Choline Modulation of the Aβ P1-40 Channel Reconstituted into a Model Lipid Membrane

Nicotinic acetylcholine receptors (AChRs), implicated in memory and learning, in subjects affected by Alzheimer's disease result altered. Stimulation of α7-nAChRs inhibits amyloid plaques and increases ACh release. β-amyloid peptide (AβP) forms ion channels in the cell and model phospholipid membranes that are retained responsible in Alzheimer disease. We tested if choline, precursor of ACh, could affect the AβP1-40 channels in oxidized cholesterol (OxCh) and in palmitoyl-oleoyl-phosphatidylcholine (POPC):Ch lipid bilayers. Choline concentrations of 5 × 10−11 M–1.5 × 10−8 M added to the cis- or trans-side of membrane quickly increased AβP1-40 ion channel frequency (events/min) and ion conductance in OxCh membranes, but not in POPC:Ch membranes. Circular Dichroism (CD) spectroscopy shows that after 24 and 48 hours of incubation with AβP1-40, choline stabilizes the random coil conformation of the peptide, making it less prone to fibrillate. These actions seem to be specific in that ACh is ineffective either in solution or on AβP1-40 channel incorporated into PLMs.


Introduction
Alzheimer's disease (AD) is an age-related neurodegenerative disorder that is characterized by a progressive loss of memory and deterioration of higher cognitive functions. The wealth of evidence of early and major synaptic damage and loss [1,2], correlated with the severity of dementia [3][4][5], considers AD to be a disorder of synaptic function [6,7].
Cholinergic neurons show particular vulnerability in the ageing process, which causes chronic deterioration of the brain component. It is a widespread belief among researchers that AβP peptides are involved in the loss of cholinergic neurons from the basal forebrain area in AD. Moreover, the mechanisms by which AβP peptides influence/cause degeneration of basal forebrain cholinergic neurons and/or the cognitive impairment characteristic of AD remain obscure, although there is little doubt that early synaptic pathology precedes the clinical symptoms of AD [8,9].
Several hypotheses have been formulated to explain the onset of this disease such as: alterations in choline uptake, choline transport, impaired acetylcholine release, deficits in expression of nicotinic, and muscarinic receptors and channel formation by AβP oligomers. An increase in choline flux across the membranes of neuronal cells exposed to AβP has also been hypothesized to contribute to the selective vulnerability of cholinergic neurons in AD [10].
AβP is an amphiphilic peptide with a hydrophilic Nterminal domain (residues 1-28) and a hydrophobic Cterminal (residues 29-40 (−42)), the latter corresponding to a part of the transmembrane domain of APP. CD studies have shown that AβP changes its conformation from lipid-free AβP in order to bind to phospholipid vesicles. AβP consists of 48.9% random-coil, 23.5% β-sheet, and 1.7% alpha-helix; while when reconstituted in phospholipid vesicles made up of 1,2-dimyristoyl-sn-glycero-3phosphocholine, both β-sheet and α-helix contents increase by about 7%, that is, the β-sheet increased to 31.2% and the α-helix to 9.5%, respectively [11]. It has been demonstrated that AβP forms ion channels in the cell and model phospholipid membranes [12][13][14]; a P K + /P Cl − permeability ratio of 11, with calcium permeability blocked by tromethamine and aluminium, has been reported [12]. Moreover, AβP disrupted calcium homeostasis and increased calcium intracellular concentrations; these events may be responsible for cellular toxicity [15]. Recent studies indicate that the peptide's ability to form ion channels depends on its conformational structure and on the peptide/lipid structure in the physiological environment. Some authors have found that membrane components, such as cholesterol (Ch) and gangliosides, alter the affinity of AβP for phospholipid membranes. In fact, Ch and gangliosides, once associated with phospholipid membranes, lead to an increase in βsheet content and/or the rate of aggregation of AβP [16].
On the other hand, other authors have shown that, when AβP was added to a 33% Ch-containing 1,2-dimyristoyl-snglycero-3-phosphocholine vesicle, the structure of AβP was drastically altered, that is, the β-sheet structure decreased to zero while the α-helix increased to 58.8% [11]. In addition, alterations to the soluble Ch concentration and/or in Ch biosynthesis have been shown to affect the normal processing of APP, both in vivo and in vitro [17]. In a previous study, we investigated the role played by membrane composition on the interaction and self-assembly of AβP1-40 during pore formation in PLMs. This study showed that AβP1-40 has a higher propensity to form channels in OxCh PLM (as compared to other neutral phospholipid PLMs, whether they contain sterols or not), where the channels present high conductance, high frequency, anion selectivity, and long lifetime [14].
In this study, we investigated the effects of choline on the AβP1-40 ion channel in lipid bilayer membranes made up of OxCh and of POPC:Ch bilayer. OxCh membranes were used due to the evidence of a Ch-rich domain both in eukaryotic plasma membranes-such as brain and blood vessels-and in aged membrane plaque. Besides, AβP1-40 shows a higher affinity to Ch and its oxidation products than to phospholipids [14].

