Structure-Based Design and Biological Evaluation of Novel Caspase-2 Inhibitors Based on the Peptide AcVDVAD-CHO and the Caspase-2-Mediated Tau Cleavage Sequence YKPVD314

Abstract

Alzheimer’s disease (AD) was first described by Alois Alzheimer over 100 years ago, but there is still no overarching theory that can explain its cause in detail. There are also no effective therapies to treat either the cause or the associated symptoms of this devastating disease. A potential approach to better understand the pathogenesis of AD could be the development of selective caspase-2 (Casp2) probes, as we have shown that a Casp2-mediated cleavage product of tau (Δtau314) reversibly impairs cognitive and synaptic function in animal models of tauopathies. In this article, we map out the Casp2 binding site through the preparation and assay of a series of 35 pentapeptide inhibitors with the goal of gaining selectivity against caspase-3 (Casp3). We also employed computational docking methods to understand the key interactions in the binding pocket of Casp2 and the differences predicted for binding at Casp3. Moreover, we crystallographically characterized the binding of selected pentapeptides with Casp3. Furthermore, we engineered and expressed a series of recombinant tau mutants and investigated them in an in vitro cleavage assay. These studies resulted in simple peptidic inhibitors with nanomolar affinity, for example, AcVDV(Dab)D-CHO (24) with up to 27.7-fold selectivity against Casp3. Our findings provide a good basis for the future development of selective Casp2 probes and inhibitors that can serve as pharmacological tools in planned in vivo studies and as lead compounds for the design of bioavailable and more drug-like small molecules.

While the pathogenesis of Alzheimer’s disease (AD) is still uncertain, genetic studies suggest that both the formation of abnormal amyloid-β aggregates¹ and neurotoxic fibrils formed by the tau protein are part of the observed pathology. Post-translational modifications (PTMs) of tau, such as phosphorylation and acetylation, play a crucial role in normal physiological conditions of neurons. Phosphorylation serves as a regulator of tau functions like stimulation and stabilization of microtubule assembly.^2,3 In AD and other tauopathies, the tau protein undergoes hyperphosphorylation, which promotes the accumulation and mistargeting of abnormal tau in dendritic spines.^4,5 Fibrillar tau has been dissociated from neuronal death and network dysfunction, implying the contribution of nonfibrillar species in the mediation of cognitive anomalies.^6,7 We reported that tau mislocalization to dendritic spines mediates synaptic dysfunction independently of neurodegeneration,⁴ whereas we have shown that formation of Δtau314, a 35 kDa tau cleavage product, can reversibly impair cognitive and synaptic function in animal and cellular models of tauopathies.⁷ Δtau314 originates from the cleavage of tau by caspase-2 (Casp2) at aspartate-314, and therefore, blocking this truncation can be a promising therapeutic approach for AD or other tauopathies, like Huntington’s disease (HD) and Lewy body dementia.⁸⁻¹⁰ Furthermore, Casp2 is also reported to be involved in the process of the progression of nonalcoholic fatty liver disease (NAFLD) to nonalcoholic steatohepatitis (NASH) by promoting proteolytic activation of site 1 protease (S1P).^11,12 Therefore, inhibition of Casp2 would also be beneficial in preventing or treating stress-induced fatty liver diseases.¹¹

Caspases (cysteine-dependent aspartate-directed proteases) are a family of protease enzymes that cleave target proteins on the N-terminal side of aspartic acid and are typically subdivided into three groups: initiator caspases (e.g., caspase-2, -8, -9, -10), effector caspases (e.g., caspase-3, -6, -7), and inflammatory caspases (e.g., caspase-1, -4, -5).¹³ Research has classically associated caspases with apoptosis but they also appear to play multiple roles in neurodegenerative diseases, particularly tauopathies.^14,15 The caspase cleavage of tau is mainly caused by Casp2 and Casp3/Casp6.¹⁶⁻¹⁸ Extensive success in the structural biology of caspases has provided a wealth of information regarding enzyme architecture, activation, and the basis for selectivity of these enzymes for different substrates.¹⁹ All caspase catalytic domains share a similar dimeric quaternary structure, with two identical monomers associated about the twofold rotation axis to form 1 large beta sheet of 12 strands. The monomers adopt a highly homologous fold, with highly variable loops (L1–L4) that lead to the specific and varying substrate selectivity of the different caspases. Following proteolytic cleavage in the L2 loop, a rearrangement of the L2 and L3 loops occurs, resulting in a repositioning of the active site cysteine and enzyme activation. Substrates and inhibitors bind in a cleft formed atop the L3 loop, with L2 and L4 on either side of this cleft. The monomer architecture is illustrated in Figure S1 (Supporting Information).

Within the caspase family, only Casp2 requires a motif of five amino acids, instead of four, as a specific recognition sequence for efficient substrate cleavage.^20,21 The closest similarities to the cleavage specificity of Casp2 are found with Casp3, 7, and 8 where an aspartic acid is also favored at positions P1 and P4 for effective binding (Figure 1). This close homology between Casp2 and Casp3 makes the design of selective Casp2 inhibitors challenging. All caspases conduct substrate cleavage by the same mechanism using a catalytic Cys/His dyad architecture in their active site. The aspartic acid of substrates is deeply buried in the basic S1 pocket where it interacts with two arginines and one glutamine residue (e.g., Arg202, Arg361, and Gln301 Casp2). Casp3, 6, and 7 are limited to smaller aliphatic residues at P2 (e.g., alanine and valine, as shown in Figure 1), as their S2 binding sites show a narrower spatial shape recognition than other caspases. Notably, Casp2 accommodates larger substrates in S2 and also contains a negatively charged Glu52 as a unique structural feature.^16,22 Therefore, we hypothesized that the introduction of sterically demanding P2 residues, as recently shown by Maillard et al.,²³ and/or the use of positively charged amino acids (e.g., lysine and arginine) at P3 could lead to enhanced potency at Casp2 over Casp3. The S3 binding site of all caspases is characterized by a conserved arginine, which can provide ionic interactions with negatively charged amino acids (e.g., glutamic acid) of the substrates at P3 (Figure 1). Casp8 and Casp9 are characterized by a second arginine, allowing for a tighter binding of ionic groups in its S3 pocket. Casp2 alone appears to prefer a smaller hydrophobic valine at P3. The S4 binding site of the caspase family can be subdivided into three groups. Casp1, 4, 5, and 11 consist of an extended, shallow hydrophobic binding pocket, favorably engaging aromatic amino acids (e.g., tryptophan, Figure 1). In contrast, Casp2, 3, and 7 have a significantly smaller hydrophilic binding site typically preferring an aspartic acid (Figure 1). Recent work has demonstrated that a homoglutamic acid P4 sidechain can confer excellent selectivity for Casp2 versus other caspases. S4 structures of Casp6, 8, 9, and 10 show little selectivity toward specific amino acids and represent a hybrid between the two binding sites mentioned above.²² Casp2 is the only caspase that prefers an occupied binding subsite S5 for activity, for example, valine or tryptophan.²⁴ This unique requirement for its substrate guarantees efficient substrate cleavage and a substrate-selectivity preference for Casp2 over other caspases.^16,25

Figure 1.

Peptide substrate preferences for Casp1, 2, 3, 4, 5, 6, 7, 8, 9, and 10.^26,28 P1–P5 represents the respective amino acid position of the peptidic substrate or inhibitor. P1 always marks the position of the aspartic acid (D) cleavage site. P4 and P5 (for Casp2) represent the N-terminal position, respectively.

The goal of this study was the exploration of the Casp2 binding pocket to elucidate the interactions of simple pentapeptide inhibitors and the target enzyme. This would allow for further optimization of both peptidomimetics and small-molecule inhibitors. We also employed computational covalent docking studies to identify critical interactions in the binding pocket. Based on these results, we synthesized a series of pentapeptides to better understand the specific features of the active site, which would allow us to more readily achieve our goal to discover new, selective, and potent Casp2 probes. A particular focus was on the development of selective Casp2 inhibitors against Casp3, on the one hand because the canonical Casp2 inhibitor AcVDVAD-CHO (2) provides equally high affinities for Casp3 (see e.g. Table 2 and Maillard et al.²³) and on the other hand because of the high concentrations of active Casp3, compared to other caspases, which is present in almost all tissues²⁶ and, furthermore, because neuroprotective effects of a Casp3 inhibitor have already been demonstrated in an HD model.²⁷

Table 2. Binding Data (pK_i Values) of Pentapeptides 1–38 at Casp2 and Casp3^a,b.

		pKi ± SD
sequence	Comp.	Casp2	N	Casp3	N	Ki(Casp3)/Ki(Casp2)
AcYKPVD-CHO	1	4.56 ± 0.43c	3	4.73 ± 0.10	4	0.68
AcVDVAD-CHO	2	7.85 ± 0.08c	5	7.73 ± 0.09	5	1.33
AcVDPVD-CHO	3	4.82 ± 0.12c	2	5.70 ± 0.04	2	0.13
AcVDKVD-CHO	4	7.40 ± 0.02c	5	7.28 ± 0.09	4	1.32
AcVDVKD-CHO	5	7.63 ± 0.05c	4	6.91 ± 0.07	4	5.16
AcVDVVD-CHO	6	7.77 ± 0.12c	4	7.81 ± 0.13	4	0.91
AcVDKAD-CHO	7	7.35c	1	7.22 ± 0.22	3	1.35
		7.09 ± 0.11d	2			0.73
AcYKKVD-CHO	8	<4c	2	<4.5	2
AcYKVKD-CHO	9	<4c	1	<4	2
		<4d	1
AcVKVVD-CHO	10	<4d	2	5.16 ± 0.06	2	<0.07
AcVKKVD-CHO	11	<4c	1	4.78 ± 0.02	2	<0.17
		<4d	1
AcVKVKD-CHO	12	<4c	2	<4	2
AcVKPVD-CHO	13	<4c	2	4.88 ± 0.04	2	<0.16
AcYKVVD-CHO	14	<4d	2	5.31 ± 0.07	2	<0.05
AcYDKVD-CHO	15	7.39 ± 0.01c	2	7.75 ± 0.13	3	0.44
		7.18d	1			0.27
AcYDVKD-CHO	16	7.83 ± 0.04c	4	7.13 ± 0.02	4	5.06
AcYDPVD-CHO	17	<4c	1	5.61 ± 0.04	2	<0.02
		<4d	1
AcVD(Orn)VD-CHO	18	7.37 ± 0.05c	2	7.52 ± 0.07	3	0.70
		6.92d	1			0.25
AcVD(Dab)VD-CHO	19	7.44 ± 0.08c	4	7.54 ± 0.02	4	0.78
AcVD(Dap)VD-CHO	20	6.53 ± 0.07c	2	7.38 ± 0.15	3	0.14
		6.18d	1			0.06
AcVDRVD-CHO	21	7.38 ± 0.13c	3	7.21 ± 0.08	3	1.48
AcVD(AcK)VD-CHO	22	7.70 ± 0.02c	2	7.96 ± 0.22	3	0.56
		7.41d	1			0.29
AcVDV(Orn)D-CHO	23	7.45 ± 0.07c	4	6.60 ± 0.08	4	7.07
AcVDV(Dab)D-CHO	24	7.26 ± 0.07c	4	5.82 ± 0.18	4	27.7
AcVDV(Dap)D-CHO	25	6.77 ± 0.17c	4	6.14 ± 0.09	4	4.30
AcVDVRD-CHO	26	7.37 ± 0.04c	2	7.12 ± 0.30	3	1.81
		7.07d	1			0.91
AcVDV(AcK)D-CHO	27	7.53 ± 0.18c	2	7.60 ± 0.20	3	0.85
		7.30d	1			0.49
AcTDKTD-CHO	28	6.84 ± 0.17c	2	7.03 ± 0.06	3	0.63
		6.53d	1			0.31
AcVDKID-CHO	29	7.57 ± 0.04c	2	7.57 ± 0.19	3	1.00
		7.16d	1			0.39
AcVDKFD-CHO	30	7.24 ± 0.06c	2	7.34 ± 0.08	3	0.79
		6.78d	1			0.27
AcVDFVD-CHO	31	7.00 ± 0.17c	2	8.02 ± 0.26	3	0.10
		6.50d	1			0.03
AcVDKPD-CHO	32	7.66 ± 0.03c	2	7.70 ± 0.25	3	0.91
		7.38d	1			0.48
AcVDPPD-CHO	33	<4c	2	<4.5	2
AcVDPKD-CHO	34	<4c	1	<4	2
		<4d	1
AcIDVKD-CHO	35	7.68 ± 0.04c	4	6.80 ± 0.10	4	7.53
AcYDVVD-CHO	36	7.50 ± 0.03d	3	8.43 ± 0.07	3	0.12
AcLDESD-CHO	37	6.85 ± 0.03c	3	7.40 ± 0.02	3	0.28
AcDEVD-CHO	38	6.60 ± 0.32c	3	8.68 ± 0.26	3	0.008

^aSelectivity expressed as K_i(Casp3)/K_i(Casp2).

