Results and discussion

Synthesis of compounds

Ugi-4CR was used to prepare a series of 11 novel peptide-like compounds, which all have an α-acylaminoacetamide skeleton but differ in the acyl part (R1: phenyl and methyl groups), substituents at the acylamino nitrogen atom (R2: benzyl and iso-butyl groups), substituent at α-acetamide carbon atom (R3: methyl group, compound 11), and α-acetamide nitrogen atom (R4: morpholinylethyl, tert-butyl, cyclohexyl, and p-toluensulfonylmethyl groups). All compounds were synthesised by both conventional and microwave-assisted synthesis (Figure 2).

Figure 2.: Synthesis of peptidomimetics 1 - 11 and their structure (the amine, aldehyde, isocyanide, and acid components are depicted in blue, green, magenta, and red, respectively)

In the initial approach, a conventional Ugi-4CR was employed to synthesize the compounds under standard conditions reported in the literature, in which the formaldehyde, amine, carboxylic acid, and isocyanide components were mixed in methanol at room temperature for 24 h [43]. However, as conventional Ugi-4CR for some of the desired compounds yielded complex mixtures with low yields, we turned to microwave-assisted synthesis. First, reaction conditions for the microwave synthesis of compound 5 were optimized by varying solvent, temperature, and reaction time (Table 1).

Entry | Solvent | Temperature, °C | Reaction time, min | Yield, %
1 | MeOH | 120 | 10 | 71
2 | EtOH | 120 | 10 | 61
3 | i-PrOH | 120 | 10 | 40
4 | MeOH | 130 | 10 | 58
5 | MeOH | 140 | 10 | 39
6 | MeOH | 120 | 5 | 49
7 | MeOH | 120 | 15 | 60
8 | MeOH | 110 | 15 | 83

Initially, solvent screening revealed that methanol was optimal for the reaction (120 °C, 10 minutes, entries 1 to 3). Increasing the reaction temperature to 130 or 140 °C resulted in yields of 58 and 39 %, respectively, compared to 120 °C (entries 4 and 5). A shorter reaction time of 5 minutes at 120 °C also lowered the yield to 49 % (entry 6), as well as a longer reaction time of 15 minutes (entry 7). Further optimization with methanol as the solvent and a reaction time of 15 minutes established 110 °C as the optimal temperature, yielding an 83 % yield (entry 8).

Compounds 1 to 5 were synthesized using the established reaction conditions. Compound 9 was prepared analogously, substituting acetic acid for benzoic acid. To minimize side-product formation during the synthesis of compounds 6 to 8, the reaction temperature was increased to 130 °C and the reaction time shortened to 4 minutes. Compound 10, synthesized from acetic acid, benzylamine, formaldehyde, and tert-butyl isocyanide, required a reaction temperature of 130 °C and a reaction time of 10 minutes.

For the synthesis of compound 11, utilizing acetone as the carbonyl component, several reaction conditions were investigated. As ketones typically require pre-condensation with amines to form an imine before the addition of carboxylic acid and isocyanide in Ugi-4CR reactions [44], imine formation from acetone and benzylamine via thin-layer chromatography within 5 minutes was confirmed. Microwave irradiation at 140 °C for 15 minutes was initially investigated using both methanol and acetonitrile as solvents; methanol proved superior. Lowering the temperature to 130 or 120 °C, with a corresponding increase in reaction time to 25 minutes, resulted in decreased yields of 29 and 10 %, respectively.

The structures of all prepared compounds were deduced and confirmed by 1D and 2D 1H and 13C NMR experiments and HRMS. Because these compounds contain amide bonds around which slow rotation is possible, they can exist as a mixture of cis- and trans-conformers [45].

Inhibition of cholinesterases

All synthesized compounds were evaluated for their ability to inhibit both human acetylcholinesterase (hAChE) and human butyrylcholinesterase (hBChE). Reversible inhibition was observed for all compounds against both enzymes. Initial screening for anticholinesterase activity was performed using commercially available horse butyrylcholinesterase (eqBChE) to determine IC50 values, summarized in Table 2.

