Development of tacrine clusters as positively cooperative systems for the inhibition of acetylcholinesterase

Abstract The synthesis of four tetra-tacrine clusters where the tacrine binding units are attached to a central scaffold via linkers of variable lengths is described. The multivalent inhibition potencies for the tacrine clusters were investigated for the inhibition of acetylcholinesterase. Two of the tacrine clusters displayed a small but significant multivalent inhibition potency in which the binding affinity of each of the tacrine binding units increased up to 3.2 times when they are connected to the central scaffold.


Introduction
Multivalent interactions (or multivalency) constitute a widespread recognition phenomenon in living organisms to establish interactions between carbohydrates and proteins, which are essential for the adhesion of viruses and bacteria to the surface of a cell in addition to cell adhesion 1 . The power of multivalent interactions is that when several binding modules are connected to a central scaffold and bind cooperatively to a target, the binding affinity of the multivalent ligand on a valency-corrected basis (rp/n) can be dramatically increased (i.e. the binding affinity of the multivalent ligands is stronger than the sum of its mono-valent ligands alone), which is known as the cluster effect or multivalent effect 2 . A wellresearched field in bioorganic chemistry is the synthesis of multivalent glycoconjugates for investigation of the multivalent effect for carbohydrate-protein (lectin) interactions 2b, 3 . On a valency-corrected basis, such multivalent assemblies of carbohydrates have achieved an affinity enhancement of an astonishing six orders of magnitude for the binding to lectins 4 . A much less explored field is multivalent enzyme inhibition, which has been associated with the fact that most enzymes possess a single deep active site that is expected to be less accessible for multimeric ligands than several binding pockets on the surface of lectins 5 . In fact, such pockets on the surface of lectins give rise to efficient chelating binding with multivalent glycoconjugates 6 , and therefore multivalent effect for enzyme inhibition has been disregarded 5b . To the best of our knowledge, if we neglect bivalent enzyme inhibitors, multivalent enzyme inhibition potency has only been achieved for a few groups of enzymes including, glycosidases 2a,3a,7 , glycosyltransferases 8 , carbonic anhydrases 9 , and very recently for cholinesterases 10 . In this context, it is worth mentioning that a 36 valent inhibitor has been demonstrated to give rise to an astonishing affinity enhancement of ca 4700-fold on a valency-corrected basis for the inhibition of a-mannosidase 2a,11 , which emphasise the power of multivalent enzyme inhibition. However, there is no general linear correlation between valency and enzyme inhibition potency on a valency-corrected basis as observations have been made in which the inhibition on valency-corrected basis decrease by valency 9a,12 . Another parameter to consider in the design of efficient multivalent inhibitors is the choice of the scaffold where various types of scaffolds implement different spatial orientations of the inhitopes, which can affect the inhibition 8,13 . The length of the linkers connecting the central scaffold with its inhitopes has also been identified as an important parameter for efficient multivalent enzyme inhibition 14 .
Alzheimer's disease (AD) is a multifactorial progressive neurological disorder that represents the most common form of dementia 15 . Currently, there is no cure available for this devastating disease due to a lack of exact knowledge of its causes 16 . The cholinergic hypothesis suggests that the level of the neurotransmitter acetylcholine (ACh) is insufficient in the Alzheimer brain, which causes cognitive loss 17 . Therefore, inhibition of cholinesterases [acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE)] and thereby increasing the concentration of ACh in the brain is an attractive target for the treatment of AD 18 . One example of a cholinesterase inhibitor drug for palliative treatment of AD is tacrine (1) (Figure 1), which unfortunately was discontinued in 2013 as it results in liver damage 19 . When the structure of AChE was solved by X-ray crystallography 20 , an active gorge, lined with aromatic residues, penetrating ca 20 Å into the enzyme was recognised hosting: (1) the active site including the catalytic triad and catalytic anionic binding subsite (CAS) nearby the bottom of the active gorge and (2) the peripheral anionic binding site (PAS) located at the interface of the active gorge.
When the structure of tacrine complexed with AChE was solved by X-ray crystallography, it was concluded that it binds to CAS in the solid state 21 . To establish interactions with both CAS and PAS, a bivalent strategy was pursued in which two tacrine rings were connected via a heptamethylene linker to obtain bis(7)tacrine (2) (Figure 1) that is a ca 1000-fold stronger AChE inhibitor than tacrine, which was associated with simultaneous interactions with CAS and PAS 22 . The finding of the enhanced AChE inhibition by bis(7)tacrine (2) triggered an avalanche of reported bivalent AChE inhibitors 18,23 .
The tetrameric structure of AChE that contains four catalytic subunits 24 led us to propose tetra-tacrines 3a-6a ( Figure 1) as multivalent AChE inhibitors. We argued that when the tacrine rings are attached to the central sugar scaffold via linkers of optimal length they would employ the chelation effect 5b to bind simultaneously to the active gorges of the AChE tetramer to form a stable AChE: tetra-tacrine complex. An alternative mechanism to achieve multivalent inhibition by tetra-tacrines 3a-6a is due to a statistical binding effect 5b caused by the increased effective concentration of the tacrine rings nearby the active gorges of AChE. Thus, in this paper we present: (1) the synthesis of the tetratacrines 3a-6a (Figure 1), (2) the synthesis of the mono-tacrines 3b-6b, and (3) the multivalent inhibition potencies of tetratacrines 3a-6a against AChE by comparing them with reference compounds 3b-6b.