^bData shown are mean values ± SD of N independent experiments, each performed in duplicate or triplicate. Data were analyzed by nonlinear regression and were best fitted to sigmoidal concentration–response curves.

^cData represent pK_i values determined with Casp2.

^dData represent pK_i values determined with cpCasp2.

Results and Discussion

Design Rationale

A detailed examination of structural data describing complexes of AcVDVAD-CHO (2) bound to both Casp3 (PDBid: 2h65)²⁹ and Casp2 (PDBid: 3r6g)²⁰ provides an excellent basis for understanding the high potency of this inhibitor scaffold toward both Casp2 and Casp3 and helps to identify some of the challenges and opportunities for selective inhibitor design. AcVDVAD-CHO (2) binds to Casp2 and Casp3 nearly identically (Figure 2). Reaction with the C-terminal aldehyde results in the formation of a covalent bond to the catalytic cysteine (Cys163 in Casp3 and Cys155 in Casp2). The P1 aspartic acid is firmly anchored in the S1 subsite by H-bonds and ionic interactions with Arg207, Arg64, and Gln161 (Casp3 numbering). The remainder of the inhibitor lies stretched out over the L3 loop, with the peptide backbone of the P3 valine and P5 valine making H-bonds with the backbone of L3 at Arg207 and Ser209 that resemble those of an antiparallel β-sheet (Figure 2C). All these interactions are largely conserved in complexes of most caspases with peptidic inhibitors and substrates; they confer broad potency but not selectivity. AcVDVAD-CHO (2) also exploits the P4 aspartic acid to increase affinity for Casp3 and Casp2. In both complexes, the Asp carboxylate forms a bridge connecting L3 to L4, with H-bonds to Asn208 of L3 and Phe250 of L4. The L4 loop is quite different in other caspases, and the L3 asparagine is only conserved in caspases 2, 3, and 8 in humans.

Figure 2.

Comparison of experimentally determined Casp3 and Casp2 complexes with the same peptidic inhibitor (AcVDVAD-CHO). (A) Complex with Casp3 (PDB-id: 2h65). (B) Complex with Casp2 (PDB-id: 3r6g). Sequence differences Y204-A and T62-E result (Casp3 numbering) in an expansion of the P2 subsite (encircled with dashed line) and a change in electrostatic environment there. (C) Overlay of AcVDVAD-CHO complexes with Casp3 (yellow) and Casp2 (cyan). Only the Casp3 backbone and sidechains are shown for clarity. Although most H-bonds are conserved in peptidic complexes with all caspases, H-bonds to the backbone of the L4 loop at Phe250 (Casp3 numbering) from the P4 aspartic acid are also present in Casp2 but are not likely more widely conserved. Both complexes are also stabilized by H-bonds from the P4 Asp carboxylate to L3 Asn208 (Asn232 in Casp2) and the P5 valine carbonyl O to the backbone of L3 at Ser209/Thr233. (D) Overlay of models of AcVDVAD-CHO binding derived by docking into Casp3 and Casp2. Peptidic inhibitors were docked into respective protein targets (Casp3, salmon, yellow inhibitor; Casp2-gray, and cyan inhibitor). The rms deviation in modeled vs experimentally observed ligand atom positions is 0.78 and 0.84 Å, respectively.

Comparison of the binding sites of the two structures exposes other differences not exploited by the AcVDVAD-CHO (2) scaffold. The most prominent difference noted by Tang et al.²⁰ is the presence of Glu52 in Casp2 that introduces a negatively charged functional group in the proximity of the S2 and S3 subsites (Figure 2B). This amino acid is an uncharged threonine in Casp3 (Figure 2A). Another significant difference is the presence of a much smaller alanine (Ala228) in Casp2 in place of the tyrosine (Tyr204) of Casp3. This difference results in a significant expansion of the S2 subsite, which we have hoped to target with larger P2 substituents.

In order to accelerate the lead optimization process, a covalent inhibitor docking protocol employing CovDock of the Schrödinger Software Suite was implemented in an attempt to predict relative binding affinities for Casp2 and Casp3 of peptides with a reactive aldehyde (−CHO) warhead at the C-terminus. To confirm that computational methods can recapitulate a reasonable binding pose, covalent docking studies were first performed with ligands AcVDVAD-CHO (2), AcLDESD-CHO (37), and AcDEVD-CHO (38) previously co-crystallized with caspases 2 and 3.^18,28 The resulting docking scores reflect the expected high affinity these peptide inhibitors show experimentally for both caspases (Table 1). Moreover, the predicted poses recapitulate the crystallographically determined structures with good fidelity (Figure 2D).

Table 1. Molecular Modeling of Casp2 and Casp3 and Potential Ligands^a.

sequence	comp	Casp2 cdock affinity	Casp3 cdock affinity
AcYKPVD-CHO	1	–9.348	–9.616
AcVDVAD-CHO	2	–14.178	–12.873
AcVDKVD-CHO	4	–14.613	–13.274
AcVDRVD-CHO	21	–15.055	–12.768
AcVDV(Dab)D-CHO	24	–15.523	–13.871
AcLDESD-CHO	37	–13.773	–13.381
AcDEVD-CHO	38	–13.447	–12.573

^aCalculation using covalent docking in Schrödinger 2021.1. Casp2 = PDBid: 1pyo, Casp3 = PDBid: 3edq. Proteins and ligands prepared as described in the section Molecular Modeling. Cdock affinity (kcal/mol).

This docking protocol was then applied to pentapeptides related to AcVDVAD-CHO that include a positively charged amino acid in place of the P3 valine or the P2 alanine in order to virtually quantify any potential benefit in Casp2 affinity or selectivity of installing a group that can make a favorable electrostatic interaction with the Glu52 unique to Casp2 (Figure 3A). Scoring did suggest that AcVDKVD-CHO (4), but especially AcVDRVD-CHO (21) and AcVDV(Dab)D-CHO (24) might have higher affinity for Casp2 (Table 1). Significant affinity enhancement was not predicted in modeling with Casp3, suggesting that this may indeed be a viable approach to achieve more selective inhibition. Therefore, a series of pentapeptide analogues of AcVDKVD-CHO (4) and AcVDVKD-CHO (5) with different basic amino acids arginine, ornithine, diaminobutyric acid (Dab), and diaminopropionic acid (Dap) were synthesized (Figure 3B). In order to test the hypothesis that the charge is beneficial, and not just the larger sidechains, inhibitors AcVD(AcK)VD-CHO (22) and AcVDV(AcK)D-CHO (27) were also synthesized as negative controls with a nonbasic terminal amide.

Figure 3.

(A) Schematic diagram of the Casp2 binding site (created with BioRender.com).^30,31 P1–P5 represents the respective amino acid position of the peptidic inhibitor. S1–S5 represents the respective subsite in the enzyme binding pocket at Casp2. (B) Substitution of AcVDVXD-CHO at P2 and AcVDXVD-CHO at P3 with various basic amino acids and interaction with Glu52 in the active site of Casp2; negative control with acetylated lysine.

An alternate design rationale arose out of consideration of the cleavage site of the tau protein (GSVQIVYKPVD|LSKVTSKCG–; D|L cleavage site), which creates Δtau314 associated with AD. Because Casp2 is the only caspase to catalyze this cleavage, it was anticipated that AcYKPVD-CHO (1) may be a selective inhibitor. Upon applying covdock protocols, however, AcYKPVD-CHO docking scores predict that binding to both Casp2 and Casp3 enzymes would be significantly weaker [−9.348 (Casp2), −9.616 (Casp3)] than binding predicted for the canonical inhibitor AcVDVAD-CHO (2) (Table 1). It was hoped that systematic exploration of hybrid pentapeptides that incorporate some features of 1 onto the potent scaffold of 2 could lead to more potent and selective inhibitors. AcYKPVD-CHO (1) and AcVDVAD-CHO (2) and hybrid molecule AcVDPVD-CHO (3) (Figure 4A) served as starting points in the design of a new pentapeptide series with which the gaps in structure–affinity relationships (SAR) between the three lead compounds can be explored. A series was developed that provides the motifs V_(P5)D_(P4), Y_(P5)K_(P4), V_(P5)K_(P4), or Y_(P5)D_(P4) for the amino acid positions P5–P4 and the motifs P_(P3)V_(P2)D_(P1), K_(P3)V_(P2)D_(P1), or V_(P3)K_(P2)D_(P1) for the positions P3–P1 (Figure 4B). To complete the series, individual changes were subsequently made at various positions. As with the previous series, all inhibitors contain an aldehyde (−CHO) that serves as an electrophilic warhead that reacts with the proteolytic cysteine (Figure 3A) in the active site in a reversible covalent mechanism.

Figure 4.

A) Caspase-2 lead compounds AcYKPVD-CHO (1), AcVDVAD-CHO (2), and AcVDPVD-CHO (3). (B) Substitution pattern of the first peptide series derived from the lead structures.

All synthesized peptides were pharmacologically characterized in vitro with respect to their Casp2 and Casp3 affinity (Table 2). Selected compounds were tested in a panel assay for selectivity within the broader caspase family (Table 3).

