endogenous H,S, because it requires
a pH much above physiological levels (Ubuka et al., 1978;
Stipanuk and Beck, 1982). These observations indicate that CBS
must be the major enzyme that produces endogenous brain H,S.
Physiological concentrations of H,S induce LTP only when it is
applied associatively with a weak tetanic stimulation, which alone
does not induce LTP (Figs. 4, 5). In addition, occlusion experi-
ments (Fig. 6) show that the potentiation induced by H,S with a
weak tetanic stimulation shares common mechanisms with LTP
induced by a strong tetanic stimulation. Therefore, H,S facilitates
LTP at active, but not quiescent, synapses, suggesting that H,S is
involved in associative learning as defined by Hebb (1949).
Although H,S, like NO and CO, facilitates the induction of
LTP, the mechanism of action of H,S is different from those of
NO and CO. First, long-term enhancement by NO or CO does not
require NMDA receptor activation (Zhuo et al., 1993), whereas
the LTP-facilitating effect of H,S is not observed under the
blockade of NMDA receptors, suggesting that H,S may not act as
a retrograde messenger. Second, NO and CO increase the intra-
cellular cyclic GMP (Garthwaite, 1991; Bredt and Snyder, 1992;
Verma et al., 1993), whereas H,S does not (our unpublished
data). Finally, we found that physiological concentrations of H,S
selectively increase NMDA receptor-mediated responses (Fig. 7).
High concentrations of H,S (>320 PM) inhibit synaptic trans-
mission in the hippocampus (Fig. 3). Although in the presence of
taurine, NaHS inhibits the tetrodotoxin-sensitive sodium channels
(Warenycia et al., 1989b), the suppression of EPSPs and popula-
tion spikes by high concentrations of NaHS in the present study is
unlikely to be attributable to the inhibition of sodium channels
because NaHS does not inhibit the presynaptic fiber volleys (Fig.
3B, asterisks). The lethal concentration of H,S in the brain is only
twice as much of an endogenous concentration of H,S in the rat
(Warenycia et al., 1989a). The suppressive effect of H,S on
synaptic transmission in the CNS may be partly responsible for the
dizziness and unconsciousness caused by acute sublethal H,S
exposure (Reiffenstein et al., 1992).
An additional experiment to support the role of endogenous
H,S in the induction of LTP is to test whether the induction of
s I
Weak tetanus
Weak tetanus
1 I I I 1 1 I
-10 0 10 20 30 40
50 60 70
Time (min)
Figure
8. Pretreatment with D’IT does not occlude the LTP-facilitating
effect of H,S. DTP (1 mM)
was applied during the time indicated by
solid bars.
After treatment with DTT, a weak tetanic stimulation (1.5 pulses at 100 Hz)
induced a small LTP. NaHS (130
PM)
with a weak tetanic stimulation (1.5
uulses at 100 Hz) induced an additional LTP even after treatment with DTT.
A weak tetanic stimulation by itself does not cause LTP.
Abe and Kimura
l
Hydrogen Sulfide as an Endogenous Neuromodulator
J. Neurosci., February 1, 1996, 76(3):1066-1071 1071
LTP is blocked by the inhibitors of ‘HzS production. Although
hydroxylamine and amino-oxyacetate suppress the production of
HzS (Fig. 2), they could not be used to test the role of endogenous
HzS in LTP for the following reasons. Amino-oxyacetate (0.5-l
mM) suppresses the baseline field EPSP by 23.3 + 4.7% (n = 4).
In addition to the inhibitory effect on CBS, hydroxylamine pro-
duces NO (Southam and Garthwaite, 1991). AdoMet, a specific
activator for CBS, did not significantly potentiate the induction of
LTP because it is unable to enter the cell. In addition, AdoMet
must be added to brain homogenates at least 15 min before the
reaction with CBS, suggesting that it only activates CBS with time.
The development of more specific and potent inhibitors for HzS-
producing enzymes is required.
