Acetylcholinesterase from Electrophorus electricus (Type VI-S; EC 3.1.1.7) (eelAChE), acetylcholine iodide (≥97%, ACh), choline chloride (≥99%, Ch), galanthamine hydrobromide from Lycoris sp (≥94% HPLC), 9-amino-1,2,3,4-tetrahydroacridine hydrochloride hydrate (tacrine hydrochloride, ≥99%), donepezil hydrochloride (≥98% HPLC), edrophonium chloride, caffeine, ammonium acetate, Tris hydrochloride, glutaraldehyde (25% aqueous solution), and sodium phosphate monobasic were purchased from Sigma-Aldrich (St. Louis, United States). 3-(Aminopropyl)trietoxysilane (≥99%, APTES) was purchased from Acros Organics (Geel, BE). HPLC grade methanol (MeOH), acetonitrile (ACN) and ethanol (EtOH) were acquired from Merck-Millipore (Burlington, United States). Ultrapure water was obtained with a Merck-Millipore Milli-Q^® system (Burlington, United States). Fused silica capillary (0.375 mm o. d. x 0.1 mm i. d.) was acquired from Polymicro Technologies (Phoenix, United States). All the solutions were filtered through a Merck-Millipore nylon membrane filter (0.22 μm) prior to use (Burlington, United States).

Bulbs of the species Hippeastrum calyptratum (Amaryllidaceae) were collected in the city of Cunha (São Paulo, Brazil), identified by Mr. Mauro Peixoto and Dr. Jullie Dutilh (University of Campinas, Unicamp, Brazil) and a voucher deposited at the Herbarium of the Instituto plantarum (HPL 13043). An ethanolic crude extract was obtained by cold maceration until exhaustion, evaporated under reduced pressure and stored in a freezer. The sample was (10 mg ml⁻¹) resuspended in ethanol and centrifuged at 10,000 rpm for 5 min before performing the analyses. Access to genetic heritage was registered at Sistema Nacional de Gestão do Patrimônio Genético e do Conhecimento Tradicional Associado (SisGen) under code ABE20A2.

The instrument consisted of modules from the 1,290 Infinity line (Agilent Technologies, Santa Clara, United States): an autosampler (G4226A), a thermostatted column compartment (G1316C), a binary pump (G4220A, referred to as enzyme buffer pump), a high-speed pump (G7120A), a diode-array UV detector (DAD) (G4212A, with 1.0 μL flow cell, G4212-60008), a 8-port/2-position high-pressure switching valve (Agilent G4236A, Duo Valve) equipped with two identical sample loops (Agilent 5,067–5,440, calibrated loop kit for 2D-LC) and an additional pump from the Prominence series (model LC-20AD, Shimadzu, Tokyo, Japan). The system was controlled by the 2D-LC acquisition software (A.01.04 [017]) from Agilent and the Prominence pump was manually activated. The LC system was coupled to a Bruker Amazon Speed ion trap mass spectrometer (IT-MS) equipped with an electrospray source, controlled by the Bruker Compass Hystar software (version 4.5) (Bruker Daltonics Inc., Billerica, United States). Data were acquired and analyzed with the Compass DataAnalysis software (version 4.3). MS parameters were configured as follows: capillary voltage = 5000 V; end-plate voltage = 550 V, drying gas flow rate = 11.0 L min⁻¹, drying temperature = 350°C, and nebulizer pressure = 60 psi. Analyses were conducted in the manual MS(n) mode with positive ionization (m/z from 40 to 1,500) and isolation width was set at ± m/z 0.5. A syringe pump (Model 11 Plus, Harvard Apparatus, Holliston, United States) was also used for different purposes, described in the following sections.