Table 3. Binding Data (pK_i Values) of Selected Pentapeptides (1, 2, 4–6, 16, 19, 23–25, and 35) at Casp1, Casp2, Casp3, Casp6, Casp7, and Casp9^a.

		pKi ± SD
sequence	Comp.	Casp1	N	Casp2b	N	Casp3b	N	Casp6	N	Casp7	N	Casp9	N
AcYKPVD-CHO	1	<4	2	4.56 ± 0.43	3	4.73 ± 0.10	4	<4	2	4.53 ± 0.19	2	<4.5	2
AcVDVAD-CHO	2	7.06 ± 0.06	2	7.85 ± 0.08	5	7.73 ± 0.09	5	<4	2	7.08 ± 0.06	2	5.57 ± 0.31	2
AcVDKVD-CHO	4	5.97 ± 0.12	2	7.40 ± 0.02	5	7.28 ± 0.09	4	<4	2	7.28 ± 0.10	2	4.77 ± 0.50	2
AcVDVKD-CHO	5	6.18 ± 0.20	2	7.63 ± 0.05	4	6.91 ± 0.07	4	<4	2	6.25 ± 0.07	2	<4.5	2
AcVDVVD-CHO	6	6.61 ± 0.13	2	7.77 ± 0.12	4	7.81 ± 0.13	4	4.93 ± 0.01	2	7.63 ± 0.07	2	5.52 ± 0.72	2
AcYDVKD-CHO	16	6.34 ± 0.01	2	7.83 ± 0.04	4	7.13 ± 0.02	4	<4	2	7.04 ± 0.01	2	4.72 ± 0.03	2
AcVD(Dab)VD-CHO	19	6.04 ± 0.02	2	7.44 ± 0.08	4	7.54 ± 0.02	4	<4	2	7.34 ± 0.11	2	5.17 ± 0.19	2
AcVDV(Orn)D-CHO	23	5.82 ± 0.04	2	7.45 ± 0.07	4	6.60 ± 0.08	4	<4	2	5.92 ± 0.01	2	<4.5	2
AcVDV(Dab)D-CHO	24	5.43 ± 0.16	2	7.26 ± 0.07	4	5.82 ± 0.18	4	<4	2	4.61 ± 0.36	2	<4	2
AcVDV(Dap)D-CHO	25	5.48 ± 0.08	2	6.77 ± 0.17	4	6.14 ± 0.09	4	<4	2	5.12 ± 0.09	2	<4	2
AcIDVKD-CHO	35	6.05 ± 0.03	2	7.68 ± 0.04	4	6.80 ± 0.10	4	<4	2	6.25 ± 0.11	2	<4.5	2

^aData shown are mean values ± SD of N independent experiments, each performed in duplicate or triplicate. Data were analyzed by nonlinear regression and were best fitted to sigmoidal concentration–response curves.

^bData at Casp2 and Casp3 are also shown in Table 2.

Chemistry

The aspartic acid-loaded amino-Merrifield resin was synthesized according to previously described protocols³²⁻³⁴ with small modifications in the sequence of the synthesis steps in order to improve the yield in our lab (for further information, see Supporting Information, Scheme S1). The pentapeptides were prepared by manual solid-phase peptide synthesis (SPPS) on an aspartic acid-loaded amino-Merrifield resin according to the standard Fmoc strategy. The synthesis, in general, can be subdivided into four steps (see Scheme 1). The Fmoc deprotection of the N-terminal amino acid of the resin follows the single coupling of an amino acid to the modified amino-Merrifield resin. The coupling of the amino acids was performed by using 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU)/N,N-diisopropylethylamine (DIPEA),²³ or Oxyma/N,N′-diisopropylcarbodiimide (DIC)³⁵ (Scheme 1). The pentapeptides were prepared by repeating the deprotection and coupling steps. Once the desired pentapeptide was prepared on resin, the N-terminus was acetylated by using acetic anhydride. Subsequently, the peptides were cleaved from the resin, and the sidechains of the amino acids were deprotected by treatment with trifluoroacetic acid (TFA) 90% in water. Structures of the pentapeptides 1 and 3–36 are shown in Figures S2–S35 in the Supporting Information. Chemical stability of pentapeptides 3, 15, 22, and 32 was analyzed at room temperature for 7 days and is shown in the Supporting Information (Figures S140–S143). In addition, for selected compounds (1–2, 4–5, 24, and 35), an in silico study of physicochemical, pharmacokinetic, and bioavailability parameters was performed using the SwissADME web tool.³⁶ As expected, the peptidic structures showed unfavorable properties in almost all areas (Table S3, Figure S157, Supporting Information).

Scheme 1. Preparation of the Pentapeptides on the Modified Amino-Merrifield Resin: (a) Piperidine 20% in DMF, 35 °C, 15 min; (b) Amino Acid (5 equiv), HATU/DIPEA (5/5 equiv) or Oxyma/DIC (5/5 equiv), DMF/NMP (8/2 v/v), 35 °C, 45 min (Single Coupling); (c) Acetic Anhydride (10 equiv), DIPEA (10 equiv), DMF, rt, 30 min; and (d) TFA (90% in Water), rt, 1 h.

Pharmacology

The results of the fluorescence-based enzyme inhibition assay using Casp1, Casp2, Casp3, Casp6, Casp7, and Casp9 are shown in Tables 2 and 3. The alternation of Val in position 5, coming from the canonical inhibitor AcVDVAD-CHO (2), with Tyr and Ile, resulted in a series of peptides that did not show any significant changes with respect to affinity. AcVDVKD-CHO (5), AcYDVKD-CHO (16), and AcIDVKD-CHO (35) have high inhibitory affinity at Casp2 (pK_i value of 7.63–7.83) with a slight tendency toward selectivity for Casp2 (selectivity ranging from a factor of 5.06–7.53) as opposed to Casp3. Furthermore, AcVDKVD (4) and AcYDKVD-CHO (15) showed high affinity at Casp2 (pK_i: 7.40 and 7.39/7.18). However, they have a lack of selectivity (factor of 1.32 and 0.44/0.27) in comparison to the abovementioned peptides (Table 2). This suggested that Val at position 5 is not crucial for our inhibitors and can be exchanged with Tyr and Ile, whereas the substitution of Asp at position 4 by Lys, coming from the tau cleavage sequence YKPVD, led to a loss of affinity at Casp2 and Casp3 for all our inhibitors. AcVKVVD-CHO (10), AcVKKVD-CHO (11), and AcVKVKD-CHO (12) did not show any affinity at Casp2 in comparison to AcVDVVD-CHO (6), AcVDKVD-CHO (4), and AcVDVKD-CHO (5) (Table 2). The same applies to AcYKKVD-CHO (8) and AcYKVKD-CHO (9) in contrast to AcYDKVD-CHO (15) and AcYDVKD-CHO (16) (Table 2). Moreover, the peptides AcYKVVD-CHO (14) and AcVKPVD-CHO (13) did not show any affinity for Casp2 (Table 2). The peptide AcYKPVD-CHO (1), derived from the optimal cleavage sequence of Casp2 for the tau protein, as well as the peptides AcVDPVD-CHO (3), AcVKPVD-CHO (13), AcYDPVD-CHO (17), AcVDPPD-CHO (33), and AcVDPKD-CHO (34) demonstrated that Pro at position 3 is also not tolerated at Casp2 and Casp3 (Table 2). For example, AcVDKPD-CHO (32) (Casp2 pK_i: 7.66/7.38; Casp3 pK_i: 7.70) has high affinity at Casp2 and Casp3, whereas AcVDPKD-CHO (34) (Casp2 pK_i: <4; Casp3 pK_i: <4) led to a great loss in affinity by switching Pro from position 2 to position 3.

To characterize the potential Glu52 interaction demonstrated by covalent docking, an additional peptide series based on AcVDKVD-CHO (4) and AcVDVKD-CHO (5) was investigated. Starting from 4 (Casp2 pK_i: 7.40; Casp3 pK_i: 7.28), the chain length of the basic amino acid Lys was shortened piecewise. This resulted in AcVD(Orn)VD-CHO (18) and AcVD(Dab)VD-CHO (19), which showed almost the same affinity as 4 for Casp2 (pK_i: 7.37/6.92 and 7.44) and Casp3 (pK_i: 7.52 and 7.54). In contrast, AcVD(Dap)VD-CHO (20) demonstrated a significantly lower affinity at Casp2 (pK_i: 6.53/6.18), whereas the affinity at Casp3 (pK_i: 7.38) remained the same. The exchange of Lys at position 3 by Arg resulted in AcVDRVD-CHO (21), which displayed similar affinity (Casp2 pK_i: 7.38, Casp3 pK_i: 7.21). Likewise, the acetylation of Lys in position 3, resulting in 22, demonstrated no shift in affinity (Casp2 pK_i: 7.70/7.41; Casp3 pK_i: 7.96). In addition, we demonstrated that the exchange of Val at P2 or P5 by Thr, Phe, Ala, and Ile of probe 4 is tolerated. AcTDKTD-CHO (28), AcVDKID-CHO (29), AcVDKAD-CHO (7), and AcVDKFD-CHO (30) showed high affinity at Casp2 and Casp3, without any notable trend toward selectivity (Table 2). The same variations of the chain length were performed with Lys at position 2 of compound 5. AcVDV(Orn)D-CHO (23) (Casp2 pK_i: 7.45; Casp3 pK_i: 6.60) and AcVDV(Dab)D-CHO (24) (Casp2 pK_i: 7.26; Casp3 pK_i: 5.82) showed comparable affinities at Casp2, while Casp3 affinity decreased with further chain length shortening. In contrast, AcVDV(Dap)D-CHO (25) showed a loss in affinity at Casp2 (same as for 20), while affinity at Casp3 slightly increased (Casp2 pK_i: 6.77; Casp3 pK_i: 6.14) in comparison to 24. Substitution of Lys by Arg resulted in 26, which showed nearly the same affinity as 5 at Casp2 and Casp3 (Table 2). Acetylation of the Lys sidechain of 5 resulted in AcVDV(AcK)D-CHO (27), showing the same affinity at Casp2 and slightly higher affinity for Casp3 (Casp2 pK_i: 7.53/7.30; Casp3 pK_i: 7.60). Because the two peptides with acetylated lysine sidechains (22 and 27) showed high affinity, both at Casp2 and at Casp3, the hypothesis of ionic interactions between Lys at position 2 or 3 and Glu52 in the subsite of Casp2 must be doubted. This was additionally confirmed by exchanging Lys with Phe, resulting in AcVDFVD-CHO (31) that displays no drastic loss in Casp2 affinity (pK_i: 7.00/6.50). Because of these negative controls, it must be assumed that the interactions in the S2 and S3 subsite of Casp2 are not mainly due to ionic interactions. The alternation from Lys at position 3 to position 2 goes hand in hand with a moderate gain in Casp2 selectivity. Compound 5 showed a selectivity ratio of 5.16, whereas 4 showed no selectivity (Table 2). This effect increased with shorter chains, whereas 23 had a selectivity factor of 7.07 compared to 18 (0.70/0.25). The maximum gain in selectivity was reached for 24 (27.7), having diamino butyric acid, at position P2. Further shortening led to a loss of selectivity, as demonstrated for 25 (4.30). Interestingly, the gain in selectivity was not achieved by an increase of Casp2 affinity, but by a steady decrease of Casp3 affinity. In molecular modeling studies of Casp2 and Casp3, the trends for Casp2 selectivity were relatively well confirmed for 24 (Table 1), whereas the docking scores for 4 and 21 were not entirely consistent with the in vitro data (Table 1). Selected peptides were subjected to a panel assay to investigate their affinity for Casp1, Casp6, Casp7, and Casp9 (Table 3). While AcYKPVD-CHO (1) showed no significant affinity for those enzymes, the canonical inhibitor AcVDVAD-CHO (2) was active at Casp1 (pK_i: 7.06), Casp7 (pK_i: 7.08), and Casp9 (pK_i: 5.57) (Table 3). The lead structures for the Glu52 engagement, AcVDKVD-CHO (4) and AcVDVKD-CHO (5), exhibited inhibitory affinity at Casp1 (pK_i: 5.97 and 6.18), Casp7 (pK_i: 7.28 and 6.25), and Casp9 (pK_i: 4.77 and <4.5), and no affinity was shown at Casp6 (pK_i: <4). AcVD(Dab)VD-CHO (19) showed almost the same affinities in the panel as the respective lead compound 4, whereas AcVDV(Orn)D-CHO (23), AcVDV(Dab)D-CHO (24), and AcVDV(Dap)D-CHO (25) demonstrated lower affinities at Casp1, Casp 6, Casp 7, and Casp 9 compared to lead compound 5 (Casp1 pK_i: 5.82, 5.43, and 5.48; Casp7 pK_i: 5.92, 4.61, and 5.12; Casp9 pK_i: <4.5, <4 and <4). Among all synthesized peptides, 24 showed the best selectivity profile for Casp2 (Tables 2 and 3), displaying a 27.7-fold selectivity over Casp3, 67.3-fold over Casp1, 444-fold over Casp7, and at least 1827-fold over Casp6 and Casp9. Further panel results are summarized in Table 3.