In addition to changes in enzyme activity, there are at least two
substances that change the concentration of HzS; they are cysteine
and AdoMet. Cysteine is a source of HzS. Its concentration may
be changed when the glutamate/cystine transporter (Murphy et
al., 1989) is locally inhibited by the increased extracellular gluta-
mate, an excitatory neurotransmitter. AdoMet activates CBS re-
sulting in the increase in HzS production (Fig. 2), and the con-
centration of AdoMet is changed by testosterone (Manteuffel-
Cymborowska et al., 1992). Therefore, it is possible that H,S is
involved in the modulation of synaptic activities regulated by
steroid hormones and neurotransmitters.
It can be concluded that HzS is produced in the brain largely by the
activity of CBS and that HzS may be involved in associative learning.
REFERENCES
Aizenman E, Lipton DA, Loring RH (1989) Selective modulation of
NMDA responses by reduction and oxidation. Neuron 2:1257-1263.
Beauchamp RO, Bus JS, Popp JA, Boreiko CJ, Andjelkovich DA (1984)
A critical review of the literature on hydrogen sulfide. Crit Rev Toxic01
13:25-97.
Bliss TVP, Collingridge GL (1993) A synaptic model of memory: long-
term potentiation in the hippocampus. Nature 361:31-39.
Braunstein AE, Goryachenkova EV, Tolosa EA, Willhardt IH, Ye-
fremova LL (1971) Specificity and some other properties of liver serine
sulphhydrase: evidence for its identity with cystathionine P-synthase.
Biochim Biophys Acta 242:247-260.
Bredt DS, Snyder SH (1992) Nitric oxide, a novel neuronal messenger.
Neuron 8:3-11.
Collingridge GL, Kehl SJ, McLennan H (1983) Excitatory amino acids in
synaptic transmission in the Schaffer collateral -commissural pathway of
the rat hippocampus. J Physiol 334:33-46.
Erickson PF, Maxwell IH, Su IJ, Baumann M, Glode LM (1990) Se-
quence of cDNA for rat cystathionine y-lyase and comparison of de-
duced amino acid sequence with related Escherichia
coli
enzymes.
Biochem J 269:335-340.
Finkelstein JD, Kyle WE, Martin JJ, Pick AM (1975) Activation of
cystathionine synthase by adenosylmethionine and adenosylethionine.
Biochem Biophys Res Commun 66:81-87.
Garthwaite J (1991) Glutamate, nitric oxide and cell-cell signalling in the
nervous system. Trends Neurosci 14:60-67.
Goodwin LR, Francom D, Dieken FP, Taylor JD, Warenycia MW, Rei-
ffenstein RJ, Dowling G (1989) Determination of sulfide in brain tissue
by gas dialysis/ion chromatography: postmortem studies and two case
reports. J Anal Toxicol 13:105-109.
Griffith OW (1987) Mammalian sulfur amino acid metabolism: an over-
view. In: Methods in enzymology 143, pp 366-376. New York:
Academic.
Haley JE, Wilcox GL, Chapman PF (1992) The role of nitric oxide in
hippocampal long-term potentiation. Neuron 8:211-216.
Harris EW, Ganong AH, Cotman CW (1984) Long-term potentiation in
the hippocampus involves activation of N-methyl-o-aspartate receptors.
Brain Res 323:132-137.
Hebb DO (1949) The organization of behavior. New York: Wiley.
Kang H, Schuman EM (1995) Long-lasting neutrophin-induced en-
hancement of synaptic transmission in the adult hippocampus. Science
267:1658-1662.
Kombian SB, Reiffenstein RJ, Colmers WF (1993) The actions of hydro-
gen sulfide on dorsal raphe serotonergic neurons in vitro. J Neuro-
physiol 70:81-96.
Kraus JP (1990) Molecular analysis of cystathionine P-synthase - a gene
on chromosome 21. In: Molecular genetics of chromosome 21 and
Down syndrome, pp 201-214. New York: Wiley.
Maines MD (1988) Heme oxygenase: function, multiplicity, regulatory
mechanisms, and clinical applications. FASEB J 2:2557-2568.
Manteuffel-Cymborowska M, Chmurzynska W, Grzelakowska-Sztabert B
(1992) Tissue-specific effects of testosterone on S-adenosylmethionine
formation and utilization in the mouse. Biochim Biophys Acta
1116:166-172.
