Dipeptide-Derived Alkynes as Potent and Selective Irreversible Inhibitors of Cysteine Cathepsins

Abstract

The potential of designing irreversible alkyne-based inhibitors of cysteine cathepsins by isoelectronic replacement in reversibly acting potent peptide nitriles was explored. The synthesis of the dipeptide alkynes was developed with special emphasis on stereochemically homogeneous products obtained in the Gilbert-Seyferth homologation for C≡C bond formation. Twenty-three dipeptide alkynes and 12 analogous nitriles were synthesized and investigated for their inhibition of cathepsins B, L, S, and K. Numerous combinations of residues at positions P1 and P2 as well as terminal acyl groups allowed for the derivation of extensive structure-activity relationships, which were rationalized by computational covalent docking for selected examples. The determined inactivation constants of the alkynes at the target enzymes span a range of >3 orders of magnitude (3-10 133 M^-1 s^-1). Notably, the selectivity profiles of alkynes do not necessarily reflect those of the nitriles. Inhibitory activity at the cellular level was demonstrated for selected compounds.

Figure 1

Cathepsin B-inhibitory dipeptide nitriles 1a (reported by Greenspan et al.) and 1b as lead compounds for the design of dipeptide-derived alkynes 2a and 2b, respectively, as potentially irreversible inhibitors of cathepsin B. Protease subsite-targeting moieties P1–P3 are highlighted for nitrile 1a.

Scheme 1. Synthesis of Dipeptide Nitriles 1a–e

Reagents and conditions: (a) allyl bromide, K₂CO₃, acetone, reflux, 2 h; (b) NaI, acetone, 5 h; (c) Boc-L-serine, NaH, DMF, 0 °C, 15 min, rt, 30 min; (d) iBCF, NMM, NH₃, THF, −10 °C, 10 min, rt, 30 min; (e) TFA/CH₂Cl₂ (1:1), 2 h; (f) N-Boc-amino acid, DIPEA, PyBOP, THF, 3 h; (g) acyl chloride, TEA, CH₂Cl₂, 2 h, or carboxylic acid, DIPEA, PyBOP, THF, 3 h; (h) cyanuric chloride, DMF, 2 h; (i) Pd(PPh₃)₄, morpholine, CH₂Cl₂, 30 min.

Scheme 2. Synthesis of Dipeptide Alkyne 18 as a Diastereomeric Mixture

See Scheme 6 for the structural definition of stereochemically pure compounds. Reagents and conditions: (a) PyBOP, DIPEA, NaBH₄, THF, 1 h; (b) Dess-Martin periodinane, CH₂Cl₂, 4 h; (c) dimethyl (1-diazo-2-oxopropyl)phosphonate, K₂CO₃, MeOH, 0 °C, 2 h, rt, overnight; (d) TFA/CH₂Cl₂ (1:1), 2 h; (e) N-Boc-3-methyl-l-phenylalanine, PyBOP, DIPEA, THF, 3 h; (f) 2,4-difluorobenzoyl chloride, NMM, CH₂Cl₂, 2.5 h; (g) NaOH, THF/MeOH (3:1), overnight.

Figure 2

¹H NMR spectrum of dipeptide alkyne 15 in CDCl₃ and detailed view of relevant signals (right), whose doubling indicates the presence of a diastereomeric mixture.

Figure 3

Carboxy-functionalized P1 residues intended as occluding loop-targeting moieties in the dipeptide-derived cathepsin B inhibitor. Structure of the P1 residue described by Greenspan et al. (left; X = N) and Schmitz et al. (middle; X = N) and derived alternative P1 residue (right) for the design of cathepsin B-directed dipeptide alkynes (X = CH).