Crystallography

Eight novel complexes of peptidic covalent inhibitors with Casp3 have been crystallographically characterized in the course of this work. These are complexes with compounds 1, 3, 4, 18, 21, 22, 24, and 31. We sought also to obtain crystals with Casp2, but protein production limits have hampered these efforts. Three different unique crystal forms (space groups P2₁, P2₁2₁2₁, and P6₃) arise from identical crystallization conditions. Each form includes an activated heterotetramer in the crystallographic asymmetric unit. Diffraction quality varies within this collection of structures from 1.67 to 3.25 Å, although even in the worst cases, unequivocal confirmation exists of the covalent bond between the C-terminal aspartic acid of the peptides and the catalytic cysteine (Cys163). In most structures, clear electron density identifies the position of each bound peptide across its entire length (see omit maps included in Table S1). The only exceptions are complexes with the two proline-containing peptides 1 (AcYKPVD-CHO) and 3 (AcVDPVD-CHO). For compound 1, no density is present to permit positioning of any of the peptide upstream of the P3 proline. From the absence of density, we can infer that the peptide has no preferred conformation for the AcYK moieties. Some density exists to allow for positioning of the P3 proline and P4 aspartic acid in the complex with 3, but the P4 aspartic acid sidechain and the P5 valine backbone have exchanged places, presumably necessitated by conformational limitations imposed by the P3 proline. There is no density for the N-terminal acetate. The absence of order is not entirely unexpected, given the very low affinity (25–52 μM) of the peptides with the P3 proline. The compound 22 (AcVD(AcK)VD-CHO) backbone is well ordered across its full length, but there is no observable density for the acetyl group at the end of the lysine sidechain (Table S1b). This group is essentially dangling into solvent and not held in any specific position.

The complexes with peptides without proline are all quite similar, despite variations in the P3 or P2 amino acid. All complexes conserve the constellation of H-bonds around the P1 aspartic acid involving Arg64, Arg207, and Gln161, and the anti-parallel H-bonds between the peptide backbone at P3 and the Arg207 mainchain are exactly as seen in the complex with AcVDVAD-CHO (2, cf. Figure 2A). As these peptides all also preserve the P4 aspartic acid, H-bonds joining the P4 Asp to the backbone of the L4 loop at Phe250 are also strictly conserved. None of these interactions involve sidechains varied in the different inhibitors. The only significant difference in the complexes is a shift in a portion of the L4 loop of Casp3, specifically residues 251–255 (Figure 5). The effect of this shift is to widen the cavity between the L3 and L4 loops where the S5 subsite resides. The shift was first observed in the complex with AcVDPVD-CHO (3), where it can be attributed to the absence of the H-bonds from P4 Asp to Phe250 and the need for a broader S5 pocket to accommodate the P5 valine in its place. However, the L4 shift is actually most pronounced in the AcVDRVD-CHO (21) and AcVD(AcK)VD-CHO (22) complexes, where Phe252 has moved 1.6 Å from its position in the AcVDVAD-CHO complex. The hydrophobic contacts between the P5 valine and Phe250 and Phe252 are lost with the movement, but hydrogen bonds to both Phe250 and Asn208 by the P4 aspartic acid of L3 are preserved (and even improved), as the P5 valine moves 1.6 Å in the opposite direction to more closely contact L3, which does not move. Two conclusions may be drawn from the examination of these complexes: (1) single amino acid substitution in the peptide inhibitors can induce changes in binding unrelated to direct contacts between the amino acid and the subsite in which it resides and (2) the protein itself is at least as flexible and responsive to ligand binding as the ligands. This should serve as a cautionary tale when attempting to dock ligands into these proteins using protocols that largely assume the protein conformation is fixed.

Figure 5.

L4 loop movements in Casp3-inhibitor complexes. Three complexes are shown to represent the magnitude of the shift that occurs in L4 upon ligand binding (AcVDFVD-CHO, blue; AcVDRVD-CHO, salmon; AcVDPVD-CHO, gray). Inhibitors differ only in the P3 amino acid; although the P3 alpha-carbon position is essentially unchanged in these complexes, the L4 loop shifts significantly. Changes are also observed in P4–P5 of ligands, as the gap between L3 and L4 loops widens; ligands seem to prefer to associate more closely with L3, as hydrophobic contacts with L4 are lost. None of the structures resulting during this study show significant movement in Phe256.

The complex with compound AcVDV(Dab)D-CHO (24) is interesting, as it reveals a structural basis for the selectivity of this compound for Casp2 over Casp3. Diffraction data for this complex are the poorest of all those determined in this work (3.25 Å), but it is nevertheless sufficient to show that the inhibitor binds with the P1 Asp anchored as expected, and the ligand backbone positioned as in all other complexes to make H-bonds along the entire length toward the valine at P5 (Figure 6). In order to provide for a reasonable fit to the electron density, however, the sidechain of the P2 Dab is modeled in an unrealistically high-energy conformation, with χ₁ = 22°. This model likely represents an average of multiple strained conformational states. There really is no energetically favorable conformation for the Dab sidechain, given the boundaries of the S2 pocket imposed by the sidechains of Phe256 and Tyr204, which appear not to have moved from the usual positions. This strain is not expected to be present in a complex of this same inhibitor with Casp2, due to the differences noted earlier regarding this subsite (Figure 2B), so the binding affinity of this compound for Casp2 should be better than that for Casp3 (as observed).

Figure 6.

Casp3/AcVDV(Dab)D-CHO (24) complex. Backbone interactions conserved throughout the peptide complexes anchor the ligand in the expected position, but sidechains (Phe256 and Tyr204) lining the P2 subsite limit extension of the Dab, forcing it to adopt a strained conformation. Solvent-accessible surfaces of both Casp3 (gray) and the ligand are shown to illustrate the tight packing that exists against the ceiling of the P2 subsite (dashed red line).

Site-Directed Mutageneses of Tau Protein—The Role of P5–P2 Residues in Caspase-2-Mediated Tau Cleavage

To better understand the contributions of amino acid residues at P5, P4, P3, and P2 to caspase-2-catalyzed cleavage of tau, one of its naturally occurring substrates, we engineered and expressed a series of recombinant tau mutants in addition to the wild-type form. Results from the time-dependent in vitro cleavage assay showed that while a K-to-D mutation at P4 or a P-to-V mutation at P3 did not alter tau cleavage, combining these two mutations led to a 25-fold increase in the cleavage product after a 4 h reaction. Meanwhile, a triple mutation of tau at P5–P3 (i.e., YKP-to-VDV) resulted in a 6-fold increase in the cleavage product after a 4 h reaction (Figure 7). These findings are consistent with the results of the in vitro characterization of peptidic inhibitors (cf. 1, 2, 6, 14, 17, and 36) observed in the pharmacological experiments.

Figure 7.

Contributions of P5–P2 residues to in vitro Casp2-catalyzed tau cleavage. (A) Representative Western blots (tau-13) showing in vitro Casp2-catalyzed cleavage of a variety of tau mutants in a time-dependent manner. The emergence and yield of the ∼35 kDa cleavage product (ending at D314, arrow) differed among investigated tau mutants. (B) Quantification levels of the cleavage product were normalized to levels of full-length tau [asterisk in (A)] at T = 0; experiments were performed in duplicates; means (open symbols) and standard deviations (SDs, error bars) are shown; repeated measures ANOVA was performed to compare effects of tau mutants (F(5, 6) = 279.60, P < 0.0001), followed by Tukey’s post hoc test (***, p < 0.001, tau Y310V K311D P312V V313A vs tau WT, tau K311D, or tau P213V; #, p = 0.05–0.06, tau K311D P312V vs tau WT, tau K311D, or tau P213V).

Conclusions

In this study, we aimed to elucidate the caspase-2 binding site by a series of new pentapeptides and the resulting SAR. In particular, the differences with respect to caspase-3 should be investigated to get closer to the goal of highly potent and selective caspase-2 inhibitors. For this purpose, 35 pentapeptides were synthesized and tested for their inhibitory affinity at Casp2 and Casp3. Selected inhibitors 1, 2, 4, 5, 6, 16, 19, 23, 24, 25, and 35 were subjected to a panel assay screening to investigate affinities at Casp1, Casp6, Casp7, and Casp9 in order to get more information about the selectivity within the caspase family. Our ligands are structurally derived from the canonical inhibitor AcVDVAD-CHO and the tau cleavage sequence YKPVD. We observed that by the introduction of lysine at P2 [AcVDVKD-CHO (5), Casp2 pK_i: 7.63, Casp3 pK_i: 6.91], we gained slight selectivity over Casp3 without losing affinity at Casp2. The selectivity advantage was increased by further shortening the chain length. The maximum selectivity could be achieved by introducing Dab at position 2 (AcVDV(Dab)D-CHO (24); Casp2 pK_i: 7.26; Casp3 pK_i: 5.82; Casp3/Casp2 selectivity 27.7). Negative control with acetylated lysine AcVDV(AcK)D-CHO (27) has shown that the gain in selectivity is not caused by ionic interactions between lysine and glutamine (Glu52) in the binding site of Casp2, as initially suspected, and is achieved by a decrease in Casp3 affinity rather than an increase in Casp2 affinity. Molecular docking studies of 1, 2, 4, 21, 24, 37, and 38 were performed to refine our understanding of the molecular interactions in the binding site of Casp2 and Casp3. In addition, co-crystallizations of 1, 3, 4, 18, 21, 22, 24, and 31 with Casp3 were made, allowing us to target the differences between Casp2 and Casp3 to further develop selective inhibitors. Furthermore, an in vitro cleavage assay with a series of recombinant tau mutants [e.g., K311D P312V (YDVVD³¹⁴)] was performed and showed results that are in agreement with the results of the in vitro Casp2 binding data. The inhibitor AcYKPVD-CHO (1) deriving from the optimal tau cleavage sequence has not shown to be an active Casp2 inhibitor. The same applies to all other inhibitors containing proline at position 3 (e.g., AcVDPKD-CHO, 34) or lysine at position 4 (e.g., AcVKVKD-CHO, 12). For that reason, other natural cleavage sequences should be investigated for their potential to inhibit the enzyme activity. These natural cleavage sequences of Casp2 still provide a large pool of potential selective inhibitors. In summary, the results of this study provide a reasonable basis for the development of further selective Casp2 inhibitors.