Single Channel Measurement.
Channel activities were recorded in a lipid bilayer membrane made up of OxCh in n-decane (1:1, v:v) (Fluka) or POPC:Ch (65 : 35, w/w) in 1% n-decane. OxCh was obtained following the method of Tien et al. [18]. Bilayers were formed across a 300 μm hole in a Teflon partition separating two Teflon chambers (volume 4000 μl) which held symmetrical 50 mM KCl solutions, pH = 7, temperature 23 ± 1 • C. The aqueous solutions were used unbuffered. The salts used in the experiments were of analytical grade. In all peptide experiments performed, the conductance and capacitance of each membrane was tested by applying a voltage of ±100 mV for 10-15 minutes under stirring to ensure that the membrane was stable.
A stock solution of AβP1-40 (Sigma or EZBiolab) was prepared by dissolving AβP1-40 powder (0.1 mg) in 100 μl of bidistilled sterile water under stirring for 3 minutes. From this solution, 5 μl were withdrawn and diluted in 45 μl of bidistilled sterile water under stirring for 3 minutes. Both solutions were stored at −20 • C until use and 8.7 μl of the second solution was added to the cis-side of the membrane, to obtain the final concentration of 5×10 −8 M. After AβP1-40 ion channel formation, few μl of scalar dilution of choline (Fluka) stock solutions (1×10 −3 ; 1×10 −5 ; 1×10 −7 ; 1×10 −8 M) were added to obtain the final desired concentration on the cis or trans side of the Teflon chambers. The solutions were stirred after each addition of AβP1-40 or choline for 1 minute.
In single-channel experiments, the membrane current was monitored with an oscilloscope and recorded on a chart recorder for further data analysis by hand. The cis and trans chambers were connected to the amplifier head stage by Ag/AgCl electrodes in series with a voltage source and a highly sensitive current amplifier. The single-channel instrumentation had a time resolution of 1-10 msec depending on the magnitude of the single-channel conductance. The polarity of the voltage was defined according to the side where AβP was added (the cis-side). A trans-negative potential (indicated by a minus sign) means that a negative potential was applied to the trans side, the compartment opposite the one where AβP was added.
The phenomenology of AβP1-40 incorporation and choline action was studied as follows: (i) to define the voltage-dependent characteristics of AβP1-40, we measured the amplitude of channel events at each membrane potential applied, (ii) to define the channel lifetime, from records extending over prolonged periods, the channel duration was measured considering the time between the opening and closing of each channel. The single channel data were obtained from at least two and sometimes four experiments (more than 100 single channels for each experiment) performed on different days. A histogram of channel conductance distribution for each experiment was constructed and fitted by a Gaussian distribution function (Graph-Pad Prism version 3.0; GraphPad Software Inc., http://www.graphpad.com/). Results are expressed as mean ± SE. The average lifetime of the conductance unit was estimated by the formula where N is the number of channels that remain open for a time equal to or greater than a certain time t, A 1 and A 2 are the zero time amplitudes, and τ 1 and τ 2 are related to the fast and slow components of the time constant, respectively. The single-exponential distribution is included in the formula (A 2 = 0). To choose between the two models, we performed an appropriate statistical test (F-test, Graph Pad Prism version 3.0; Graph Pad Software, Inc, http://www.graphpad.com/) (iii) to identify the charge on the ion carrying the current, we measured the shift in the reversal potential induced by a change from a symmetrical to an asymmetrical KCl solution system. When the membrane conductance reached a virtually stable value, after addition of AβP1-40 to the cis chamber,we added choline to the cis or trans side. After stabilizing the membrane conductance, the KCl concentration was raised to 100 mM by adding concentrated salt solution to the cis side of the chamber.
The reversal potential was determined by changing the holding potential of ±4 mV step by step, and the potential at which the current was zero was taken as the reversal potential for the open channel.
The permeability ratio was calculated using the Goldman-Hodgkin-Katz equation

