(Received for publication, August 9, 1996, and in revised form, September 19, 1996)
From the Division of Neuroscience and Department of Psychiatry, UCLA School of Medicine, Neuropsychiatric Institute, Brain Research Institute, University of California, Los Angeles, California 90024-1759 and the West Los Angeles Veterans' Administration Medical Center, Los Angeles, California 90024
Prions cause neurodegenerative disease in animals and humans. Recently it was shown that a 21-residue fragment of the prion protein (106-126) could be toxic to cultured neurons. We report here that this peptide forms ion-permeable channels in planar lipid bilayer membranes. These channels are freely permeable to common physiological ions, and their formation is significantly enhanced by "aging" and/or low pH. We suggest that channel formation is the cytotoxic mechanism of action of amyloidogenic peptides found in prion-related encephalopathies and other amyloidoses. The channels reported here are large enough and nonselective enough to mediate cell death through discharge of cellular membrane potential, changes in ionic homeostasis, and specifically, influx of calcium, perhaps triggering apoptosis.
Prions are proteinaceous infectious agents that cause
transmissible and genetic neurodegenerative diseases, such as scrapie and bovine spongiform encephalopathy in animals and kuru and
Creutzfeldt-Jakob disease in humans (for review see Ref. 1). The
cellular form of the prion protein (PrPc),1
which does not cause disease, can be converted into a pathogenic form
(PrPsc), and this transition appears to take place without any covalent
modifications to the molecule (1). Regions of the prion protein that
are predicted to be -helical actually form
-sheet when
synthesized and aggregate into amyloid fibrils similar to those found
in prion induced encephalopathies (2). Recent studies have implicated
channel formation as a possible pathogenic mechanism of other
amyloidoses, such as Alzheimer's disease (3, 4) and type II diabetes
mellitus (5). Previous studies have shown that a 21-amino acid fragment
of the prion protein (PrP 106-126; Fig. 1) could be
toxic when chronically exposed to primary rat hippocampal cultures at
micromolar concentrations (6). Other peptides from the prion protein
were not found to be neurotoxic. Amyloid deposition is a frequent but
not universal feature of prion-related encephalopathies (1). A major
question in prion induced diseases and other amyloidoses has been
whether amyloid plays a role in cell death. Recent studies have
suggested that the full-length
-amyloid peptide (AB) from
Alzheimer's disease or the neurotoxic fragment AB (25-35) could form
ion-permeable channels in lipid bilayer membranes (3, 4). Additionally, we have recently shown that the amyloidogenic peptide amylin, found in
the islets of Langerhans of patients with type II diabetes mellitus, can also form ion permeable channels at cytotoxic
concentrations (5). We therefore set out to examine the effects of the
neurotoxic fragment of the prion protein on planar bilayer
membranes.
PrP 106-126 (>95% purity) was purchased from Bachem
Bioscience Inc. (King of Prussia, PA), and azolectin (soybean
phosphatide extract, granulated, 45% phosphocholine content) was
purchased from Avanti Polar Lipids, Inc. (Birmingham, AL). Only a
single peak was observed on high pressure liquid chromatography
analysis of the peptide in three solvent systems. Lyophilized PrP
106-126 was dissolved in deionized water at a concentration of 2 mg/ml, distributed into 25-µl aliquots, and stored at 200 °C.
The peptide was thawed before addition to the lipid bilayer membranes
and never frozen again.
Solvent-free and painted planar lipid bilayer membranes were formed as described previously (5, 7, 8). Usually, after the initial incorporation of PrP 106-126, free peptide in the aqueous solution was washed out. Membranes used in experiments were stable and had conductances of less than 10 pS up to voltages of ± 100 mV for a period of at least 10 min prior to peptide addition.
Recording EquipmentVoltage clamp conditions were employed and contact with the aqueous phases was made using Ag/AgCl electrodes with agar salt bridges. Electrode asymmetry was always less than 1 mV. Membrane formation was verified by monitoring membrane capacitance and resistance. Data were digitized and stored on VHS tape and played back for later analysis. An Axopatch 1C amplifier with head stage CV-3B was used for measuring membrane current. For data acquisition, a digital tape recorder and video cassette recorder allowed recording of large amounts of data. A storage oscilloscope was used for monitoring membrane capacitance and single-channel recordings. The cis-solution (peptide containing) was taken as the virtual ground, and the sign of the membrane voltage corresponded to the trans-side of the membrane.