Experimental Section

General

Unless otherwise listed, chemicals and solvents were purchased from commercial suppliers and used as received. Canonical inhibitor AcVDVAD-CHO (2), AcLDESD-CHO (37), and AcDEVD-CHO (38) were purchased from Bachem (Torrance/CA, USA). All the solvents were of analytical grade or distilled prior to use. Dimethylformamide (DMF) and N-methyl-2-pyrrolidone (NMP) were purchased from Iris Biotech (Marktredwitz, Germany). Sodium borhydride, oxalyl chloride, methanol, dichloromethane, diethyl ether, toluene, ethyl acetate, tetrahydrofuran, and hydrochloric acid 37% were obtained from Fisher Scientific/Acros Organics (Schwerte, Germany). If lower concentrations of hydrochloric acid were required, these were diluted accordingly. Isobutyl chloroformate, tert-butylcarbazat, p-toluenesulfonic acid, trans-4-(aminomethyl)cyclohexanecarboxylic acid, DIC, Oxyma and N-methylmorpholine were purchased from TCI (Eschborn, Germany). 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI), 1-hydroxybenzotriazole hydrate (HOBt), HATU, DIPEA, and TFA were obtained from ABCR (Karlsruhe, Germany). 1,1′-Carbonyldiimidazole was purchased from Fluorochem (Derbyshire, United Kingdom). Aminomethylated polystyrene HL (100–200 mesh), acetic anhydride, and triethylamine were purchased from Merck (Darmstadt, Germany). Benzyl alcohol, dimethylsulfoxide, acetic acid, and acetonitrile for high-performance liquid chromatography (HPLC) were obtained from Sigma-Aldrich (Taufkirchen, Germany). Protected amino acids Fmoc-Tyr(tBu)-OH, Fmoc-Arg(Pbf)-OH, Fmoc-Lys(Boc)-OH, Fmoc-Phe-OH, Fmoc-Pro-OH, Fmoc-Val-OH, Fmoc-Asp(OtBu)-OH, Fmoc-Ala-OH, Fmoc-Orn(Boc)-OH, Fmoc-Dab(Boc)-OH, Fmoc-Dap(Boc)-OH, Fmoc-Lys(Ac)-OH, Fmoc-Thr(tBU)-OH, and Fmoc-Ile-OH were procured from Carbolution Chemicals (St. Ingbert, Germany). Deuterated solvents for nuclear magnetic resonance (NMR) spectroscopy were purchased from Deutero (Kastellaun, Germany). The orbital shaker (Multi Reax) was from Heidolph (Schwabach, Germany). The frits had a pore size of 35 μm and were procured from Roland Vetter Laborbedarf (Ammerbuch, Germany). The infrared lamp was obtained from Medisana (Neuss, Germany), and the thermostat was obtained from PEARL.GmbH (Buggingen, Germany). The syringes used (Injekt Luer Solo) were obtained from Braun (Melsungen, Germany). For the preparation of stock solutions, buffers, and HPLC eluents, millipore water was used. NMR spectra (¹H NMR, ¹³C NMR, ¹⁹F NMR, DEPT, ¹H COSY, HSQC, HMBC) were recorded on a Bruker AVANCE-300 (7.05 T, ¹H: 300 MHz, ¹³C: 75.5 MHz, ¹⁹F: 188) or AVANCE-400 (9.40 T, ¹H: 400 MHz, ¹³C: 100.6 MHz, ¹⁹F: 282) NMR spectrometer (Bruker, Karlsruhe, Germany). All chemical shifts are reported in the δ-scale as parts per million (ppm) relative to the solvent’s residual peaks as the internal standard. Moreover, the multiplicity, coupling constant (J), and number of protons are stated. Multiplicities are specified with the following abbreviations: s (singlet), d (doublet), t (triplet), q (quartet), quint (quintet), m (multiplet), as well as combinations thereof. High-resolution mass spectrometry (HRMS) was performed on a Q-TOF 6540 ultrahigh definition (UHD) LC/MS system (Agilent Technologies) using an electrospray ionization (ESI) source or on an AccuTOF GCX GC/MS system (Jeol, Peabody, MA) using an electron ionization source. Preparative HPLC was performed with a system from Waters (Eschborn, Germany) consisting of a Waters 2545 binary gradient module, a Waters 2489 UV/vis-detector, and a Waters Fraction Collector 3, and the column was a YMC Triart C18 (150 × 10 mm, 5 μm) (YMC, Dinslaken, Germany) at a flow rate of 20 mL/min or a HPLC from Knauer (Berlin, Germany) consisting of two K-1800 pumps and a K-2001 detector, and the column was a Phenomenex Gemini (250 × 21 mm, 5 μm) (Phenomenex, Aschaffenburg, Germany) at a flow rate of 15 mL/min. As a mobile phase, mixtures of MeCN and 0.1% aqueous (aq.) TFA were used. The UV detection was carried out at 220 nm. For sample preparation, all compounds were dissolved in a mixture of water/acetonitrile (95/5 v/v) and filtered with PTFE filters (25 mm, 0.2 μm) (Phenomenex, Aschaffenburg, Germany). The purified peptides were lyophilized using a CHRIST Alpha 2–4 LD freeze dryer (Osterode am Harz, Germany) equipped with a RZ 6 rotary vane vacuum pump (Vacuubrand, Wertheim, Germany). Analytical purity control was performed on a 1100 HPLC system from Agilent Technologies equipped with an Instant Pilot controller, a G1312A binary pump, a G1329A ALS autosampler, a G1379A vacuum degasser, a G1316A column compartment, and a G1315B diode array detector. The column was a Phenomenex Kinetex XB-C18 column (250 × 4.6 mm 2, 5 μm) (Phenomenex, Aschaffenburg, Germany) or a Phenomenex Gemini NX-C18 column (250 × 4.6 mm, 5 μm). The oven temperature during HPLC analysis was 30 °C. As a mobile phase, mixtures of MeCN/aqueous TFA were used. Absorbance was detected at 220 nm. The injection volume was 20–80 μL at compound concentrations of 1 mM. The following linear gradient was applied: MeCN/TFA (0.05%) (v/v) 0 min: 10:90, 25 min: 95:5, 35 min: 95:5; flow rate: 1.0 mL/min, t₀(Kinetix XB-C18) = 2.75 min, t₀(Gemini NX-C18) = 2.99 min (t₀ = dead time). Retention (capacity) factors k were calculated from the retention times t_R according to k = (t_R – t₀)/t₀. The purities of the compounds were calculated by the percentage peak area of the chromatograms. The tested compounds have been screened for PAINS and aggregation by publicly available filters (http://zinc15.docking.org/patterns/home, http://advisor.docking.org).^37,38 Compounds have not been previously reported as PAINS or aggregators. None of the data showed abnormalities, for example, high Hill slopes, what could be a hint for PAINS.³⁸

Computational Chemistry

Molecular Modeling

Covalent docking studies were performed using the “covalent docking (CovDock)” module within the Schrödinger small-molecule drug discovery software suite (Schrödinger, LLC, New York, NY, Version 2021.1 unless otherwise noted).^30,39−41 The crystal structures of Casp2 (PDBid: 1pyo) and Casp3 (PDBid: 3edq) were prepared using the module “protein preparation wizard” in Maestro with the default protein parameters. Hydrogen atoms were added, and water molecules that were beyond 5 Å from heterocyclic groups were deleted. The covalently bound ligands AcLDESD-CHO: 1pyo and AcLDESD-CHO: 3edq where D-CHO represents “aspartic acid aldehyde” were included in the protein structures during their preparation for the covalent docking. In the experiments below, the protein structures prepared in this fashion are denoted by PP/PDBid, for example, PP:1pyo. Hydrogen bonds were optimized, the partial charges were assigned, and the protein structure was energy-minimized using the OPLS3e force field. Following this preparation, the covalent bond connecting the ligand to the protein was broken, and the now-separated aldehyde (reactive functional group) and cysteine (nucleophilic reaction group) were reconstituted by adjusting bond orders, adding hydrogens, and minimizing these groups in place. This free ligand (the “workspace ligand”) was employed to create the covalent docking grid used in the covalent docking and scoring (vide infra). The individual target receptors were set up using the following reactive cysteine residues (A: 155, 1pyo; A: 163, 3edq). The “reaction type” SMARTS string {[H]C=O} was built as a customized nucleophilic addition to a double bond. The “box center” for the docking grid was set using the “centroid of (the) workspace ligand”. Docking was performed in the “pose prediction (thorough)” mode. A “minimization radius” of 3.0 Å was used, “perform MM-GBSA scoring” was selected, and three (3) “output poses per ligand reaction site” were selected (only the lowest energy pose is reported). Ligands for covalent docking experiments were drawn in ChemDraw, imported into Maestro as sdf, and refined into 3D structures using the “ligand preparation” module and its default parameters. These 3D structures, with the appropriate tautomers and charges, were directly used in “covalent docking” experiments and are designated as PL/ligand name, for example, PL: AcLDESD-CHO. “cdock affinity” is reported in kcal/mol.^30,31

Compound Characterization

The pentapeptides were characterized using the following methods: HRMS, ¹H NMR spectroscopy (for spectra, see Supporting Information), and ¹³C NMR spectroscopy (for spectra, see Supporting Information). Additionally, two-dimensional (2D) NMR spectra like COSY, HSQC (for spectra, see Supporting Information), and HMBC were made of the pentapeptides. For purity and stability control, HPLC (RP-HPLC) analysis was performed (for chromatogram, see Supporting Information) with a minimum purity standard of ≥95%.

Synthesis and Analytical Data

General Procedure for the Synthesis of the Pentapeptides

The aspartic acid-loaded semicarbazide amino-Merrifield resin (45) (300 mg, 0.188 mmol, 1 equiv) was weighed into a fritted 10 mL syringe. Subsequently, 5 mL of a mixture of piperidine 20% in DMF was drawn up to remove the N-terminal Fmoc-protecting group. The syringe was shaken on an orbital shaker at 35 °C for 15 min. The orbital shaker was covered with a box, which was insulated from the inside with the aluminum foil. An infrared lamp was placed on an aperture on top. In order to keep the temperature constant at 35 °C, the lamp was controlled by a thermostat. The liquid was then removed with the aid of a vacuum flask, and the residual resin was washed with DMF (3 × 8 mL). The coupling of the amino acids to the N-terminus was performed by two different methods. The corresponding amino acid (5 equiv) and HATU (357 mg, 0.94 mmol, 5 equiv) (method A) or Oxyma (134 mg, 0.94 mmol, 5 equiv) (method B) were weighed in two separate Erlenmeyer flasks. Subsequently, both were dissolved in 3–4 mL of a mixture of DMF/NMP (8/2 v/v). Then, DIPEA (164 μL, 0.94 mmol, 5 equiv) (method A) or DIC (151 μL, 0.94 mmol, 5 equiv) (method B) was added to the solution of HATU (method A)/Oxyma (method B). Subsequently, both solutions were drawn up with the resin-loaded syringe and shaken at 35 °C for 45 min. The liquid was then removed with the aid of a vacuum flask, and the residual resin was washed with DMF (3 × 8 mL).

These two steps (coupling and deprotection) were repeated until the desired pentapeptide was built up. Then, the N-terminal Fmoc-protecting group was removed using piperidine in DMF (20%). Subsequently, the N-terminus of the pentapeptide was acetylated by dissolving acetic anhydride (178 μL, 1.88 mmol, 10 equiv) and DIPEA (328 μL, 1.88 mmol, 10 equiv) in 6–8 mL DMF. The solution was drawn up with the syringe and shaken for 30 min at room temperature. After completion, the liquid was removed, and the resin was washed with DMF (2 × 8 mL), methanol (2 × 8 mL), dichloromethane (2 × 8 mL), and finally with diethyl ether (2 × 8 mL). The pentapeptide was cleaved off the resin, and the sidechains were deprotected by drawing up 6 mL of 90% TFA in water. The syringe was shaken for 1 h at room temperature. The liquid was then poured into a round-bottomed flask. The step was repeated again, and then, the cleavage cocktail was diluted with 50 mL of water and freeze-dried. The crude product was purified by HPLC, yielding the desired pentapeptides.