Effect of Choline on AβP1-40 Channel Conductance.
In this study, we evaluated the effect of choline on the AβP1-40 ion channel incorporated into PLMs made up of OxCh or POPC : Ch. First of all, in order to exclude any nonspecific and destabilizing effect of choline per se on PLMs used, we performed experiments by leaving choline in the medium facing the membrane for up to 24 hours. The stability of the PLM was tested by applying a voltage of ±100 mV for 10-15 minutes under stirring and monitoring constant values for conductance (25 pS) and capacitance (0.32 μF/cm 2 ); choline, over the range of concentrations used in the present work, caused no variations in membrane conductance and capacitance in bare membranes.
The incorporation of AβP1-40 (added to the cis side of the medium facing the membrane) into the lipid bilayer leads to nonrandom discrete square events that fluctuate between conductive and nonconductive states, compatible with channel-type openings, and closures with different conductance levels, lifetime and frequency. The pattern is similar to that found in our previous study [14], that is, we found higher conductance for the AβP1-40 channel incorporated into OxCh PLM. Figure 1 shows examples of chart recordings of AβP1-40 channel formation in OxCh PLMs.
When the AβP1-40 channels fluctuated in the open state in OxCh PLMs, choline (5×10 −11 M) was added to the cis or trans side of the medium facing the membrane. In a short time (about 5/30 minutes, cis/trans, resp.) choline determined an increase in AβP1-40 ion channel activity characterized by more frequent multiple levels of conductance. Furthermore, these patterns were more evident at positive applied voltages. Figure 2 reports the histograms of conductance distribution for each experimental condition used in OxCh PLMs. All the histograms show single-peaked conductance distributions. The central value of conductance (Λ c ) obtained by the Gaussian best-fit characterizes the conductance state of AβP1-40 channels in various membranes. Figure 3(a) reports the central value of Λ c ± SE at the different applied voltages for different experimental conditions in OxCh PLMs. In the presence of choline, no matter which side it is added, the Λ c values are significantly higher at positive applied voltages than at negative ones (P < .0001 and P ≤ .0012 for choline added on the cis and on the trans side, resp.). If we compare the effect of choline on AβP1-40 channel conductance, it can be seen that choline added on the trans side increases channel conductance more than when added to the cis side. Figure 4 shows typical examples of single-channel chart recordings of AβP1-40 channel formation with associated conductance distribution histograms in OxCh PLMs without and with successive additions of choline to the cis-side of the medium facing the membrane or when choline in a ratio of about 1: 3 to the peptide (1.5×10 −8 M) was added when the channel was fluctuating in the open state. It can be seen that single-channel activity sometimes occurred in highly variable steps, yet the frequency of channels increases by increasing the choline concentration; we also observed alternating periods of paroxystic channel activity, during which it is impossible to make a rigorous analysis of the number of channels, followed by quiescent periods that are more frequents at higher choline concentration and often followed by membrane destabilization until rupture. Furthermore, a t-test showed that the central channel conductance of AβP1-40 is statistically increased when choline at different concentrations is added to the cis-side (see capture of Figure 4). Yet, the same result was found when choline at different concentrations was added to the trans side (data not shown).  Our results indicate that the presence of choline (on the cis or trans side) does not modify the voltage dependence of AβP1-40 channels in OxCh or POPC : OxCh PLMs (Figures  3(a) and 6(a)).