AChE immobilization was carried out by infusing the enzyme into a fused silica capillary using a syringe pump. A capillary (1,000 mm × 0.1 mm i. d.) previously activated with APTES and modified with glutaraldehyde was used as support for enzyme immobilization. The immobilization procedure was performed according to previously reported protocols (Vanzolini et al., 2013) using a 50 U mL⁻¹ AChE solution. When not in use the reactor was maintained at 4°C, with both extremities immersed in Tris buffer 50 mM, pH 8. A negative control capillary (referred to as cIMER-blank) was prepared by following the same immobilization procedure except for addition of the AChE solution.

The AChE-cIMER catalytic activity was determined by measuring the formation of the product of ACh enzymatic hydrolysis by AChE, which is choline (detected by MS as [M + H]⁺, m/z 104.0). To this end, a simplified system configuration as illustrated in Figure 1 was used, which includes a syringe pump, a switching valve equipped with two identical sample loops, the AChE-cIMER (100 cm × 0.375 mm x 0.1 mm i. d.), the enzyme buffer pump, and the IT-MS instrument.

The syringe pump was used to fill the loops with the enzyme substrate (ACh solution prepared in 15 mM ammonium acetate solution pH 5, at 100 μl min⁻¹). The valve was regularly activated (referred to as modulation time), connecting the flow path between the enzyme buffer pump and the previously filled loop, delivering the substrate to the AChE-cIMER in ammonium acetate solution (15 mM, pH 8), followed by MS detection.

The offline AChE-cIMER inhibition assay was performed using the configuration shown in Figure 2. An additional pump was necessary to deliver the enzyme substrate (140 μM ACh solution, prepared in 15 mM ammonium acetate at pH 5) at 50 μL min⁻¹. The syringe pump was used to deliver individual solutions of acetylcholinesterase inhibitors in different concentration ranges: galantamine (0.27–27 μM); tacrine (4.2 × 10⁻³–4.2 μM) and donepezil (2.3 × 10^–3 to 2.3 μM), which were all prepared in 15 mM ammonium acetate solution at pH 8. The syringe pump flow was 50 μL min⁻¹. Substrate and inhibitor solutions were combined through a T-piece placed before the switching valve (Figure 2), which resulted in a final flow rate of 100 μL min⁻¹ reaching the sample valve. The valve was equipped with two 60 μL sample loops and switched every 36 s (modulation time = 36 s), to deliver 60 μL of the substrate + inhibitor mixture to the IMER at each modulation. In Position 1, the substrate + inhibitor solution was stored in loop 1, whereas the enzyme buffer pump (15 mM ammonium acetate pH 8.0, at 600 μL min⁻¹) conditioned the AChE-cIMER and kept the MS baseline steady. In Position 2, the content of loop 1 was directed to the AChE-cIMER for analysis, while loop 2 was filled with more substrate + inhibitor solution. Total analysis time was 10 min. A negative control experiment was performed by replacing the cIMER with the cIMER-blank.

The inhibition percentage for each inhibitor concentration was calculated using Eq. 1 and peak areas from extracted ion chromatograms for the molecular ion of protonated choline (Ch, m/z 104.0): i n h i b i t i o n % = [ 1 − A i − C i A 0 − C 0 ] x 100 (1) where A _i and C ₁ are the areas obtained in the presence of inhibitors with the AChE-cIMER and cIMER-blank, respectively, and A ₀ and C ₀ are the areas obtained in the absence of inhibitors with the AChE-cIMER and cIMER-blank, respectively. All samples were analyzed in triplicate and results are presented as means ± one standard deviation (GraphPad Prism version 5.0).

The 2D-LC-MS system was configured with the previously described modules (see Instrumentation and software section) as illustrated in Figure 3. The first dimension (¹D) included a binary pump, autosampler, thermostatted column compartment, a Gemini C18 analytical column (150 mm × 1.0 mm i. d., 5 μm, 110 Å, Phenomenex, Torrance, United States) and diode array detector. The second dimension (²D) consisted of the enzyme buffer pump, the Prominence pump (added to the system for exclusive delivery of enzyme substrate solution), and the immobilized acetylcholinesterase enzyme reactor (1,000 mm × 0.1 mm i. d.), which was finally coupled to the ion trap mass spectrometer instrument (IT-MS). Both dimensions were interfaced with an 8-port/2-position high-pressure switching valve equipped with two 60-µL sampling loops.