(S)-3-((S)-2-((S)-1-(Acetyl-l-tyrosyl-l-lysyl)pyrrolidine-2-carboxamido)-3-methylbutanamido)-4-oxobutanoic Acid Hydrotrifluoroacetate (1)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (15.2 mg, 3%): RP-HPLC: 99%, (t_R = 6.83, k = 1.28). ¹H NMR (400 MHz, D₂O): δ 6.99 (d, J = 8.0 Hz, 2H), 6.70 (d, J = 8.5 Hz, 2H), 4.92 (dd, J = 4.6, 3.2 Hz, 1H), 4.42 (t, J = 6.5 Hz, 1H), 4.35 (t, J = 7.9 Hz, 1H), 4.26–4.20 (m, 1H), 4.20–4.09 (m, 1H), 3.92 (dd, J = 7.6, 4.7 Hz, 1H), 3.52–3.40 (m, 2H), 2.93–2.82 (m, 3H), 2.81–2.64 (m, 2H), 2.53–2.35 (m, 1H), 2.27–2.11 (m, 1H), 1.95–1.74 (m, 7H), 1.66–1.42 (m, 4H), 1.29–1.20 (m, 2H), 0.91–0.77 (m, 6H). ¹³C NMR (101 MHz, D₂O): δ 173.96, 173.93, 172.96, 172.89, 170.73, 170.72, 154.45, 130.46, 127.86, 115.36, 89.68, 60.16, 59.81, 55.52, 51.54, 50.68, 47.92, 46.72, 39.22, 36.17, 34.54, 30.48, 30.23, 29.38, 27.45, 26.36, 24.55, 21.68, 21.56. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₁H₄₇N₆O₉⁺, 647.3399; found, 647.3413: C₃₁H₄₆N₆O₉ × C₂HF₃O₂ (760.77).

(S)-3-((S)-2-Acetamido-3-methylbutanamido)-4-((S)-2-(((S)-1-(((S)-1-carboxy-3-oxopropan-2-yl)amino)-3-methyl-1-oxobutan-2-yl)carbamoyl)pyrrolidin-1-yl)-4-oxobutanoic Acid (3)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (17.4 mg, 16%): RP-HPLC: 97%, (t_R = 7.27, k = 1.42). ¹H NMR (400 MHz, D₂O): δ 4.94–4.85 (m, 2H), 4.37–4.28 (m, 1H), 4.21–4.08 (m, 1H), 4.02–3.84 (m, 2H), 3.76–3.58 (m, 2H), 2.90–2.78 (m, 1H), 2.73–2.53 (m, 2H), 2.56–2.37 (m, 1H), 2.21–2.04 (m, 1H), 2.00–1.73 (m, 8H), 0.97–0.71 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.27, 175.16, 174.37, 173.91, 173.44, 173.11, 173.03, 170.31, 170.24, 89.71, 60.60, 60.54, 59.87, 59.81, 59.56, 51.52, 51.47, 48.28, 48.06, 35.23, 33.94, 33.78, 29.98, 29.41, 24.55, 21.62, 18.33, 17.57. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₅H₄₀N₅O₁₀⁺, 570.2770; found, 570.2776: C₂₅H₃₉N₅O₁₀ (569.61).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (4)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (45.1 mg, 20%): RP-HPLC: 99%, (t_R = 6.53, k = 1.18). ¹H NMR (400 MHz, D₂O): δ 4.91 (dd, J = 6.0, 4.7 Hz, 1H), 4.62–4.58 (m, 1H), 4.26–4.08 (m, 2H), 3.99–3.89 (m, 2H), 2.92–2.79 (m, 3H), 2.76–2.63 (m, 2H), 2.51–2.35 (m, 1H), 2.00–1.86 (m, 5H), 1.77–1.49 (m, 4H), 1.39–1.16 (m, 2H), 0.80 (dt, J = 9.5, 5.4 Hz, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.23, 175.13, 174.62, 173.95, 173.72, 173.59, 173.44, 173.36, 173.30, 172.98, 172.87, 172.02, 171.96, 89.70, 59.95, 59.92, 59.76, 53.56, 53.50, 51.58, 51.50, 50.07, 39.19, 35.14, 34.08, 33.91, 30.38, 30.15, 30.10, 29.84, 26.24, 21.93, 21.68, 18.30, 17.61. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₅N₆O₁₀⁺, 601.3192; found, 601.3201: C₂₆H₄₄N₆O₁₀ × C₂HF₃O₂ (714.69).

(4S,7S,10S,13S,16S)-13-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (5)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (39.9 mg, 30%): RP-HPLC: 98%, (t_R = 6.29, k = 1.10). ¹H NMR (400 MHz, D₂O): δ 4.94–4.90 (m, 1H), 4.65–4.61 (m, 1H), 4.31–4.02 (m, 2H), 4.00–3.89 (m, 2H), 2.91–2.79 (m, 3H), 2.74–2.56 (m, 2H), 2.51–2.34 (m, 1H), 2.02–1.88 (m, 5H), 1.77–1.49 (m, 4H), 1.39–1.22 (m, 2H), 0.87–0.76 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.31, 175.13, 174.58, 174.19, 173.65, 173.29, 173.22, 173.14, 173.12, 172.20, 172.11, 89.70, 59.87, 59.65, 53.59, 51.50, 50.06, 39.22, 35.24, 30.45, 29.90, 26.20, 22.01, 21.65, 18.46, 18.40, 18.33, 17.70, 17.62. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₅N₆O₁₀⁺, 601.3192; found, 601.3198: C₂₆H₄₄N₆O₁₀ × C₂HF₃O₂ (714.69).

(4S,7S,10S,13S,16S)-7-(Carboxymethyl)-16-formyl-4,10,13-triisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid (6)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (32.5 mg, 18%): RP-HPLC: 99%, (t_R = 8.88, k = 1.96). ¹H NMR (400 MHz, MeOD): δ 4.78–4.67 (m, 1H), 4.61–4.51 (m, 1H), 4.35–4.24 (m, 1H), 4.23–4.07 (m, 3H), 2.97–2.83 (m, 1H), 2.83–2.59 (m, 2H), 2.57–2.42 (m, 1H), 2.20–2.02 (m, 3H), 2.00 (s, 3H), 1.06–0.85 (m, 18H). ¹³C NMR (101 MHz, MeOD): δ 173.67, 172.72, 172.41, 172.26, 172.05, 171.82, 170.98, 97.00, 59.45, 59.29, 59.18, 50.56, 47.88, 34.92, 33.03, 30.35, 30.24, 30.19, 21.02, 18.31, 17.40, 17.26. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₅H₄₂N₅O₁₀⁺, 572.2926; found, 572.2933: C₂₅H₄₁N₅O₁₀ (571.63).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4-isopropyl-13-methyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (7)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (45.1 mg, 35%): RP-HPLC: >99%, (t_R = 4.81, k = 0.60). ¹H NMR (300 MHz, D₂O): δ 4.89 (dd, J = 4.5, 3.1 Hz, 1H), 4.60–4.54 (m, 1H), 4.19–4.00 (m, 3H), 3.88 (dd, J = 7.1, 2.3 Hz, 1H), 2.91–2.78 (m, 3H), 2.75–2.32 (m, 3H), 1.97–1.82 (m, 4H), 1.77–1.45 (m, 4H), 1.34–1.17 (m, 5H), 0.82–0.75 (m, 6H). ¹³C NMR (75 MHz, D₂O): δ 175.25, 175.16, 174.64, 174.09, 173.74, 173.23, 173.17, 172.31, 172.19, 89.70, 59.94, 53.65, 51.46, 50.04, 49.86, 39.15, 35.08, 34.01, 30.33, 29.84, 26.22, 21.94, 21.65, 18.27, 17.62, 16.61. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₄H₄₁N₆O₁₀⁺, 573.2879; found, 573.2886: C₂₄H₄₀N₆O₁₀ × C₂HF₃O₂ (686.64).

(4S,7S,10S,13S,16S)-7,10-Bis(4-aminobutyl)-16-formyl-4-(4-hydroxybenzyl)-13-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Dihydrotrifluoroacetate (8)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (75.2 mg, 44%): RP-HPLC: 99%, (t_R = 5.42, k = 0.97). ¹H NMR (300 MHz, D₂O): δ 7.00–6.91 (m, 2H), 6.71–6.60 (m, 2H), 4.87 (dd, J = 6.8, 4.7 Hz, 1H), 4.31 (t, J = 7.6 Hz, 1H), 4.19–3.86 (m, 4H), 2.90–2.75 (m, 6H), 2.69–2.30 (m, 2H), 1.95–1.77 (m, 4H), 1.73–1.37 (m, 8H), 1.34–1.07 (m, 4H), 0.83–0.71 (m, 6H). ¹³C NMR (75 MHz, D₂O): δ 175.33, 175.25, 174.07, 173.67, 173.49, 173.41, 173.35, 173.25, 173.19, 172.87, 172.74, 154.46, 130.47, 127.92, 115.44, 89.69, 59.61, 59.56, 55.51, 53.67, 53.61, 52.99, 51.59, 51.49, 39.18, 36.15, 34.25, 34.01, 30.62, 30.31, 26.35, 26.27, 22.06, 21.90, 21.60, 18.31, 17.70, 17.62. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₂H₅₂N₇O₉⁺, 678.3821; found, 678.3824: C₃₂H₅₁N₇O₉ × C₄H₂F₆O₄ (905.85).

(4S,7S,10S,13S,16S)-7,13-Bis(4-aminobutyl)-16-formyl-4-(4-hydroxybenzyl)-10-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Dihydrotrifluoroacetate (9)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (66.3 mg, 39%): RP-HPLC: 99%, (t_R = 5.48, k = 0.83). ¹H NMR (300 MHz, D₂O): δ 6.99–6.92 (m, 2H), 6.70–6.63 (m, 2H), 4.91–4.86 (m, 1H), 4.33 (t, J = 7.5 Hz, 1H), 4.18–4.03 (m, 3H), 3.90–3.81 (m, 1H), 2.86–2.73 (m, 6H), 2.68–2.31 (m, 2H), 1.94–1.77 (m, 4H), 1.69–1.39 (m, 8H), 1.33–1.10 (m, 4H), 0.79 (t, J = 6.1 Hz, 6H). ¹³C NMR (75 MHz, D₂O): δ 175.26, 175.06, 174.04, 173.69, 173.18, 173.07, 163.21, 162.74, 154.44, 130.46, 127.90, 115.47, 89.64, 59.65, 55.48, 53.66, 53.05, 51.49, 39.17, 36.15, 34.23, 33.89, 30.51, 30.06, 26.30, 26.24, 21.94, 21.60, 18.40, 18.34, 18.00, 17.94. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₂H₅₂N₇O₉⁺, 678.3821; found, 678.3831: C₃₂H₅₁N₇O₉ × C₄H₂F₆O₄ (905.85).

(4S,7S,10S,13S,16S)-7-(4-Aminobutyl)-16-formyl-4,10,13-triisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (10)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (54.9 mg, 42%): RP-HPLC: >99%, (t_R = 7.23, k = 1.41). ¹H NMR (300 MHz, D₂O): δ 4.87 (dd, J = 9.0, 4.7 Hz, 1H), 4.22 (t, J = 7.4 Hz, 1H), 4.18–4.02 (m, 1H), 3.98–3.84 (m, 3H), 2.82 (t, J = 7.6 Hz, 2H), 2.69–2.29 (m, 2H), 2.00–1.75 (m, 6H), 1.72–1.44 (m, 4H), 1.39–1.10 (m, 2H), 0.86–0.70 (m, 18H). ¹³C NMR (75 MHz, D₂O): δ 175.62, 175.47, 174.40, 173.74, 173.45, 173.31, 173.16, 173.05, 172.76, 172.64, 89.68, 59.74, 59.67, 59.57, 53.22, 51.56, 39.19, 34.19, 30.40, 30.25, 30.15, 29.96, 26.23, 22.05, 21.57, 18.38, 18.02, 17.97, 17.80, 17.74. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₇H₄₉N₆O₈⁺, 585.3606; found, 585.3617: C₂₇H₄₈N₆O₈ × C₂HF₃O₂ (698.74).