Materials and methods
General procedures DMF has dried over 4 Å molecular sieves (oven-dried). All reactions were carried out under an argon atmosphere unless otherwise specified. Microwave reactions were performed in a CEM Discover-SP, max power 300 W. TLC analyses were performed on Merck silica gel 60 F254 plates or Sigma-Aldrich aluminium oxide 60 F254 (neutral) plates using a UV light for detection. Silica gel NORMASIL 60 V R 40-63 mm or Aluminium oxide Sigma-Aldrich 58 Å pore size was used for flash column chromatography. NMR spectra were recorded on a Bruker Avance NMR spectrometer; 1 H NMR spectra were recorded at 400.13 or 850.13 MHz, 13 13 C NMR). High-resolution mass spectra (HRMS) were recorded from on a Qexactive spectrometer in positive electrospray ionisation (ESI) mode.

Synthetic protocols
General procedure for the preparation of compounds 3b-6b A mixture of propargyl alcohol (7) (2.4 mmol, 7 equiv.), azide 8, 9, 10, or 11 (0.2 mmol, 1 equiv.), and copper (II) sulphate pentahydrate (0.3 equiv.) in DMF (3 ml) in a foil-covered round bottom flask was added sodium ascorbate (0.6 equiv.). The mixture was kept stirring at room temperature overnight under Ar atmosphere. The solvent was then removed under reduced pressure and the concentrate was purified by silica gel flash column chromatography.
General procedure for the preparation of compounds 3a-6a A mixture of the alkyne 13 (0.2 mmol, 69.3 mg, 1 equiv.), azide 8, 9, 10, or 11 (4.8 mmol, 1.2 equiv. per reactive group of the alkyne), and copper (II) sulphate pentahydrate (0.3 equiv. per reactive group of the alkyne) in DMF (5 ml) was added sodium ascorbate (0.6 equiv. per reactive group of the alkyne). The mixture was irradiated in a microwave at 300 W and 115 C for 45 min. Water (10 ml) was added and the crude mixture was extracted with dichloromethane (3 Â 20 ml). The organic phases were combined, dried with MgSO 4 , and filtered. Evaporation of the solvent by reduced pressure yielded a crude material that was purified by column chromatography.