Samples were analyzed using the online 2D-LC-MS platform for AChE-cIMER inhibition assay according to the following procedure:

When the valve was in position 1, sample was injected (10 μL) into a Gemini C18 analytical column (150 mm × 1.0 mm i. d., 5 μm, 110 Å, Phenomenex, Torrance, United States) held at 40°C. The binary mobile phase consisted of (A) 20 mM ammonium acetate at pH 6 and (B) methanol. A solvent gradient was employed from 5-100-100% B in 0-10-12 min for the known inhibitors and 0-40-42 min for the H. calyptratum crude extract with equilibration time of 5 min (post-run time). The flow rate was 100 μL min⁻¹. The RP-LC effluent was split 1:1 (1 part to waste and 1 part to the switching valve) and combined with the substrate solution (ACh, 140 μM in 15 mM ammonium acetate at pH 5) at 50 μL min⁻¹ before the switching valve by means of a T-piece (total flow of 100 μL min⁻¹ reaching the valve).

While the first fraction combined with the enzyme substrate fills the first sample loop (60 μL) the ²D pump (enzyme buffer pump) delivering the enzyme buffer (15 mM ammonium acetate at pH 8) at 600 μL min⁻¹ runs through the second loop (60 μL) keeping the AChE-cIMER conditioned and MS base line steady. After 36 s (modulation time) the valve switches position and, when in position 2, the content of the first loop (fraction + substrate solution) is flushed out by the enzyme buffer pump (15 mM ammonium acetate at pH 8) at 600 μL min⁻¹ towards the AChE-cIMER for MS analysis while the second loop is filled with the next following fraction eluting from the analytical column + substrate solution. Total analysis time was 10 min.

Acetylcholinesterase activity was investigated after the immobilization procedure using the system configuration illustrated in Figure 1. The immobilization procedure did not affect the enzymatic activity and the amount of enzyme used in the immobilization procedure (50 U mL⁻¹) proved to be enough to deliver a reliable, reproducible, and detectable signal for the protonated molecular ion of choline (m/z 104.0), the product of the enzymatic reaction (Figure 4). It was possible to monitor AChE-cIMER activity even with substrate concentration as low as 6.25 × 10⁻⁵ μM (Supplementary Figure S1A). Spontaneous hydrolysis of the substrate was not detectable—only noise signal corresponding to the Ch product ion (m/z 104.0) was detected when the substrate solution was infused through the cIMER-blank (Supplementary Figure S2). As expected, the presence of ACN and EtOH at concentrations as low as 20% caused the enzyme to completely lose its catalytic activity (data not shown), while MeOH was well tolerated at concentrations up to 50%. However, at 50% MeOH a small loss of enzymatic activity was observed as a decrease in signal intensity for the choline ion. Nevertheless, the enzymatic activity was recovered and remained reproducible with each new injection. The enzymatic reaction proved to be fast and reproducible, even with modulation times as low as 0.3 min (18 s) for all the ²D flow rates tested. However, on the basis of peak shape, the following were chosen as final conditions: modulation time of 36 s, a 60 μL sample loop, and ²D flow rate of 600 μl min⁻¹.

The AChE-cIMER stability was evaluated over time according to its ability to maintain its catalytic activity by monitoring the choline signal corresponding to the enzymatic reaction product (Ch, m/z 104.0) using the configuration described in Figure 3 and the optimized final conditions. After 3 months of daily use, the Ch signal did not decrease, remaining visible with an intensity above x10⁶ and was still active 2 years after the immobilization.