(4S,7S,10S,13S,16S)-7,10-Bis(4-aminobutyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Dihydrotrifluoroacetate (11)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (76.0 mg, 48%): RP-HPLC: >99%, (t_R = 5.23, k = 0.90). ¹H NMR (300 MHz, D₂O): δ 4.88 (dd, J = 6.0, 4.7 Hz, 1H), 4.25–4.04 (m, 3H), 3.97–3.85 (m, 2H), 2.84 (t, J = 7.6 Hz, 4H), 2.71–2.32 (m, 2H), 1.98–1.78 (m, 5H), 1.72–1.45 (m, 8H), 1.41–1.14 (m, 4H), 0.84–0.72 (m, 12H). ¹³C NMR (75 MHz, D₂O): δ 175.25, 175.15, 174.43, 173.83, 173.48, 173.36, 172.89, 172.76, 89.68, 59.79, 59.64, 59.59, 53.49, 53.43, 53.27, 51.58, 51.47, 39.18, 33.92, 30.45, 30.36, 30.27, 29.93, 26.25, 22.06, 22.02, 21.60, 18.36, 17.79, 17.70, 17.62. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₈H₅₂N₇O₈⁺, 614.3872; found, 614.3871: C₂₈H₅₁N₇O₈ × C₄H₂F₆O₄ (841.80).

(4S,7S,10S,13S,16S)-7,13-Bis(4-aminobutyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Dihydrotrifluoroacetate (12)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (91.0 mg, 58%): RP-HPLC: >99%, (t_R = 5.02, k = 0.67). ¹H NMR (400 MHz, D₂O): δ 4.95–4.91 (m, 1H), 4.34–4.07 (m, 3H), 4.02–3.86 (m, 2H), 2.88 (t, 4H), 2.74–2.36 (m, 2H), 2.00–1.84 (m, 5H), 1.78–1.50 (m, 8H), 1.40–1.19 (m, 4H), 0.87–0.77 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.15, 174.47, 173.93, 173.88, 173.73, 173.70, 173.10, 173.04, 89.70, 59.87, 59.54, 53.71, 53.29, 51.53, 39.23, 39.19, 34.29, 30.31, 30.21, 29.92, 26.32, 26.20, 22.11, 22.02, 21.60, 18.40, 17.81. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₈H₅₂N₇O₈⁺, 614.3872; found, 614.3877: C₂₈H₅₁N₇O₈ × C₄H₂F₆O₄ (841.80).

(S)-3-((S)-2-((S)-1-(Acetyl-l-valyl-l-lysyl)pyrrolidine-2-carboxamido)-3-methylbutanamido)-4-oxobutanoic Acid Hydrotrifluoroacetate (13)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (59.9 mg, 46%): RP-HPLC: >99%, (t_R = 6.26, k = 1.28). ¹H NMR (300 MHz, D₂O): δ 4.88 (dd, J = 4.6, 1.4 Hz, 1H), 4.52–4.45 (m, 1H), 4.36–4.24 (m, 1H), 4.17–4.06 (m, 1H), 3.93–3.83 (m, 2H), 3.78–3.66 (m, 1H), 3.55–3.44 (m, 1H), 2.85 (t, J = 7.4 Hz, 2H), 2.71–2.32 (m, 2H), 2.27–2.07 (m, 1H), 1.98–1.49 (m, 12H), 1.44–1.21 (m, 2H), 0.88–0.72 (m, 12H). ¹³C NMR (75 MHz, D₂O): δ 175.24, 175.18, 174.34, 173.98, 173.95, 173.73, 173.15, 173.04, 171.74, 89.68, 60.30, 60.27, 59.90, 59.85, 59.61, 51.53, 51.44, 51.28, 48.07, 39.21, 33.94, 33.85, 29.93, 29.42, 26.34, 24.68, 21.87, 21.58, 18.37, 17.77. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₇H₄₇N₆O₈⁺, 583.3450; found, 583.3454: C₂₇H₄₆N₆O₈ × C₂HF₃O₂ (696.72).

(4S,7S,10S,13S,16S)-7-(4-Aminobutyl)-16-formyl-4-(4-hydroxybenzyl)-10,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (14)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (58.8 mg, 41%): RP-HPLC: 98%, (t_R = 7.33, k = 1.44). ¹H NMR (300 MHz, D₂O): δ 6.99–6.90 (m, 2H), 6.70–6.63 (m, 2H), 4.87 (dd, J = 9.1, 4.7 Hz, 1H), 4.34 (t, J = 9.0 Hz, 1H), 4.19–4.04 (m, 2H), 3.97–3.84 (m, 2H), 2.90–2.30 (m, 6H), 1.81 (s, 5H), 1.66–1.39 (m, 4H), 1.25–1.03 (m, 2H), 0.85–0.68 (m, 12H). ¹³C NMR (75 MHz, D₂O): δ 175.33, 175.15, 173.99, 173.21, 173.08, 173.04, 172.75, 172.63, 154.44, 130.45, 127.89, 115.47, 89.66, 59.76, 59.53, 55.38, 52.95, 51.50, 39.17, 36.21, 33.98, 30.66, 30.28, 30.02, 26.27, 21.88, 21.59, 18.37, 17.74. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₁H₄₉N₆O₉⁺, 649.3556; found, 649.3566: C₃₁H₄₈N₆O₉ × C₂HF₃O₂ (762.78).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4-(4-hydroxybenzyl)-13-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (15)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (17.4 mg, 27%): RP-HPLC: >99%, (t_R = 6.27, k = 1.09). ¹H NMR (400 MHz, D₂O): δ 7.03–6.98 (m, 2H), 6.75–6.68 (m, 2H), 4.91 (dd, J = 6.3, 4.7 Hz, 1H), 4.52–4.45 (m, 1H), 4.36 (t, J = 7.5 Hz, 1H), 4.25–4.06 (m, 2H), 3.96 (dd, J = 7.8, 1.6 Hz, 1H), 2.93–2.37 (m, 8H), 1.99–1.87 (m, 1H), 1.85 (s, 3H), 1.74–1.49 (m, 4H), 1.34–1.17 (m, 2H), 0.85–0.75 (m, 6H). ¹³C NMR (101 MHz, D₂O): δ 175.21, 175.10, 174.21, 173.90, 173.41, 173.36, 173.27, 172.96, 172.85, 171.67, 171.60, 154.50, 130.48, 127.93, 115.50, 89.70, 59.73, 55.55, 53.63, 53.56, 51.58, 51.49, 49.91, 39.20, 36.00, 35.07, 33.91, 30.26, 30.11, 26.26, 22.00, 21.95, 21.67, 18.33, 18.29, 17.72, 17.67. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₀H₄₅N₆O₁₁⁺, 665.3141; found, 665.3155: C₃₀H₄₄N₆O₁₁ × C₂HF₃O₂ (778.70).

(4S,7S,10S,13S,16S)-13-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4-(4-hydroxybenzyl)-10-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (16)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (38.6 mg, 26%): RP-HPLC: >99%, (t_R = 6.28, k = 1.09). ¹H NMR (400 MHz, D₂O): δ 7.04–6.99 (m, 2H), 6.75–6.70 (m, 2H), 4.92 (dd, J = 4.6, 1.5 Hz, 1H), 4.52 (t, J = 6.7 Hz, 1H), 4.38 (t, J = 7.6 Hz, 1H), 4.24–4.07 (m, 2H), 3.96–3.85 (m, 1H), 2.93–2.37 (m, 8H), 2.00–1.87 (m, 1H), 1.85 (s, 3H), 1.79–1.46 (m, 4H), 1.42–1.18 (m, 2H), 0.83–0.74 (m, 6H). ¹³C NMR (101 MHz, D₂O): δ 175.29, 175.11, 174.18, 174.15, 173.78, 173.39, 173.36, 173.28, 173.20, 171.87, 171.78, 154.51, 130.49, 127.93, 115.53, 89.71, 59.87, 55.46, 53.56, 51.50, 49.91, 39.21, 36.07, 35.12, 34.20, 33.90, 29.89, 26.19, 22.00, 21.95, 21.63, 18.40, 17.81. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₀H₄₅N₆O₁₁⁺, 615.3141; found, 665.3146: C₃₀H₄₄N₆O₁₁ × C₂HF₃O₂ (778.74).

(S)-3-((S)-2-Acetamido-3-(4-hydroxyphenyl)propanamido)-4-((S)-2-(((S)-1-(((S)-1-carboxy-3-oxopropan-2-yl)amino)-3-methyl-1-oxobutan-2-yl)carbamoyl)pyrrolidin-1-yl)-4-oxobutanoic Acid (17)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (55.3 mg, 47%): RP-HPLC: 99%, (t_R = 7.69, k = 1.80). ¹H NMR (300 MHz, D₂O): δ 7.00–6.89 (m, 2H), 6.72–6.59 (m, 2H), 4.86 (dd, J = 4.6, 2.1 Hz, 1H), 4.73 (t, J = 7.0 Hz, 1H), 4.32 (t, J = 7.9 Hz, 1H), 4.20–4.00 (m, 2H), 3.86 (d, J = 7.7 Hz, 1H), 3.54–3.09 (m, 2H), 2.85–2.29 (m, 6H), 2.11–1.90 (m, 1H), 1.96–1.66 (m, 7H), 0.87–0.69 (m, 6H). ¹³C NMR (75 MHz, D₂O): δ 175.24, 175.12, 174.00, 173.84, 173.06, 172.98, 172.53, 169.59, 169.49, 154.44, 130.49, 127.79, 115.37, 89.68, 60.50, 60.43, 59.83, 59.77, 55.20, 51.48, 51.43, 47.74, 47.61, 36.16, 35.45, 33.90, 29.87, 29.30, 24.36, 21.57, 17.73, 17.69. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₉H₄₀N₅O₁₁⁺, 634.2719; found, 634.2727: C₂₉H₃₉N₅O₁₁ (633.66).

(4S,7S,10S,13S,16S)-10-(3-Aminopropyl)-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (18)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (72.5 mg, 55%): RP-HPLC: >99%, (t_R = 5.91, k = 0.97). ¹H NMR (400 MHz, D₂O): δ 4.90 (dd, J = 4.7, 1.9 Hz, 1H), 4.62–4.55 (m, 1H), 4.30–4.06 (m, 2H), 4.00–3.86 (m, 2H), 2.95–2.36 (m, 6H), 1.98–1.84 (m, 5H), 1.79–1.48 (m, 4H), 0.84–0.74 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.20, 175.13, 174.65, 173.85, 173.76, 173.41, 172.86, 172.69, 171.99, 89.70, 59.95, 59.78, 53.14, 53.05, 51.58, 51.45, 50.09, 38.86, 35.11, 34.09, 33.88, 30.08, 29.80, 27.90, 23.11, 21.66, 18.37, 18.28, 17.75, 17.57. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₅H₄₃N₆O₁₀⁺, 587.3035; found, 587.3042: C₂₅H₄₂N₆O₁₀ × C₂HF₃O₂ (700.67).

(4S,7S,10S,13S,16S)-10-(2-Aminoethyl)-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (19)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (69.7 mg, 54%): RP-HPLC: >99%, (t_R = 5.93, k = 0.98). ¹H NMR (400 MHz, D₂O): δ 4.91 (d, J = 4.9 Hz, 1H), 4.58 (t, J = 1.6 Hz, 1H), 4.40–4.09 (m, 2H), 3.99–3.88 (m, 2H), 2.96–2.37 (m, 6H), 2.12–2.01 (m, 1H), 2.00–1.87 (m, 6H), 0.86–0.75 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.11, 174.70, 173.83, 173.39, 172.85, 172.27, 172.20, 171.96, 171.69, 89.70, 59.93, 51.60, 51.49, 51.13, 50.16, 36.18, 35.10, 33.92, 30.01, 29.79, 28.74, 21.67, 18.37, 18.30, 17.74, 17.54. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₄H₄₁N₆O₁₀⁺, 573.2879; found, 573.2886: C₂₄H₄₀N₆O₁₀ × C₂HF₃O₂ (686.64).