Cholinesterase assays
For the assessment of enzymatic inhibition, commercially available acetylcholinesterase from Electrophorus electricus (type V-S, Sigma Aldrich) was used, conducting minor modifications on Ellman's protocol 25 . Stock solutions of inhibitors were prepared in DMSO, being the solvent content of 1.25% (V/V) in the final assay solutions. Enzymatic activities were measured in a UV-Vis instrument (Hitachi U-2900) using PS cuvettes containing 0.1 mM phosphate buffer (pH 8.0), 5,5 0 -dithiobis(2-nitrobenzoic acid) (DTNB, 0.88 mM, buffer solution), acetylthiocholine iodide as a model substrate, inhibitor, properly diluted aqueous enzyme solution, and water up to 1.2 ml. Solutions of the enzymes were prepared so as to keep the reaction rate within 0.12-0.15 Abs/min when [S] ¼ 4 Â K M . The formation of the chromophore was monitored during 125 s at 405 nm and 25 C.
Calculation of IC 50 values was accomplished by plotting %I vs. log[I] and adjusting to a second-order equation. Substrate concentration was kept at 121 mM, using 2-4 independent assays, each of them, being run in duplicate.
For the calculation of the kinetic parameters of the free enzyme, and in the presence of 3a, five different substrate concentrations, ranging from 1 =4 K M to 4 Â K M were used. Cornish-Bowden method 26 provided the mode of inhibition of 3a; for that purpose, two different plots were used: 1/v vs.
[I]. Mixed inhibition was found for such compound, which means that it binds both, the free enzyme (K ia ) and the enzymesubstrate complex (K ib ). Kinetic parameters (K M , V max , K M app, V max app ) were obtained through non-linear regression analysis (least squares fit) using the GraphPad Prism 8.01 software and inhibition constants were calculated using the following equations: Data are expressed as the mean ± SD.

Synthesis
The presence of 1,2,3-triazole moieties in the linker between the pharmacophores in bivalent cholinesterase inhibitors has been found to establish interactions with residues in AChE 27 . Therefore, we considered it unsuitable to employ tacrine (1) as a reference compound for the evaluation of the multivalent inhibition potency of 3a-6a. Instead, for each tetra-tacrine 3a, 4a, 5a, and 6a a mono-tacrine reference compound 3b, 4b, 5b, and 6b, respectively, was prepared to contain the 1,2,3-triazole moiety and the same number of CH 2 -groups between the tacrine ring and the hydroxyl group as the corresponding tetra-tacrine contains CH 2groups between its tacrine rings and central scaffold. Reference compounds 3b-6b were obtained when propargyl alcohol (7) underwent Cu(I) catalysed alkyne-azide 1,3-dipolar cycloaddition (CuAAc) with azide armed tacrine derivatives 8 28 and 9-11 23a (Scheme 1). The synthesis of the tetravalent tacrine architectures 3a-6a commenced from commercially available methyl a-D-glucopyranoside (12), which was subjected to propargylation upon treatment with propargyl bromide and sodium hydride to provide 13 (Scheme 2). In the final step, tetra-alkyne 13 was armed with four tacrine inhitopes when it underwent Cu(I) catalysed alkyne-azide 1,3-dipolar cycloaddition with azides 8, 9, 10, and 11 to obtain tetramers 3a, 4a, 5a, and 6a, respectively, with variable length of the linker between the central sugar scaffold and their inhitopes. The formation of 1,4-regiosisomeric triazole moieties in 3a-6a was supported with 13 C-NMR spectroscopy where the carbon atoms in 5-position in the triazole moieties consistently appeared in the range 124.8 to 122.7 ppm, which agrees with reported data for such isomers 29 . The carbons in 5-position (CH-triazole) were in turn identified through HMBC correlation with the CH 2 -protons (2 0 -H, 3 0 -H, 4 0 -H, and 6 0 -H) between the triazole moieties and the central sugar scaffold (Figure 2).

