*2.1. Biochemical Characterisation of the Interaction between Aβ<sup>40</sup> and Aβ<sup>42</sup> with P-gp*

The baculovirus system provides high capacity expression of membrane proteins, including human P-gp with full retention of function including substrate binding [19–21]. Figure 1a demonstrates that P-gp was purified to near homogeneity from High-Five membranes and eluted specifically in the presence of 400 mM imidazole. The efficiency of reconstitution was routinely assessed as described previously [21] and the concentration of P-gp obtained in the present investigation was 0.092 ± 0.036 mg/mL (*n* = 4). The total yield was 234 ± 86 µg of purified P-gp per 100 mg of crude High-Five membrane, indicating that the expression was >0.2% of total membrane protein. These parameters are similar to previously published values [21].

The tryptophan quenching assay has been widely used in the field to describe the selectivity and apparent potency of ligand/substrate interaction with P-gp. The affinity is classified as apparent since it measures the binding step and subsequent interaction with a proximal tryptophan residue. Consequently, this assay was used to provide further evidence to support a putative interaction between Aβ peptides and P-gp. A recent publication from our group and collaborators using molecular docking [14] indicated that this interaction is complex, and the kinetics are likely to be slow. Consequently, the assay conditions were configured with an elevated temperature (37 ◦C) and a longer incubation time than typically adopted for ligand binding studies [22]. As shown in Figure 1b,c, both peptides were able to produce a dose-dependent quenching of the endogenous tryptophan fluorescence intensity, with no shift in the maximal wavelength. The extent of reduction in signal was 20–40% of that produced by the P-gp modulator nicardipine. Numerous researchers [19,23–25] have demonstrated considerable variability in the extent of tryptophan quenching by drugs and short peptide substrates of P-gp. These observations suggest that the Aβ peptides have a direct interaction with P-gp that is similar to established drug substrates and inhibitors. This is supported by an earlier study demonstrating that the Aβ peptides quench the fluorescence of an extrinsic probe (MIANS) covalently attached to P-gp [26]. MIANS was attached at the nucleotide-binding domains of P-gp and this indicates a long-distance allosteric interaction, whereas the intrinsic tryptophan quenching is thought to involve residues proximal to the drug binding site and central cavity [23].

**2021**, , x FOR PEER REVIEW 3 of 21

β

β β

β β μ μ β β μ **Figure 1.** The interaction between Aβ<sup>40</sup> and Aβ<sup>42</sup> peptides with purified, reconstituted P-gp. (**a**) Fractions obtained from the metal affinity purification of P-gp were separated on 8% SDS-PAGE and detected with Instant BlueTM protein stain. Fractions (40 µL) from each stage of the purification were loaded onto the gel. F<sup>T</sup> corresponds to the unbound detergent extracted material, 20 mM and 60 mM represent washing stages, and 400 mM is the elution phase. (**b**) Representative fluorescence emission spectra were obtained for the endogenous tryptophan residues of purified, reconstituted P-gp (15 µg) with excitation at 295 ± 10 nm at 37 ◦C. Spectra were recorded in the absence or presence of a series of Aβ<sup>40</sup> concentrations. (**c**) Fluorescence emission spectra obtained as in (**b**), but with varying concentrations of Aβ<sup>42</sup> peptide. (**d**) ATPase activity of purified, reconstituted P-gp (0.2–0.5 µg) in the presence of varying concentrations of nicardipine. The activity was measured over 40 min at 37 ◦C and normalised to the value obtained in the absence of drug (basal). Data were fitted by the general dose-response relationship using non-linear least squares regression and values represent mean ± SEM obtained from four independent preparations. (**e**) ATPase activity of purified, reconstituted P-gp (0.2–0.5 µg) in the presence of varying concentrations of Aβ<sup>40</sup> and Aβ<sup>42</sup> peptide. The activity was measured over 40 min at 37 ◦C and normalised to the value obtained in the absence of drug (basal). Data were fitted by a hyperbolic relationship using non-linear least squares regression and values represent mean ± SEM obtained from three independent preparations.

Binding to the transporter represents the first step in the process of translocation across the membrane and therefore an activity assay was chosen to further explore the interaction between Aβ peptides and P-gp. Transported substrates of P-gp increase the rate of ATP hydrolysis [27–30], whereas pure inhibitors reduce the activity [31,32]. Consequently, measuring substrate effects on ATPase activity provides a rigorous assessment of the interaction with P-gp at the level of initial binding and the subsequent coupling event. Figure 1d shows the dose-dependent stimulation of ATPase activity by nicardipine. The overall activity of purified, reconstituted P-gp was characterised with a basal level of 419 ± 30 nmol/min/mg and a maximal activity of 1815 ± 139 nmol/min/mg for the preparations generated (*n* = 4) in this investigation. These values are in agreement with previously published values [21] and represent a 4.3-fold stimulation by nicardipine. Both Aβ peptide isoforms were also able to elicit a stimulation in the ATPase activity of purified, reconstituted P-gp as shown in

Figure 1e. The estimated maximal degree of stimulation by Aβ<sup>40</sup> was 1.5-fold, potentially suggesting that it is a relatively weak substrate of P-gp. Aβ<sup>42</sup> produced a comparatively greater stimulation of approximately 2.2-fold, which is equivalent, or greater, than established substrates vinblastine, paclitaxel, and rhodamine 123 [19]. It was not possible to generate a more extensive range of concentrations for the peptides (see Methods Section 5.5) and consequently, the affinity cannot be reliably estimated.

Overall, the two assays provide further evidence that Aβ<sup>40</sup> and Aβ<sup>42</sup> interact with P-gp and the use of purified, reconstituted protein demonstrates a direct mechanism. In addition, the nature of the effect on ATP hydrolysis may indicate their interaction is akin to a transported substrate, for which there is considerable evidence using cellular systems (for review see [16]).