Scheme 3. Synthesis of Dipeptide Alkyne 28 Containing a Carboxy-Functionalized 1,2,3-Triazolyl Moiety at P1

Reagents and conditions: (a) triflyl azide, CuSO₄·5H₂O, K₂CO₃, MeOH/H₂O (3:1), overnight; (b) methyl propiolate, CuSO₄·5H₂O, sodium ascorbate, DMSO/H₂O (2:1), overnight; (c) DIPEA, NaBH₄, PyBOP, THF, 1 h; (d) Dess-Martin periodinane, CH₂Cl₂, 1 h; (e) dimethyl (1-diazo-2-oxopropyl)phosphonate, K₂CO₃, MeOH, 0 °C, 2 h, rt, 3 h; (f) TFA/CH₂Cl₂ (1:1), 2 h; (g) Boc-3-methyl-l-phenylalanine, PyBOP, DIPEA, THF, 3 h; (h) 4-fluorobenzoyl chloride, NMM, CH₂Cl₂, 2 h; (i) NaOH, THF/MeOH (3:1), overnight.

Figure 4

(A) ¹H NMR spectrum of dipeptide alkyne 25 in CD₃Cl and detailed view of selected signals (right). (B) HPLC chromatogram of 25 (left) and ESI-MS spectra corresponding to the peaks (right).

Scheme 4. Synthesis of Dipeptide Nitriles 35a–c with a Carboxy-Functionalized 1,2,3-Triazolyl Residue at P1

Reagents and conditions: (a) triflyl azide, CuSO₄·5H₂O, K₂CO₃, MeOH/H₂O (3:1), overnight; (b) methyl propiolate or methyl butynoate, CuSO₄·5H₂O, sodium ascorbate, DMSO/H₂O (2:1), overnight; (c) iBCF, NMM, NH₃, THF, −15 °C, 10 min, rt, 30 min; (d) TFA/CH₂Cl₂ (1:1), 2 h; (e) Boc-3-methyl-l-phenylalanine, DIPEA, PyBOP, THF, 3 h; (f) 4-fluorobenzoyl chloride, TEA, CH₂Cl₂, 2 h; (g) cyanuric chloride, DMF, 3 h; (h) NaOH, THF/MeOH (3:1), overnight or pig liver esterase, KH₂PO₄ buffer (0.2 M, pH 7.0)/acetone (10:1), 6–10 days.

Scheme 5. Synthesis of Serine-Based Dipeptide Nitrile 43 with a Carboxy-Functionalized 1,2,3-Triazolyl Residue at P1

Reagents and conditions: (a) methyl-2-azidoacetate, CuSO₄·5H₂O, sodium ascorbate, DMSO/H₂O (1:2), 0 °C, 1 h, rt, overnight; (b) iBCF, NMM, NH₃, THF, −15 °C, 10 min, rt, 30 min; (c) TFA/CH₂Cl₂ (1:1), 2 h; (d) Boc-3-methyl-l-phenylalanine, PyBOP, DIPEA, THF, 3 h; (e) 4-fluorobenzoyl chloride, TEA, CH₂Cl₂, 2 h; (f) cyanuric chloride, DMF, 2 h; (g) LiOH, THF/H₂O (5:1), 0 °C, 5 min; procedure following ref (95).

Scheme 6. Synthesis of Dipeptide Alkynes 2a–m

Reagents and conditions: (a) acetyl chloride, MeOH, reflux, 2 h; (b) Boc₂O, TEA, THF, 0 °C, 45 min, rt, overnight; (c) 2,2-dimethoxypropane, BF₃·OEt₂, acetone, 3 h; (d) LiAlH₄, THF, 45 min; (e) oxalyl chloride, DIPEA, DMSO, CH₂Cl₂, −78 °C, 80 min, 0 °C, 10 min; (f) dimethyl (1-diazo-2-oxopropyl)phosphonate, K₂CO₃, MeOH, 0 °C, 4 h; (g) HCl (4 M)/MeOH (3:5), reflux, 1 h; (h) Boc₂O, TEA, THF, 0 °C, 45 min, rt, overnight; (i) NaH, DMF, 0 °C, 5 min, rt, 1.5 h; (j) TFA/CH₂Cl₂ (1:1), 2 h; (k) N-Boc-amino acid, DIPEA, PyBOP, THF, 3 h; (l) acyl chloride, TEA, CH₂Cl₂, 2 h, or carboxylic acid, DIPEA, PyBOP, THF, 3 h; (m) Pd(PPh₃)₄, morpholine, CH₂Cl₂, 30 min.