At neutral pH, PrP 106-126 induced a conductance in solvent
containing membranes when added to concentrations greater than or equal
to 20 µM (Fig. 2A). The induced
conductance was voltage-independent (Fig. 2B) and was due to
the formation of ion-permeable channels (Fig.
3A). Several single channel conductances were
observed, including the most common conductances of 20, 40, and 60 pS
(Fig. 3B). In a 10-fold gradient of sodium chloride across
the membrane, the reversal potential was 15 millivolts (dilute side
positive), indicating that the channels were permeable both to sodium
and chloride (PNa/PCl
approximately equals 2.5). Other selectivity measurements indicated
that virtually all common physiologic ions were permeable through the
channel in the sequence Ca+2 > Na+ > K+ > Li+ > Rb+ > Cs+ > C1.
Macroscopic conductance induced by PrP
106-126 in planar lipid bilayers membranes. A, the increase
in membrane current (and therefore conductance) in a planar bilayer
membrane to which PrP 106-126 had been added (arrow) is shown. The
membrane voltage was held constant at 50 mV, and as the current
increased (indicated by downward deflection), the noise (due to channel
flickering) also increased. Note that the current steadily increased
after PrP 106-126 addition and reached a steady state only after 5-8 min. (not shown). Planar lipid membranes were formed as described below. PrP 106-126 (>95% purity) was purchased from
Bachem Bioscience Inc., and azolectin (soybean phosphatide extract,
granulated, 45% phosphocholine content) was purchased from Avanti
Polar Lipids, Inc. Lyophilized PrP 106-126 was dissolved in deionized
water at a concentration of 2 mg/ml, distributed into 25-µl aliquots, and stored at
20 °C. The peptide was thawed before addition to the
lipid bilayer membranes and never frozen again. Solvent-free painted
planar lipid bilayer membranes were formed as described previously (5,
8). The formation of solvent-free bilayers was carried out in a Teflon
chamber with two compartments (0.3 ml each) separated by a thin (30 µm) Teflon film with an aperture diameter of 150-200 µm. Initially
the level of salt solutions in both compartments was raised to a level
just below the aperture, and then 15 µl of a 10 mg/ml solution of
lipid in hexane was spread at the surface of the aqueous phase of both
compartments. 20 µl of 25 mg/ml hexadecane in n-pentane
was applied to the Teflon film. After evaporation of hexane and
pentane, the planar lipid membrane was made by raising the salt
solution levels over the aperture and joining the two compartment lipid
monolayers over the aperture. The painted bilayer membrane was formed
from 15 mg/ml solution of lipids in n-heptane at the end of
Teflon tubing with 500- or 1000 µ- diameter. The construction of the
chamber allowed substitution of the solution in one (cis)
compartment within several seconds. After the membrane had turned
black, the solution in the cis-side was substituted with
peptide-containing solution. Usually, after the initial incorporation
of PrP 106-126, free peptide in the aqueous solution was washed out.
Membranes used in experiments were stable and had conductances of less
than 10 pS up to voltages of ±100 mV for a period of at least 10 min prior to peptide addition. Voltage clamp conditions were employed, and
contact with the aqueous phases was made using Ag/AgCl electrodes with
agar salt bridges. Electrode asymmetry was always less than 1 mV.
Membrane formation was verified by monitoring membrane capacitance and
resistance. Data were digitized and stored on VHS tape and played back
for later analysis. An Axopatch 1C amplifier with head stage
CV-3B was used for measuring membrane current. For data acquisition, a
digital tape recorder and video cassette recorder allowed recording of
large amounts of data. A storage oscilloscope was used for monitoring
membrane capacitance and single-channel recordings. The
cis-solution was taken as the virtual ground, and the sign
of the membrane voltage corresponded to the trans-side of
the membrane. B, steady state I-V plot of the multichannel containing membrane induced by 40 µM prion 106-126 at pH
7.5. The current-voltage relation for a membrane that had been doped with 10 µM PrP 106-126 is shown. Note that the
relationship is essentially linear (Ohmic). Conditions were as in
a.