Compound | eqBChE | hAChE | hBChE | Ki(AChE)/ Ki(BChE)
IC50, μM | Ki / μM | KS / mMa | Ki / μM | KS / mMa
1 | 10.0 ± 0.6 | 151 ± 8 | 0.31 ± 0.03 (m) | 5.6 ± 0.3 | 0.59 ± 0.07 (m) | 25.2
2 | 95.7 ± 1.7 | 210 ± 13 | 0.34 ± 0.03 (m) | 59 ± 3 | 0.46 ± 0.05 (m) | 3.6
3 | 39.3 ± 4.3 | 185 ± 13 | 0.46 ± 0.06 (m) | 21 ± 1 | 1.1 ± 0.13 (m) | 8.8
4 | 63.0 ± 2.6 | 196 ± 15 | 0.27 ± 0.03 (c) | 49 ± 2 | 1.2 ± 0.21 (m) | 4.0
5 | 115.9 ± 8.5 | 559 ± 21 | 0.94 ± 0.13 (m) | 1.9 ± 0.1 | 0.65 ± 0.06 (c) | 279
6 | 425.4 ± 27.8 | 434 ± 20 | 0.41 ± 0.03 (c) | 70 ± 4 | 0.33 ± 0.03 (m) | 6.2
7 | 224.0 ± 3.3 | 334 ± 18 | 0.29 ± 0.02 (c) | 55 ± 2 | 0.46 ± 0.04 (c) | 6.0
8 | 45.0 ± 1.8 | 338 ± 12 | 0.38 ± 0.02 (c) | 2.1 ± 0.1 | 0.87 ± 0.13 (m) | 169
9 | 269.8 ± 0.1 | 206 ± 12 | 0.36 ± 0.03 (m) | 47 ± 2 | 0.44 ± 0.04 (c) | 4.4
10 | 104.7 ± 0.7 | 199 ± 8 | 0.43 ± 0.04 (m) | 20 ± 1 | 1.3 ± 0.26 (c) | 9.9
11 | 445.7 ± 9.3 | 138 ± 10 | 0.26 ± 0.03 (m) | 51 ±2 | 0.56 ± 0.05 (c) | 2.7
Galantamine [47] |  | 0.52 ± 003 |  | 1.08 ± 0.08 |  | 0.5
Donepezil [47] |  | 0.024 ± 0.007 |  | 2.33 ± 0.73 |  | 0.01

Despite differences in potency and selectivity, the fundamental inhibitor mechanisms for eqBChE, hBChE, and hAChE are generally conserved. Commercially available equine BChE was used as an enzyme of choice for initial inhibitor screening due to its high homology in primary structure with human BChE [46]. The primary structure of equine and human BChE differ in 15 amino acids, and comparison of the crystal structure of human BChE and modelled structure of equine BChE suggests that only Thr69, located close to the peripheral site of equine BChE and 150 nm far from the acetylation site, could affect the different inhibitory activity determined for carbamates [46]. Therefore, eqBChE was used as a useful, preliminary model to efficiently flag compounds worthy of further investigation against both human cholinesterases.

The compounds demonstrated varying potency against equine (eqBChE), with inhibition ranging from 10.0 μM (compound 1) to 445.7 μM (compound 11), with compound 1 being the most potent, followed by compounds 8 (45.0 μM) and 3 (39.3 μM).

Following confirmation that all compounds inhibited eqBChE in the micromolar range (Table 2), their ability to inhibit human BChE and human AChE was assessed by determining enzyme-inhibitor dissociation constants, _K_i (Table 2).

All compounds reversibly inhibited both hBChE and hAChE, with _K_i values ranging from 1.9 to 70 μM for hBChE and 138 to 559 μM for hAChE. Compounds 5 and 8 were the most potent hBChE inhibitors and compounds 1 and 11 (_K_i = 138 μM) hAChE inhibitors. Benzoic acid, as an acid component in the Ugi reaction, generally provided higher inhibition potency toward hBChE than acetic acid derivatives 9 to 11. Notably, benzoic acid derivative 1 exhibited four to eight times higher affinity for hBChE than compounds 2, 3, and 4, which contained meta-substituted bromo, chloro, or nitro groups. Electron-withdrawing groups on the benzene ring (Br, Cl, NO2) of the benzoic acid generally increased the BChE _K_i values, indicating that unfavourable electrostatic interactions were introduced besides the expected steric factors.