Figure 5

Structure of the two symmetry-independent molecules of compound 48 as observed in the crystalline state by X-ray diffraction analysis with the atom labeling scheme. Thermal displacement ellipsoids are shown at the 50% probability level. The hydrogen bond between the two molecules is shown as a dashed line.

Scheme 7. Synthesis of Nitrile- and Alkyne-Based Dipeptide Alkynes 56a–f and 62, respectively

Reagents and conditions: (a) propargylamine or aminoacetonitrile (in THF/H₂O/NaOH), iBCF, NMM, THF, −25 °C, 10 min, rt, 30 min; (b) TFA/CH₂Cl₂ (1:1), 2 h; (c) acidic chloride, TEA, CH₂Cl₂, 2 h, or carboxylic acid, DIPEA, PyBOP, THF, 2 h; procedure following (d) isobutyl bromide, tetrabutylammonium iodide, EtOH/3 M NaOH (1:1), 3 days; (e) Boc₂O, 1 day; (f) KMnO₄, acetic acid, 2.5 h; (g) propargylamine, iBCF, NMM, THF, −25 °C, 10 min, rt, 30 min; (h) 4-fluorobenzoyl chloride, TEA, CH₂Cl₂, 2 h.

Figure 6

Determination of the type of inhibition of cathepsin B by dipeptide nitrile 1a. (A) Initial rates of substrate turnover as a function of substrate concentration (x axis) and inhibitor concentration (legend). (B) Double-reciprocal plot of 1/v_i vs 1/[S] for different inhibitor concentrations (transformed data set identical to that of panel A). The lines intersect in the fourth quadrant, which allows the characterization of 1a as a noncompetitive inhibitor with an α value of >1. Measurements were performed as duplicate determinations in assay buffer (pH 6.0) containing 200 μM Z-RR-AMC, 25 ng/mL cathepsin B, and 1.5% DMSO. Measured values ± SEM are shown.

Figure 7

Inhibition of cathepsin B by dipeptide alkyne 2b. (A) Turnover of Z-RR-AMC by cathepsin B in the presence of increasing concentrations of dipeptide alkyne 2b. (B and C) Replots of pseudo-first-order rate constants, k_obs, and initial velocities, v₀, respectively, vs inhibitor concentration.

Figure 8

Jump-dilution inhibition experiment for nitrile 1b and alkyne 2b with cathepsin B to prove the irreversible inhibition by 2b. The inhibitors were preincubated in a volume of 30 or 180 μL in the presence of enzyme as schematically shown in Figure S46. The highly concentrated solution was diluted to 180 μL immediately before measurement. Subsequently, the reaction was started by substrate addition so that equal enzyme and inhibitor concentrations were achieved at the start of each measurement. The measurement was performed as a duplicate determination in assay buffer (pH 6.0) containing 200 μM Z-RR-AMC, 25 ng/mL cathepsin B, and 1.5% DMSO. The concentrations indicated refer to the final concentrations during the measurement after dilution.

Figure 9

Selectivity profile of stereochemically pure alkyne-based inhibitors 2a, 2b, and 28 (bottom) and corresponding nitriles (top). Shown are the negative decadic logarithms of the K_i values and the decadic logarithm of the inactivation constants k_inact/K_I. Thus, larger values are equivalent to higher inhibition potentials of the compounds. The measurement was performed in three independent experiments (each as a duplicate determination) in assay buffer (pH 6.0) with 1.5% DMSO. n.i. = no inhibition; i.e., no evidence of irreversible inhibition was discernible within the considered time and concentration ranges.

Figure 10

Influence of the side chain at P1 in depicted dipeptide nitriles on their inhibitory selectivity for cathepsins B, S, L, and K. The measurement was performed in three independent experiments (each as a duplicate determination) in assay buffer (pH 6.0) containing 1.5% DMSO.