Mudd SH, Levy HL, Skovby F (1989) Disorders of transsulfuration. In:
The metabolic basis of inherited disease (Striver CR, Beaudet AL, Sly
WS, Valle D, eds), pp 693-734. New York: McGraw-Hill.
Murphy TH, Miyamoto M, Sastre A, Schnaar RL, Coyle JT (1989)
Glutamate toxicity in a neuronal cell line involves inhibition of cystine
transport leading to oxidative stress. Neuron 2:1547-1558.
O’Dell TJ, Hawkins RD, Kandel ER, Arancio 0 (1991) Tests of the roles
of two diffusible substances in long-term potentiation: evidence for
nitric oxide as a possible early retrograde messenger. Proc Natl Acad Sci
USA 88:11285-11289.
Palmer RMJ, Ashton DS, Moncada S (1988) Vascular endothelial cells
synthesize nitric oxide from L-arginine. Nature 333:664-666.
Reiffenstein RJ, Hulbert WC, Roth SH (1992) Toxicology of hydrogen
sulfide. Annu Rev Pharmacol Toxic01 32:109-134.
Rfeffer M, Ressler C (1967) p-Cy anoalanine, an inhibitor of rat liver
cystathionase. Biochem Pharmacol 16:2299-2308.
Savage JC, Gould DH (1990) Determination of sulfides in brain tissue
and rumen fluid by ion-interaction reversed-phase high-performance
liquid chromatography. J Chromatogr 526:546-545. - -
Schuman EM, Madison DV (1991) A requirement for the intercellular
messenger nitric oxide in long-term potentiation. Science 254:1503-1506.
Southam E, Garthwaite J (1991) Comparative effects of some nitric oxide
donors on cyclic GMP levels in rat cerebellar slices. Neurosci Lett
130:107-111.
Stevens CF, Wang Y (1993) Reversal of long-term potentiation by inhib-
itors of haem oxygenase. Nature 364:147-149.
Stipanuk MH, Beck PW (1982) Characterization of the enzymic capacity
for cysteine desulphhydration in liver and kidney of the rat. Biochem J
2061267-277.
Swaroop M, Bradley K, Ohura T, Tahara T, Roper MD, Rosenberg LE,
Kraus JP (1992) Rat cystathionine p-synthase. J Biol Chem 267:
11455-11461.
Tang L-H, Aizenman E (1993) The modulation of N-methyl-o-aspartate
receptors by redox and alkvlating reagents in rat cortical neurones in
vitro. J Physiol (Land) 465:~303-323. -
Ubuka T, Umemura S, Yuasa S, Kinuta M, Watanabe K (1978) Purifi-
cation and characterization of mitochondrial cysteine aminotransferase
from rat liver. Physiol Chem Phys Med NMR 10:483-500.
Uren JR. Raain R. Chavkovskv M (1978) Modulation of cvsteine metab-
olism in n&e - ‘effecis of p;opar’gylglycine and L-cyst(ejine-degrading
enzymes. Biochem Pharmacol 27:2807-2814.
Verma A, Hirsch DJ, Glatt CE, Ronnett GV, Snyder SH (1993) Carbon
monoxide: a putative neural messenger. Science 259:381-384.
Warenycia MW, Goodwin LR, Benishin CG, Reiffenstein RJ, Francom
DM, Taylor JD, Dieken FP (1989a) Acute hydrogen sulfide poison-
ing: demonstration of selective uptake of sulfide by the brainstem
by measurement of brain sulfide levels. Biochem Pharmacol
38:973-981.
Warenycia MW, Steele JA, Karpinski E, Reiffenstein RJ (1989b) Hy-
drogen sulfide in combination with taurine or cysteic acid reversibly
abolishes sodium currents in neuroblastoma cells. Neurotoxicology
10:191-200.
Zhuo M, Small SA, Kandel ER, Hawkins RD (1993) Nitric oxide and
carbon monoxide produce activity-dependent long-term synaptic en-
hancement in hippocampus. Science 260:1946-1950.