At the same time, acetic acid derivative 9 showed comparable hBChE inhibition to compounds 2 and 4 with _meta-_substituted bromo and nitro benzoic acid. As expected, acetic derivative 11 proved to be the best inhibitor for hAChE (isostructural to ACh). The switch from formaldehyde to acetone (compound 10 to compound 11) increased the _K_i values by 2.6-fold for hBChE, but it is reversed for hAChE. This suggests that the acetone's methyl groups introduced steric bulk that is unfavourable in hBChE but had favourable hydrophobic interactions within the hAChE active site. Isocyanide and amine components in compounds that were the best inhibitors of both enzymes were different and had a lesser amount of influence on inhibition potency.

Overall, the compounds exhibited higher affinity for hBChE, with selectivity ratios (_K_i(AChE)/_K_i(BChE)) ranging from 2.7 to 279. Benzoic acid derivative 5 displayed the strongest selectivity for hBChE (ratio of 279), followed by compound 8 (ratio of 169), suggesting that these two compounds are promising candidates for further development of hBChE-selective inhibitors.

Generally, compounds from the series bind to both enzymes, occupy the catalytic centre of both enzymes and compete for the binding with substrate ATCh, and more often form additional interactions with amino acids of the peripheric aromatic site, demonstrating the mixed type of inhibition. Inhibition potency of compounds 5 and 8 towards BChE corresponded to that of currently approved Alzheimer’s disease drugs galantamine and donepezil, pointing to their potential use as BChE selective anti-AD drug.

Principal component analysis

To better understand the selectivity and joint activities of the investigated compounds, we performed data-augmented PCA of the inhibition data. PCA can reveal distinct selectivity patterns among chemical compounds by mapping their interactions with two enzymes into a lower-dimensional space, thereby separating compounds that preferentially bind one enzyme from those that bind the other. Joint activities where a compound affects both enzymes will cluster together in the PCA space, indicating a possible similar mechanism or binding profile. By visualizing compound distributions in a rotated, reduced space spanned by augmented data, PCA highlights which compounds are selective for a single enzyme and which exhibit activity towards both enzymes. Augmenting the data with the obtained min/max values for each enzyme, we created a more robust representation with clear physical meaning, where the first principal component (PC1) aligns with a predefined "maximum" direction. This is an advantage for understanding the data because the reduced space spanned was visualized based on individual enzymatic activities. Classification based on analysis performed for hAChE and hBChE data from Table 2 is presented in Figure 3. Selectivity/joint inhibition reduced space was spanned by 4 distinct points that define the selectivity directions and the joint inhibition direction. The direction from the green/red circle (good for hAChE and bad for hBChE) on the left to the red/green circle (bad for hAChE and good for hBChE) on the right along the 1st principal component represents a path in the rotated reduced space where selectivity changes from hAChE to the selectivity toward hBChE. Direction from red/red circle (bad for hAChE and bad for hBChE) on the top to green/green circle (good for hAChE and good for hBChE) on the bottom in a negative direction of the 2nd principal component represents a path in the reduced space from overall bad joint activities to good joint activities. The augmented points are not perfectly aligned perpendicularly (or horizontally) because the principal components are influenced by all points, and their directions represent statistical weights and variance across them. These data can be used in the future to establish inhibition/property regression models.

Figure 3.: Factor scores of the reversible inhibition data for compounds 1 to 11 spanned by the first two principal components. (Red/red circle: bad for hAChE and bad for hBChE, green/green circle: good for hAChE and good for hBChE, green/red circle: good for hAChE and bad for hBChE, red/green circle: bad for hAChE and good for hBChE). Molecular charge distribution for compounds 1, 3, 10, 7 and 6 visualized by electrostatic potential mapped on the total electronic density calculated at B3LYP-D3BJ/6-31G(d) level of the theory (isovalue for density: 0.01)