The offline AChE-cIMER inhibition assay was performed using the system configuration described in Figure 2. Briefly, solutions containing standard AChE inhibitors tacrine, donepezil and galanthamine at least five different concentrations were infused at 50 μL min⁻¹ and combined with the substrate solution that was being pumped at 50 μL min⁻¹. The final solution containing inhibitor + substrate was transferred to the AChE-cIMER every 36 s while the enzyme buffer pump flow rate was set at 600 μl min⁻¹.

Enzyme inhibition was visible at inhibitor concentrations as low as 1.35 μM for galanthamine (20.3% inhibition), 1.15 μM for donepezil (32.9% inhibition), and 0.21 μM for tacrine (33.4% inhibition). Meanwhile, inhibitor solutions with 10-fold higher concentrations resulted in 55.3, 64.3 and 66.6% enzyme inhibition for galanthamine, tacrine, and donepezil, respectively (Figure 5). The concentration of each inhibitor that caused an enzyme inhibition above 50% was used as a reference to prepare individual inhibitor solutions and a solution containing a mixture of the three aforementioned inhibitors for the online 2D-LC-MS platform for AChE-cIMER inhibition assay.

First, we evaluated the ability of the online 2D-LC-MS platform for AChE-cIMER inhibition screening assay to individually detect the inhibition potential for the same known inhibitors, galanthamine, tacrine, and donepezil at 108, 8.4 and 6 μM (all in 15 mM ammonium acetate at pH 8), respectively. Each inhibitor solution was injected (10 μL) into the analytical column and then eluted using the chromatographic conditions described in the experimental section. Half of the eluate was combined with the enzyme substrate solution (140 μM in 15 mM ammonium acetate pH 5) and half went to the trash. Every 36 s, a 60-μL aliquot was transferred to the second dimension toward the AChE-cIMER, and the enzyme inhibition was measured.

Figure 6 shows the resulting extracted ion chromatogram (EIC) for the protonated molecular ion choline (Ch, m/z 104.0) after the injection of individual solutions of galanthamine (Figure 6 panel 2A), tacrine (Figure 6 panel 2B) and donepezil (Figure 6 panel 2C). It is possible to observe that the choline signal intensity (Panels 2A, 2B and 2C) decreased when compared to the control (analysis performed in the absence of inhibitor; Panels 1A, 1B and 1C) and this decrease corresponds to the exact modulation where each inhibitor was present. In previous works the enzymatic activity was strongly impaired in the presence of the organic modifier, although the MeOH concentration was kept as low as 25% (Ingkaninan et al., 2000b; Yuan et al., 2020). Here, the enzymatic activity was only slightly affected when concentration of MeOH in the ¹D effluent reached 100% (Figure 6 Panel 1A-D), and the enzyme activity was recovered prior to the next injection during flushing with 15 mM ammonium acetate solution at pH 8.

The signal intensity of the enzymatic reaction product decreased (m/z 104.0) (Panels 2A, 2B and 2C) when compared to the control (obtained in the absence of inhibitor; Panels 1A, 1B and 1C) at the exact retention time corresponding to the modulations where the inhibitors were present (Panels 3A, 3B and 3C). The peak corresponding to galanthamine when transferred from the analytical column to the AChE-cIMER was divided into two modulations. Panel 3A shows the EIC corresponding to galanthamine (m/z 288.0) at about 6 min. In panel 2A, it is possible to observe a decrease in the signal intensity of the Ch (m/z 104.0) when compared to the control (Panel 1A) around the same retention time matching the modulations where galanthamine was present, thus proving the occurrence of enzymatic inhibition. The same reasoning can be used to analyze the results for tacrine and donepezil. The EIC corresponding to tacrine (m/z 199.9) and donepezil (m/z 380.0) can be observed close to 7.5 and 10 min, respectively and modulations corresponding to these retention times show a decrease in Ch (m/z 104.0) signal intensity. Caffeine was used to assess the assay’s specificity for true AChE inhibitors. As shown in panel 2D, caffeine (m/z 195.0) did not inhibit AChE even when tested at a high concentration of 400 µM. These results confirmed the sensitivity and specificity of the 2D-LC-MS online platform for the AChE-cIMER inhibition screening assay.