Inhibition studies
The potency of tetra-tacrines 3a-6a and mono-tacrines 3b-6b for the inhibition of Electrophorus electricus AChE were tested using the Ellman method 25 and the activities are presented in Table 1. All the tacrine-monomers 3b-6b displayed potency in the nM concentration range from IC 50 ¼ 566 nM down to IC 50 ¼ 7.1 nM for the inhibition of AChE. The mono-tacrines with longer linkers [5b (m ¼ 6) and 6b (m ¼ 8)] between the tacrine and triazole rings are significantly stronger inhibitors than 3b (m ¼ 2) and 4b (m ¼ 3) with shorter linkers, which indicates that the longer ligands establish more efficient simultaneous interactions with PAS and CAS in the active gorge. The tetra-tacrines 3a-6a also displayed potency in nM concentration range (IC 50 ¼ 12.5 nM to IC 50 ¼ 232 nM). However, for these tetra-valent inhibitors, there was no clear trend between the linker length between the tacrine and triazole rings as the strongest tetra-tacrine AChE inhibitor 5a (IC 50 ¼ 12.5 nM, m ¼ 6) behaves as a 36-fold stronger inhibitor than the weakest tetra-tacrine inhibitor 6a (IC 50 ¼ 232 nM, m ¼ 8).
The relative inhibition potency (rp) was obtained by dividing the IC 50 value of the mono-tacrine with the IC 50 value of the corresponding tetra-tacrine, which contains the same number of CH 2 -groups between the tacrine and triazole rings [for instance, rp ¼ IC 50 (3b)/ IC 50 (3a) ¼ 12.9]. The relative inhibition potencies for tetra-tacrines 3a-6a demonstrates that longer linkers between the triazole and tacrine rings have a destructive impact on the inhibition potency, as the rp-values gradually decrease from rp ¼ 12.9 for 3a (m ¼ 2) to rp ¼ 0.03 for 6a (m ¼ 8). The inhibition potencies on valency-corrected basis (rp/n) showed that tetra-tacrines 3a (rp/n ¼ 3.2, m ¼ 2) and 4a (rp/n ¼ 1.5, m ¼ 3) exhibit small but significant multivalent inhibition potencies for AChE. The rp/n-values for 5a (rp/n ¼ 0.25, m ¼ 6) and  6a (rp/n ¼ 0.008, m ¼ 8) on the other hand demonstrate that the mono-tacrines 5b (m ¼ 6) and 6b (m ¼ 8) were 75% and more than 99% less active, respectively, when they are connected to the central multivalent sugar scaffold. From a Cornish-Bowden plot ( Figure 3) for 3a, we concluded that it causes a mixed inhibition mode of AChE [K ia ¼ 31.6 ± 2.0 nM (competitive inhibition constant) and K ib ¼ 45.0 ± 5.9 nM (non-competitive inhibition constant)], which implies that it binds to the catalytic site in addition to a second binding site, for instance, PAS on the entrance of the active gorge. Thus, the multivalent inhibitory potency observed for 3a and 3b might be due to the chelation effect in which the length of the linkers in 3a and 4a are of sufficient length to allow simultaneous binding of their tacrine inhitopes to more than one active gorge in the tetrameric AChE enzyme. On the other hand, shorter linkers in the tetra-tacrines imply higher effective concentration nearby the active gorges, and thus a statistical binding effect cannot be excluded as the reason for the observed multivalent inhibition potency observed for tetra-tacrines 3a and 4a. However, as rp/n ˂ 1 for 5a and 6a, in such statistical binding effect scenario, it implies that another effect is involved, which oppose the binding of the inhitopes to the enzyme for example that longer linkers affect the position of the tacrine rings in such a way that they become less accessible for the enzyme.

Conclusions
We have applied the Cu(I)-catalysed azide-alkyne Huisgen cycloaddition reaction to obtain four tetra-tacrine clusters 3a-6a in which the tacrine rings are connected to a central scaffold via linkers of variable lengths. Two of the tetra-tacrines 3a and 4a with the shortest linkers displayed a small but significant multivalent effect in the inhibition of AChE. The observed multivalent inhibition potency is proposed to arise from the chelation or statistical binding effects.