Figure 11

Influence of the P3 substituents in depicted dipeptide alkynes on their selectivity for cathepsins B, S, L, and K. n.i. = no inhibition; i.e., no evidence of irreversible inhibition was discernible within the considered time and concentration ranges.

Figure 12

(A) Influence of the substituent at the meta position of the benzoyl residue at P3 on the inhibition of cathepsin L. The measurement was performed in three independent experiments (each as a duplicate determination) in assay buffer (pH 6.0) containing 10 μM Z-FR-AMC, 25 ng/mL cathepsin L, and 1.5% DMSO. (B) Relationship between the van der Waals radii of the substituents at the meta position of the P3 benzoyl residue and the cathepsin L inactivation constant. The data point for the CF₃ substituent was excluded from calculating the regression line.

Figure 13

Selectivity profile of dipeptide alkynes with a glycine-derived moiety at P1. The lead compound 2b is included for comparison. The measurement was performed in three independent experiments (each as a duplicate determination) in assay buffer (pH 6.0) containing 1.5% DMSO. n.i. = no inhibition; i.e., no evidence of irreversible inhibition was discernible within the considered time and concentration ranges.

Figure 14

Molecular models for covalent enzyme–inhibitor complexes predicted in silico. Cathepsins B, S, and L in cartoon and transparent surface representations are colored green, orange, and cyan, respectively. Interacting protein residues are shown as sticks, colored by atom type and labeled. Inhibitors (A–C) 1b, (D–F) 2b, and (G–I) 2k are shown as gray sticks and colored by atom type. Pocket binding sites S1–S3 are indicated by red labels. Intermolecular hydrogen bonds, salt bridges, π–π and halogen hydrogen bond interactions are depicted as black, magenta, cyan, and purple dashed lines, respectively. Figure generated in Maestro (Schrödinger).

Figure 15

Quartets of matched molecular pairs of dipeptide nitriles and alkynes, which are structurally related to each other by double-transformation cycles.

Figure 16

Nonadditivity analysis in double-transformation cycles (horizontal arrows for m-tolyl → m-iodophenyl and vertical arrows for p-fluorobenzoyl → diphenylacetyl) for (A) dipeptide nitrile and (B) dipeptide alkyne quartets Q1 and Q2, respectively, as exemplified for the inhibition of cathepsin B.

Figure 17

Analysis of expression of (A) cathepsin B and (B) cystatin B in the total cell lysate by Western blotting.

Figure 18

Dipeptide nitrile 1b and dipeptide alkyne 2b compared with the literature inhibitor CA-074 in the U251-MG cell lysate. (A) Turnover of internally quenched substrate Abz-GIVR↓AK(Dnp)-NH₂ (↓ indicates the cleavage site) as measured by increasing fluorescence intensities after preincubation for 30 min with the respective inhibitor. (B) Relative inhibition normalized to the inhibitory effect of CA-074 (mean ± SEM). The measurement was performed as a duplicate determination in assay buffer (pH 6.0) containing 0.5 mg/mL protein, 100 μM Abz-GIVRAK(Dnp)-NH₂, and 1% DMSO.

Figure 19

(A) Increasing fluorescence intensity originating from the turnover of internally quenched substrate Abz-GIVRAK(Dnp)-NH₂ shown for an examplary measurement as a duplicate determination and (B) initial velocities averaged over three measurements of v₀ = f([2b]) for inhibitor 2b on viable U87-MG cells. In panel A, the primary curves show an upward curvature presumably due to the continuously secreted enzyme. Initial velocities were determined from substrate turnover curves via linear regression over the first 600 s. For panel B, analysis of v₀ = f([2b]) was performed according to eq III. Measurements were performed in three independent experiments (each as duplicate determinations) in assay buffer (pH 6.0) containing 100 μM Abz-GIVRAK(Dnp)-NH₂, 25 ng/mL cathepsin B, and 1.5% DMSO. Shown are mean values ± SEM.