Among the investigated compounds, compound 5 shows the best selective inhibition of hBChE, and, based on the value of the 1st principal component score, compound 11 shows the worst selective inhibition of hBChE (but the best selective inhibition of hAChE). Score values from the 1st principal component are directly connected to the selectivity of investigated compounds, which is shown by green/red and red/green circles. In the analysis of joint inhibition activities, based on the 2nd principal component scores, compound 1 is the best joint inhibitor and compound 6 is the worst joint inhibitor towards both enzymes. Score values from the 2nd principal component are directly connected to the joint inhibition. To connect the molecular structure with the experimental data, the molecular charge distributions for compounds 1, 3, 10, 7 and 6 are presented in Figure 3. The compounds are listed in order of their joint inhibitory potency. Although this activity is governed by a combination of steric and electronic effects, Figure 1 reveals that regions of neutral charge distribution (depicted in green) are detrimental to inhibition (compounds 6 and 7), while positively charged groups contribute favorably (compounds 1, 3 and 10). These findings prompted further investigation via quantum-chemical docking to characterize the specific interactions occurring within the active sites.

Quantum chemical docking

Both hAChE and hBChE are serine hydrolases sharing a similar globular form, possessing a conserved catalytic triad (Ser-His-Glu) vital for choline ester hydrolysis [48]. This triad resides within a deep active site gorge, influencing substrate access and binding. The active site comprises key domains: a) an esteratic site with the catalytic serine triad (Ser, His, Glu), b) an acyl pocket - a hydrophobic region hosting the ester’s acyl group (hAChE (pdb 4ey4): Phe295, Phe297, Phe338, Trp236; hBChE (pdb 1p0i): Leu286, Phe329, Phe398, Trp231), c) a choline subsite recognizing the substrate’s quaternary ammonium group (hAChE: Trp86, Tyr337; hBChE: Trp82, Ala328), and d) an oxyanion hole formed by N-H dipoles stabilizing the substrate’s carbonyl oxygen atom (hAChE: Gly121, Gly122, Ala204; hBChE: Gly116, Gly117, Ala119). Generally, hAChE possesses a narrower, deeper, and more defined catalytic gorge optimized for ACh binding. It has a prominent peripheric aromatic site (PAS, Trp286, Tyr124, Tyr72, Tyr341, Asp74) critical for ACh binding. hBChE has a wider, more open catalytic gorge that accepts a broader range of substrates, including butyrylcholine, succinylcholine and cocaine, with a less defined PAS (Tyr332, Asp70). Overall, the differing ligand binding specificities of AChE and BChE can be largely explained by a significant reduction in aromatic residues within the hBChE catalytic gorge. While the hAChE gorge contains 10 such residues, hBChE has only four conserved [49].

Using quantum-chemical docking and the PM7 Hamiltonian, compounds 1-11 were successfully docked into the active sites of hAChE and hBChE. The resulting structures were observed to be located deep within the active site gorges (Figures 4-7 and Figures S45-S55).

Figure 4.: Placement of compound 1 inside the active site of a) AChE and b) BChE obtained by extensive quantum-chemical docking and geometry optimization at the PM7 semiempirical level of the theory. (Violet color for carbon atoms: one of the docked structures of minimal energy, non-polar hydrogen atoms are omitted for clarity)

Figure 5.: Placement of compound 5 inside the active site of a) AChE and b) BChE obtained by extensive quantum-chemical docking and geometry optimization at the PM7 semiempirical level of the theory. (Violet color for carbon atoms: one of the docked structures of minimal energy, non-polar hydrogen atoms are omitted for clarity)

Figure 6.: Placement of compound 6 inside the active site of a) AChE and b) BChE obtained by extensive quantum-chemical docking and geometry optimization at the PM7 semiempirical level of the theory. (Violet color for carbon atoms: one of the docked structures of minimal energy, non-polar hydrogen atoms are omitted for clarity)

Figure 7.: Placement of compound 11 inside the active site of a) AChE and b) BChE obtained by extensive quantum-chemical docking and geometry optimization at the PM7 semiempirical level of the theory. (Violet color for carbon atoms: one of the docked structures of minimal energy, non-polar hydrogen atoms are omitted for clarity)