Next, we evaluated whether the inhibition observed for the known inhibitors in the previous experiment (Figure 6) was a result of true inhibition or just a signal suppression effect. As shown in Figure 7, we observed no effect on the signal corresponding to the Ch ion (m/z 104.0) in the absence (Figure 7, Panels 1A–1C) and presence (Figure 7, Panels 2A–2C) of known inhibitors, even when tested at concentrations much higher than those used for the inhibition assay shown in Figure 6. This confirms that the observed decreases in peak intensity for the Ch ion (m/z 104.0) (Figure 6) were actually due to interaction between each of the known inhibitors and the AChE immobilized inside the enzyme reactor and therefore true enzyme inhibition.

A mixture containing all the three inhibitors at the same concentrations tested individually was prepared and injected it into the complete 2D-LC-MS online platform for the AChE-cIMER inhibition screening assay. RP-LC retention times for galanthamine, tacrine, and donepezil were 4.3, 5.9 and 8.5 min, respectively (Figure 8A). The enzyme inhibition profile resulting from the analysis of the inhibitor mixture was similar to the inhibition profile obtained for each inhibitor when analyzed individually. To analyze the data shown in Figure 8B, we compared the EIC signal obtained for the Ch ion (m/z 104.0) in the absence (Panel 1) and presence (Panel 2) of the known inhibitor mixture. Panel 2 shows a decrease in signal intensity corresponding to the enzymatic reaction product (m/z 104.0) matching the modulations where the known inhibitors were present. Galanthamine was the first inhibitor to elute and once again its peak was divided (not equally) between two modulations (Figure 8B panel 5). Therefore, it is possible to confirm a decrease in the intensity of the Ch signal close to 5.5 and 6 min corresponding to the modulations where galanthamine was present (Figure 8B Panel 2). Following the same reasoning, tacrine (Figure 8B Panel 4) and donepezil (Figure 8B Panel 3) were the second and third compounds to elute and their respective inhibitory effects could be verified by a decrease in the intensity of the signal corresponding to the product of the enzymatic reaction around 7.5 and 10 min coinciding with the modulations where they were present (Figure 8B Panel 2).

Lastly, an ethanolic extract solution (10 mg ml⁻¹) obtained from bulbs of H. calyptratum was used to verify the applicability of the assay using the complete 2D-LC-MS online platform (Figure 3). Figure 9A show the chromatographic profile after injection of the H. calyptratum extract while the results for the AChE inhibitor screening assay are shown in Figure 9B. Experimental details are given in Section 2.7. Figure 9B shows the EIC signal corresponding to the product of the enzyme reaction m/z 104.0 in the absence of any inhibitor (Panel 1), after the injection of the inhibitor galanthamine (GL) solution (108 μM) (Panel 2) and after the injection of the H. calyptratum ethanolic extract (Panel 4). Panel 3 and 5 corresponds to the EICs (m/z 288.0) for galanthamine when injected alone and present within the H. calyptratum ethanolic extract, respectively. It is possible to observe 3 regions (Figure 9, panel 4) where the extract showed a possible anticholinesterase activity. The first, with less intense inhibition corresponding to modulations between 5 and 10 min, the second region with a more intense and visible inhibition between 11 and 14 min and, a third region with inhibition of intermediate intensity visible between 16 and 18 min. The first region coincides with modulations and m/z corresponding to galanthamine used as a control (m/z 288.0), while the second and third may indicate one or a set of compounds with greater anticholinesterase activity within the extract. More in-depth studies should be carried out to try to identify those responsible for this activity present in the extract and are not part of the scope of this work.