Aging of the prion peptide dramatically enhanced its channel formation
activity. Thus, PrP 106-126 aged for 9 days at room temperature in 100 mM NaC1 could increase membrane conductance at
concentrations as low as 0.1 µM, indicating an activity
increase of approximately 200-fold. For peptide aged for 3 days,
activity was increased by approximately 20-fold. Single channels were
also formed by this aged prion peptide, although there were some slight differences in the distribution of single channel conductances observed
(Fig. 4).
Channel activity could also be enhanced by acidic pH. At pH 4.5 to 5.0, PrP 106-126 could induce channel activity at a concentration of 1 µM (Fig. 5). Single channel conductances
of 20, 100, and 120 pS were observed most frequently (Fig.
5B).
Channels were irreversibly associated with the membrane. Extensive washing out of the aqueous compartment containing peptide had no effect on the peptide-induced conductance.
The channels reported here are clearly due to the PrP 106-126 itself. Under the conditions of our experiments, channels were never observed in the absence of added prion peptides. The channels formed rapidly upon peptide addition to the chamber and did so at concentrations comparable with those required for neurotoxicity (6). Furthermore, the concentrations of PrP 106-126 used are comparable with the concentrations needed for channel formation by other peptide channel formers (3, 4, 5, 7, 8, 9). We suggest that channel formation accounts for the neurotoxicity of PrP 106-126 and may account for the neuronal loss observed in prion-related encephalopathies. The relative lack of ionic selectivity of these channels suggests that they would impose a substantial leakage permeability upon the cell membrane. Neurons in particular would be highly vulnerable to a leakage conductance of this kind, because they must maintain a relatively tight plasma membrane to preserve electrical signaling along that membrane. Thus, inward leakage of sodium and calcium as well as outward leakage of potassium would place a severe metabolic load on the cell and also subject it to potential ionic toxicity. Furthermore, these ionic disturbances might potentially trigger apoptosis in cells. Alternately, this leakage conductance could lead to depolarization, which might trigger disturbances via influxes of calcium through voltage-dependent calcium channels or NMDA receptors (10, 11). Calcium influx might also be triggered through influx of sodium followed by sodium/calcium exchange. The voltage independence of the conductance induced by these channels suggests that they would be permanently and stably open in the membrane. The irreversibility of channel formation means that these channels would continue to impose a metabolic strain on cells once inserted.
The enhancement of activity seen in an acid environment suggests that
aggregation of the peptide contributes to channel formation. A similar
conclusion may be inferred from our results with aging of prion
peptides. It is unclear whether this aggregation in vivo might occur in an acidic endosomal compartment. Acid pH has also been
shown to convert PrP 106-126 from -helical to
-sheet
conformation (12). This might promote peptide aggregation, amyloid
formation, and channel formation. It has also been observed that
certain peptides from the prion proteins, such as PrP 109-122, can
induce the
-helix to
-sheet transition in other peptides from the
prion protein (13). This might be a model for the conversion of PrPc to
PrPsc in vivo. It has been noted that PrPsc aggregates on
its own, and it has been suggested that PrPc is converted to PrPsc most
likely in the endosomal pathway (15).
A significant implication of our findings is that there may be a close
relationship between amyloid formation, cytotoxicity, and channel
formation. We suggest that the -sheet structures assumed by amyloid
forming peptides may also be conducive to channel formation under the
appropriate conditions. This may lead to disruptions of ionic
homeostasis and even to cell death. The fact that small
-sheet-forming peptides from three different amyloid producing diseases can be cytotoxic and form similar ion permeable channels at
cytotoxic concentrations, strongly suggests that channel formation may
be an important cause of cytotoxicity in amyloidoses. Further studies
will be required to verify this hypothesis. Screening the channels
formed for drugs capable of blocking them might provide a means of
searching for candidate therapies to ameliorate the damage done in
these amyloidoses.