The most important interactions between all α-acylaminoacetamides and hAChE/hBChE are hydrophobic, π-π interactions and/or Tshape complexes/hydrogen bonds with various amino acids. Within the active site of hAChE, 1 interacts via π-π interactions with Trp86 and Try337 and hydrogen bonds with Tyr124 and Gly121/122 (Figure 4a, Table 3). It is similarly placed in hBChE where it interacts with Trp82 and Gly116/117 (Figure 4b, Table 4). Compound 5 was interacting with hAChE via a T-shape complex with Tyr337 and a strong hydrogen bond with Phe295, CH3-π interactions with Trp286, and additional electrostatic interactions with Tyr124/341 (Figure 5a, Table 3). Within the larger active site of hBChE, 5 is better placed and interacts with Trp82 (CH3-π), T-shape complex with Tyr332 and Trp231 (Figure 5b, Table 4). Within the hAChE active site, compound 6 creates a strong hydrogen bond with Tyr124 and forms a T-shaped complex with Trp286 and CH2-π interactions with Tyr341 (Figure 6a, Table 3). In the hBChE 6 interacts electrostatically with Trp82, forms a hydrogen bond with Gly116/117, a T-shape complex with Trp231 and Phe329, and interacts additionally with Leu286 (Figure 6b, Table 4). Compound 11 is interacting with hAChE via CH3-π interactions with Trp86, a hydrogen bond with Tyr124 and Tyr337, and π-π interactions with Tyr341 (Figure 7a, Table 3). Within the hBChE active site, interactions are a three-centred hydrogen bond with Gly116/117 and CH3-π interactions with Phe329 (Figure 7b, Table 4).

Comp. | r / nm
Trp86 | Tyr124 | Trp236 | Phe338 | Tyr337 | Tyr341
1 | 0.4138(π-π) | - | - | - | 0.2471, 0.3670(H-bond, π-π) | -
2 | 0.3757(electrostatic) | 0.2663(H-bond) | - | - | 0.4583(electrostatic) | 0.3828(π-π)
3 | - | - | - | - | - | 0.4924(T-shape)
4 | 0.4763(electrostatic) | 0.3150(H-bond) | - | 0.5041(T-shape) | 0.2616(H-bond) | 0.5114(T-shape)
5 | - | 0.2396(H-bond) | - | - | 0.2797(H-bond) | 0.2966(electrostatic)
6 | - | 0.2582(H-bond) | - | - | - | 0.4080(CH3-π)
7 | - | 0.2657(H-bond) | - | - | - | 0.4350(CH3-π)
8 | - | - | - | - | 0.2659(H-bond) | 0.3765(T-shape)
9 | - | 0.2765(H-bond) | - | - | - | 0.4526(T-shape)
10 | 0.3376(CH2-π) | - | - | - | 0.2644(H-bond) | -
11 | 0.3334(CH3-π) | 0.2862(H-bond) | - | - | 0.2782(H-bond) | 0.4392(π-π)

Comp. | r / nm
Trp82 | Trp231 | Leu286 | Phe329 | Tyr332 | Phe398
1 | 0.4075(π-π) | - | - | 0.2546(CO-π) | - | -
2 | - | - | - | 0.5447(T-shape) | 0.4511(π-π) | -
3 | 0.3754(π-π) | 0.5350(electrostatic) | 0.5003(electrostatic) | - | - | -
4 | 0.5259(π-π) | 0.4528(T-shape) | - | 0.5429(T-shape) | - | -
5 | 0.3508(CH3-π) | 0.5023(T-shape) | - | - | 0.4724(T-shape) | -
6 | 0.4221(electrostatic) | 0.4537(T-shape) | 0.5322(electrostatic) | 0.5293(T-shape) | - | 0.5301(T-shape)
7 | 0.4541(T-shape) | 0.3148(CH3-π) | - | 0.5183(CH3-π) | - | -
8 | 0.4265(T-shape) | 0.5062(T-shape) | 0.5211(electrostatic) | 0.5041(T-shape) | - | -
9 | - | 0.4853(T-shape) | 0.5166(electrostatic) | 0.5388(T-shape) | - | -
10 | 0.3791(CH2-π) | - | - | - | - | -
11 | - | - | - | 0.4215(CH2-π) | - | -

Analysis of calculated standard Gibbs binding energies for optimized complexes at the PM7 semiempirical level of theory for all compounds (Figures 4 to 7 and Figures S45 to S55, Table 5) partially supported the experimental data. In the case of joint inhibition, where compound 1 is the best overall joint inhibitor and compound 6 is the worst, for hAChE, the standard Gibbs energy of binding for 1 is the 2nd-lowest, whereas for 6 it is the highest. For hBChE, the data do not fully support the experimental findings; better insight can be gained from interaction analysis within the active site.