Effect of Choline on AβP1-40 Channel Frequency and
Lifetime. The AβP1-40 frequency values are generally higher at positive than those at negative applied voltages in OxCh PLMs (Figure 3 (Figure 6(b)). Another parameter used to characterize a channel is its lifetime. Single-channel current recordings with a conspicuous number of channels were analysed to obtain cumulative open-state lifetime distributions that are reported for the different experimental conditions. In OxCh PLMs, analysis of the open-time distributions for AβP1-40 single channels is reported in Table 1, where the functions with statistically significant better description are indicated. It can be noted that at positive applied voltages the fast channel lifetime component prevails (P < .05), whereas at negative applied voltages the channel manifests both the fast and slow components of lifetime (P < .05), except for at −20 mV where it does not clearly distinguish between single and double exponentials (P = .102).
When choline is on the cis or trans side, the results of the open time distribution analysis for all positive and for low negative applied voltages indicate a statistically significant better description (P < .05) for two-exponential functions; except for −40 and −60 mV (choline cis) and −60 mV (choline trans) where there is a statistically significant better description (P < .05) for one-exponential functions, however the values of τ 1 in the presence of choline are higher than that of AβP1-40. The prevalence of dual channel populations or the higher values of τ 1 clearly indicates that choline (either cis or trans) seems to stabilize the AβP1-40 single channel.
In POPC : Ch PLMs, analysis of the open-time distribution for AβP1-40 single channels does not clearly distinguish between single and double exponentials (at applied voltages in this study) (P ≥ .078). When choline is on the cis side, the results of open-time distribution analysis indicate a statistically significant better description (P < .05) for oneexponential functions, except at applied voltages of 100 mV where it does not clearly distinguish between single and double exponentials (P = .39). When choline is on the trans side, the results of open time distribution analysis give a statistically significant better description (P < .05) for a one-exponential function at an applied voltage of 60 mV, whereas it does not clearly distinguish between single and double exponentials at an applied voltage of 80 mV (P = .37). At an applied voltage of 100 mV, the number of channels is not conspicuous enough to provide a reliable analysis of open-time distribution (Table 2). In any case, the addition of choline, either on the cis or the trans side, does not seem to modify the lifetime and stability of the ion channel. In OxCh PLMs, the P K + /P Cl − calculated by (2) was 0.50/0.38 cis/trans, respectively.
Approximately the same result was obtained with the I-V curve (Figure 7), where the amplitude of the channel events at each membrane potential was used. In fact, the reversal potential was 6.42 mV with choline in the cis chamber and 6.87 mV with choline in the trans chamber; the selective ratio P K + /P Cl − was 0.46/0.43 cis/trans, respectively. These results seem to indicate that the ion selectivity toward anions of the AβP1-40 channel in OxCh PLMs [14] is not modified by the presence of choline.
The ion selectivity of the AβP1-40 channel in POPC:Ch PLMs remains anionic; in fact, the reversal potential was 9.76 mV and the P K + /P Cl − was 0.28. When choline was present on the cis/trans side the reversal potential was 5.76/7.76 mV and the P K + /P Cl − was 0.50/0.38 respectively. Approximately the same result was obtained with the I-V curve (Figure 8). In fact, the reversal potential was 8.99 mV in the absence of choline, with a selective ratio P K + /P Cl − of 0.32. The reversal potential was 6.67 mV with choline in the cis chamber and 6.47 mV with choline in the trans chamber; with a selective ratio P K + /P Cl

Effect of Choline on AβP1-40 Secondary Structure.
To test whether choline modifies the secondary structure of AβP1-40, we carried out CD experiments using AβP1-40 samples in the absence or in the presence of choline in a molar ratio of 1 : 100 and 1 : 1000 choline : AβP. Figure 9 shows the CD spectra of AβP1-40 without and with choline measured after 5 minutes (T0), 24 hours (T24), and 48 hours (T48).
The CD spectra qualitatively indicate that AβP1-40 conformation in an aqueous environment is predominantly β-sheet and random coil, while the α-helical content is very small. Our results are in line with other authors' findings [11]. The features of the spectra show that choline at the two concentrations used does not modify the secondary structure of the peptide. Moreover, it increases the signal intensity after 24 and 48 hours of incubation. This could mean    that choline stabilizes the AβP1-40 structure, counteracting peptide aggregation. The permeability coefficients of [ 3 H]choline across the OxCh membrane, undoped and doped with AβP1-40, were: 47.1 ± 2.14 and 43.8 ± 7.9 (cm × sec −1 ×10 −6 ), respectively.

Choline Permeability across Undoped and Doped OxCh
This data indicates that choline, despite showing a high permeability coefficient across OxCh membranes, is unable to cross the AβP1-40 channel.