Comp. | hAChE | hBChE
ΔbH○ / kJ mol-1 | T298.15 K·ΔbS○ / kJ mol-1 | ΔbG○ / kJ mol-1 | ΔbH○ / kJ mol-1 | T298.15 K·ΔbS○ / kJ mol-1 | ΔbG○ / kJ mol-1
1 | −368.84 | −96.53 | −272.31 | −282.25 | −80.05 | −202.19
2 | −340.96 | −94.92 | −246.04 | −330.96 | −80.90 | −250.05
3 | −353.17 | −98.94 | −254.24 | −407.73 | −90.06 | −317.67
4 | −350.32 | −98.32 | −252.00 | −401.27 | −89.89 | −311.38
5 | −319.62 | −84.28 | −235.35 | −279.52 | −66.77 | −212.75
6 | −272.60 | −80.54 | −192.06 | −311.28 | −62.39 | −248.89
7 | −293.48 | −86.28 | −207.21 | −349.59 | −69.21 | −280.38
8 | −336.22 | −92.29 | −243.93 | −365.83 | −95.28 | −270.55
9 | −357.16 | −80.37 | −276.79 | −332.72 | −66.92 | −265.80
10 | −279.57 | −71.49 | −208.08 | −380.15 | −75.21 | −304.94
11 | −272.61 | −77.71 | −194.90 | −296.90 | −72.42 | −224.49

Physico-chemical properties, lipophilic and ligand efficiency

The potential of compounds to cross the blood-brain barrier (BBB) was assessed in silico by comparing their calculated physico-chemical properties (molecular weight (MW), hydrophobicity (log P), the number of hydrogen bond donors (HBDs) and acceptors (HBAs), rotatable bonds (RBs), and polar surface area (PSA)) against recommended thresholds for passive transport into the central nervous system (CNS) by known CNS-active drugs [50-52]. Based on the calculated values (Figure 8, Table S1), most of the tested compounds are predicted to penetrate the BBB following oral administration. Compounds 3, 5 and 8 had a higher number of PSA than recommended and therefore should have a moderate possibility for passing the BBB. Also, compound 3 had a slightly higher number of RB than recommended.

Figure 8.: Radar plot of physico-chemical properties of prepared compounds. MW stands for molecular weight, logP for hydrophobicity, HBD for number of hydrogen bond donors, HBA for hydrogen bond acceptors, RB for the number of rotatable bonds and PSA for polar surface area. The recommended values for the CNS-active drugs are presented by a red line

Analysis of the calculated log P parameter allows for correlation with initial activity and calculation of Ligand Efficiency (LE) and Lipophilic Efficiency (LLE). These metrics are crucial for identifying promising initial hits and guiding the optimization of compounds into viable leads, as detailed in Table S1. Generally, higher LE and LLE values are preferred, indicating a favourable balance between potency and lipophilicity. However, this relationship isn't always straightforward; very high lipophilicity (high log_P_) can negatively impact solubility and bioavailability, even with good LE/LLE values. Consequently, compound 1 represents the strongest lead in this series, exhibiting a relatively high p_K_i for hBChE (5.25) and a moderate log P (1.7), resulting in a good LLE value for hBChE (3.56). In contrast, compounds 5 and 8 displayed the high p_K_i values for hBChE, but their significantly higher log_P_ values (2.94 and 2.68, respectively) reduce their LLE values compared to compound 1. This suggests that reducing lipophilicity could be a beneficial optimization strategy for these compounds. With a p_K_i of 3.86, a log_P_ of 2.09, and a resulting LLE of 2.20 for hAChE, compound 11 represents a candidate that could be further optimized for potency.