Discussion
Early evidence showed that AβP1-40 is initially deposited on the membranes and acts as a "seed" for the formation of fibrils [19], although the involvement of AβP fibrillation in vivo in the process of neurodegeneration and the development of AD is not an irrefutable criterion for defining the pathology. Recent studies have shown that AβP induces toxicity on cholinergic neurons in vitro and in vivo by increasing, via distinct mechanisms, Ach turnover. In neuronal cells, choline is a precursor of ACh, but also a product of ACh hydrolysis after the neurotransmission process. The mechanisms involved in cholinergic dysfunction are under intense investigation because of their potential therapeutical implications [20]. Although the basal extracellular choline concentration is higher (about 5 μM) [21] as compared to the concentration used in the present study, the sensitivity of our experimental system, suitable for high-resolution ion current recordings, does not allow us to use choline concentration higher than 1.5 ×10 −8 M. In fact, the activation of AβP channels by choline at concentration higher than 1.5 ×10 −8 M generates an electrical signal that is so large that it exceeds the maximum level of detection by the recording system furthermore, these conditions will lead to membrane instability eventually resulting in membrane breakage. The human brain contains as much as 25% of the total pool of Ch, mainly concentrated in the myelin sheath. However, there are also considerable amounts of Ch in neuronal plasmalemma and in lipid rafts. One interesting aspect of the Ch molecule seems to be its affinity for many proteins such as porins [22,23], for peptides such as magainin-2 [24], AβP1-40 and AβP1-42 [14], and for peptides associated with myelin; examples of Ch-dependent proteins are prominin, synaptophysin, platelet-derived growth factor receptor, hemolysin, acetylcholine-receptor, peripheral myelin protein 22, and prions [25][26][27][28][29][30][31]. In particular, it has been demonstrated that Ch at a concentration of 33% in model membranes and in vesicles can favour the transition of AβP1-40 and human cytoplasmic domain of myelin protein zero into alpha helix [32]. On the other hand, membrane Ch induces the structural conformation of many peptides for incorporation. Results in OxCh and POPC : Ch planar lipid membranes (PLMs) confirm our previous findings [14] and further indicate the role played by membrane composition in the alpha-helix structure induction mechanism needed for AβP incorporation into membranes. Furthermore, we found sensitivity to choline addition of AβP channels incorporated into OxCh PLMs; indeed, choline does not modify some channel characteristics, such as voltage sensitivity, voltagedependence, and selectivity, but increases channel conductance, frequency, and stability, as indicated by the presence of two channel life-time components.
In the Type I AβP1-40 channel model proposed by Durell et al. [33], the amino acids able to form hydrogen bonds could be R5, D7, E11, K16, and D23. This has been invoked to explain the cation selectivity of the AβP1-40 channel. The AβP1-40 channel incorporated in our membranes is anion selective. Choline does not modify the anion selectivity of AβP channel incorporated in OxCh membrane, the aggregated configuration of AβP peptide could be different, although this channel shows the same diameter, as reported in our previous paper [14], found in the theoretical calculation of the model proposed by Durell et al. [33].
Incorporation in POPC PLMs containing 33% Ch seems to be driven by hydrophobic interaction, as the high hydrophobic amino acid content in the C-terminus of AβP1-40 would suggest. Furthermore, our results seem to indicate that the AβP1-40 channel is smaller than that formed in OxCh membranes, which could reflect a lower number of AβP1-40 molecules assembled to form ion channels. However, it is worth noting that in POPC : Ch PLMs, the channel maintains its anionic selectivity, thus pushing the membrane into a hyperpolarized state.
Our CD results, obtained in an aqueous environment, indicate that choline does not modify AβP1-40 peptide conformation but rather stabilizes it in a random coil conformation that is less prone to fibrillate. This result is consistent with the notion that the role played by choline is to counteract peptide aggregation, by maintaining it in the stabilized random coil form, so that in the membrane alpha-helices will arise in the peptide that tend to pack together to form ion channel rather than remain dispersed. Hydrogen bonds and ion pairs can be used to drive the association of local regions in the helices [34]. It is interesting to note that Kar et al. [9] found a concentration-dependent inhibition of high-affinity choline uptake by AβP in rat hippocampal slides, indicating a direct interaction between AβP and choline. AβP1-40 present on the extracellular side of PC12 cells as a monomer has been found to increase the conductance of choline trapped inside cells. The authors [35] suggest that AβP1-40 peptide acts as a choline carrier or that it modifies an endogenous ion channel or transporter. However, a suggestive possibility could be that choline crosses the AβP1-40 channel. To verify this possibility, permeability experiments on OxCh PLMs have been carried out. In our experimental conditions, where there are no other cellular components except for the AβP1-40 channel, choline does not cross the AβP1-40 channel.
Study of the effects of ACh on AβP1-40 by means of CD indicates that ACh has no effect on peptides in solution; moreover, single channel conductance decreases when ACh is added on the cis side when the channel is open; on the other hand, in the same experimental conditions, ACh addition on the trans side did not modify channel conductance [36]. This indicates a specific effect of choline on AβP1-40.

Conclusions
Our present findings support and extend the emerging concept that cholesterol can easily incorporate AβP1-40, thus favouring its clearance; therefore, by removing the peptide from the environment, fibrillation will be avoided. Furthermore, the anionic nature of the AβP1-40 channel formed in cholesterol-containing PLMs could be protective for the cells. Choline specifically accelerates this process. The greater effect of choline on depolarized membrane states found in this work could be of relevance for synaptic activity in that the higher ionic conductance, frequency and lifetime, as well as the anionic selectivity, all contribute to returning the membrane potential to basal condition. This could exert a kind of modulation on the synapses.
It may be speculated that owing to the high choline permeability found across the OxCh membrane, and the in vivo facilitated/active choline transport, ACh synthesis takes place quickly. The increase in ACh delivery and/or ACh esterase modulation may be beneficial to synaptic activity. Obviously, these aspects are merely speculative because we do not have any evidence that this occurs in vivo.