(4S,7S,10S,13S,16S)-10-(Aminomethyl)-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (20)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (52.3 mg, 41%): RP-HPLC: 95%, (t_R = 6.12, k = 1.04). ¹H NMR (400 MHz, D₂O): δ 4.93 (d, J = 4.7 Hz, 1H), 4.65–4.55 (m, 2H), 4.23–4.10 (m, 1H), 4.04–3.90 (m, 2H), 3.44–3.14 (m, 2H), 2.94–2.59 (m, 4H), 2.02–1.90 (m, 5H), 0.87–0.76 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.21, 174.90, 174.10, 173.95, 172.88, 172.83, 172.49, 169.50, 169.45, 89.76, 89.68, 60.06, 59.90, 51.64, 51.54, 50.70, 50.38, 39.81, 35.11, 34.12, 33.96, 29.71, 21.69, 18.37, 18.29, 17.73, 17.55. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₃H₃₉N₆O₁₀⁺, 559.2722; found, 559.2732: C₂₃H₃₈N₆O₁₀ × C₂HF₃O₂ (672.61).

(4S,7S,10S,13S,16S)-7-(Carboxymethyl)-16-formyl-10-(3-guanidinopropyl)-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (21)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (4.0 mg, 2%): RP-HPLC: 96%, (t_R = 6.71, k = 1.90). ¹H NMR (400 MHz, MeOD): δ 4.68–4.60 (m, 1H), 4.58–4.51 (m, 1H), 4.42–4.34 (m, 1H), 4.33–4.24 (m, 1H), 4.17–4.09 (m, 1H), 4.09–4.02 (m, 1H), 3.22–3.12 (m, 2H), 2.95–2.87 (m, 1H), 2.84–2.76 (m, 1H), 2.73–2.62 (m, 1H), 2.53–2.44 (m, 1H), 2.23–2.02 (m, 2H), 2.01 (d, J = 1.3 Hz, 3H), 1.96–1.82 (m, 1H), 1.82–1.68 (m, 1H), 1.62 (p, J = 6.9 Hz, 2H), 1.01–0.89 (m, 12H). HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₅N₈O₁₀⁺, 629.3253; found, 629.3262: C₂₆H₄₄N₈O₁₀ × C₂HF₃O₂ (742.71).

(4S,7S,10S,13S,16S)-10-(4-Acetamidobutyl)-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid (22)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (72.9 mg, 60%): RP-HPLC: 99%, (t_R = 7.38, k = 1.46). ¹H NMR (400 MHz, D₂O): δ 4.93–4.87 (m, 1H), 4.60 (t, J = 2.0 Hz, 1H), 4.24–4.03 (m, 2H), 3.93 (dd, J = 11.8, 7.4 Hz, 2H), 3.02 (t, J = 6.9 Hz, 2H), 2.91–2.35 (m, 4H), 2.01–1.85 (m, 5H), 1.85 (s, 3H), 1.75–1.51 (m, 2H), 1.38 (p, J = 7.3 Hz, 2H), 1.32–1.09 (m, 2H), 0.85–0.73 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.09, 174.56, 173.95, 173.91, 173.80, 173.59, 173.56, 173.36, 172.81, 172.03, 89.68, 59.90, 59.74, 53.83, 51.50, 50.02, 39.24, 35.10, 33.89, 30.53, 30.06, 29.91, 27.67, 22.40, 21.83, 21.67, 18.40, 18.31, 17.70, 17.65. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₈H₄₇N₆O₁₁⁺, 643.3297; found, 643.3312: C₂₈H₄₆N₆O₁₁ (642.71).

(4S,7S,10S,13S,16S)-13-(3-Aminopropyl)-7-(carboxymethyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (23)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (60.9 mg, 46%): RP-HPLC: >99%, (t_R = 6.19, k = 1.06). ¹H NMR (400 MHz, D₂O): δ 4.92 (dd, J = 4.6, 2.2 Hz, 1H), 4.67–4.59 (m, 1H), 4.30–4.19 (m, 1H), 4.16–4.09 (m, 1H), 4.00–3.90 (m, 2H), 3.02–2.80 (m, 3H), 2.76–2.53 (m, 2H), 2.47–2.37 (m, 1H), 2.03–1.87 (m, 5H), 1.81–1.50 (m, 4H), 0.86–0.77 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.30, 175.07, 174.55, 174.19, 173.64, 173.37, 173.15, 172.78, 172.69, 172.21, 172.14, 89.70, 89.64, 59.85, 59.79, 59.69, 53.20, 51.52, 51.51, 51.50, 50.04, 38.89, 35.20, 34.23, 33.90, 29.91, 28.16, 28.05, 23.27, 23.20, 21.67, 18.46, 18.41, 18.33, 17.73, 17.63. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₅H₄₃N₆O₁₀⁺, 587.3035; found, 587.3043: C₂₅H₄₂N₆O₁₀ × C₂HF₃O₂ (700.67).

(4S,7S,10S,13S,16S)-13-(2-Aminoethyl)-7-(carboxymethyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (24)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (10.5 mg, 8%): RP-HPLC: 95%, (t_R = 6.20, k = 1.07). ¹H NMR (400 MHz, D₂O): δ 4.94 (t, J = 4.6 Hz, 1H), 4.65–4.62 (m, 1H), 4.40–4.30 (m, 1H), 4.19–4.10 (m, 1H), 4.01–3.92 (m, 2H), 3.00–2.80 (m, 3H), 2.76–2.55 (m, 2H), 2.49–2.39 (m, 1H), 2.12–1.89 (m, 7H), 0.87–0.78 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.40, 175.10, 174.59, 174.40, 173.68, 173.39, 173.35, 172.77, 172.35, 172.30, 171.67, 89.68, 59.87, 59.82, 59.74, 51.59, 51.54, 51.26, 51.18, 50.07, 36.26, 36.22, 35.34, 34.36, 29.89, 29.03, 28.90, 21.65, 18.43, 18.39, 18.31, 17.74, 17.61. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₄H₄₁N₆O₁₀⁺, 573.2879; found, 573.2883: C₂₄H₄₀N₆O₁₀ × C₂HF₃O₂ (686.64).

(4S,7S,10S,13S,16S)-13-(Aminomethyl)-7-(carboxymethyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (25)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (47.2 mg, 37%): RP-HPLC: >99%, (t_R = 6.16, k = 1.05). ¹H NMR (400 MHz, D₂O): δ 4.94 (dd, J = 9.3, 4.6 Hz, 1H), 4.67–4.60 (m, 2H), 4.20–4.09 (m, 1H), 4.07–3.97 (m, 1H), 3.96–3.89 (m, 1H), 3.44–3.30 (m, 1H), 3.26–3.12 (m, 1H), 3.04–2.36 (m, 4H), 2.09–1.86 (m, 5H), 0.85–0.77 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.36, 174.93, 174.62, 174.59, 174.46, 174.43, 173.73, 173.56, 172.56, 172.51, 169.83, 169.38, 169.37, 89.62, 89.57, 59.93, 59.70, 59.68, 51.66, 51.53, 50.85, 50.66, 50.09, 40.11, 39.97, 35.26, 34.34, 33.94, 29.89, 21.66, 18.46, 18.43, 18.31, 17.65, 17.56, 17.54. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₃H₃₉N₆O₁₀⁺, 559.2722; found, 559.2731: C₂₃H₃₈N₆O₁₀ × C₂HF₃O₂ (672.61).

(4S,7S,10S,13S,16S)-7-(Carboxymethyl)-16-formyl-13-(3-guanidinopropyl)-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (26)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (42.5 mg, 31%): RP-HPLC: >99%, (t_R = 6.58, k = 1.39). ¹H NMR (300 MHz, D₂O): δ 4.89 (d, J = 4.6 Hz, 1H), 4.60 (t, J = 6.9 Hz, 1H), 4.31–4.02 (m, 2H), 3.98–3.84 (m, 2H), 3.06 (t, J = 6.5 Hz, 2H), 2.90–2.76 (m, 1H), 2.73–2.31 (m, 2H), 2.02–1.73 (m, 6H), 1.82–1.37 (m, 4H), 0.84–0.73 (m, 12H). ¹³C NMR (75 MHz, D₂O): δ 175.09, 174.58, 174.54, 174.09, 174.04, 173.67, 173.28, 171.97, 156.73, 59.83, 59.70, 53.37, 52.24, 51.47, 50.03, 40.46, 40.41, 35.17, 29.88, 27.69, 24.39, 21.62, 18.34, 18.30, 17.70, 17.60. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₅N₈O₁₀⁺, 629.3253; found, 629.3258: C₂₆H₄₄N₈O₁₀ × C₂HF₃O₂ (742.71).

(4S,7S,10S,13S,16S)-13-(4-Acetamidobutyl)-7-(carboxymethyl)-16-formyl-4,10-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid (27)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (49.2 mg, 41%): RP-HPLC: >99%, (t_R = 7.36, k = 1.68). ¹H NMR (300 MHz, D₂O): δ 4.88 (dd, J = 4.6, 0.9 Hz, 1H), 4.60 (t, 1H), 4.19–4.00 (m, 2H), 3.98–3.86 (m, 2H), 2.99 (t, J = 6.7 Hz, 2H), 2.90–2.31 (m, 4H), 2.01–1.84 (m, 5H), 1.81 (s, 3H), 1.68–1.49 (m, 2H), 1.41–1.10 (m, 4H), 0.83–0.72 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.19, 175.10, 174.53, 174.51, 174.03, 173.99, 173.62, 173.53, 173.45, 173.11, 173.07, 172.20, 172.08, 89.68, 59.83, 59.80, 59.62, 53.83, 51.49, 50.05, 39.18, 35.13, 34.04, 33.84, 30.73, 30.61, 29.93, 27.71, 22.43, 22.37, 21.88, 21.65, 18.45, 18.42, 18.34, 17.62. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₈H₄₇N₆O₁₁⁺, 643.3304; found, 643.3297: C₂₈H₄₆N₆O₁₁ (642.71).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-7-(carboxymethyl)-16-formyl-4,13-bis((R)-1-hydroxyethyl)-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (28)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (21.0 mg, 16%): RP-HPLC: >99%, (t_R = 3.11, k = 0.04). ¹H NMR (300 MHz, D₂O): δ 4.93–4.88 (m, 1H), 4.60–4.52 (m, 1H), 4.25–3.98 (m, 6H), 2.90–2.59 (m, 5H), 2.48–2.35 (m, 1H), 1.95 (d, J = 4.1 Hz, 3H), 1.74–1.49 (m, 4H), 1.31–1.21 (m, 2H), 1.12–0.99 (m, 6H). HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₄H₄₁N₆O₁₂⁺, 605.2777; found, 605.2783: C₂₄H₄₀N₆O₁₂ × C₂HF₃O₂ (718.64).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-13-((S)-sec-butyl)-7-(carboxymethyl)-16-formyl-4-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (29)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (95.0 mg, 69%): RP-HPLC: 98%, (t_R = 7.06, k = 1.35). ¹H NMR (400 MHz, D₂O): δ 4.94–4.89 (m, 1H), 4.63–4.55 (m, 1H), 4.27–4.09 (m, 2H), 4.06–3.96 (m, 1H), 3.93 (d, J = 6.9 Hz, 1H), 2.94–2.38 (m, 6H), 2.01–1.86 (m, 4H), 1.81–1.66 (m, 2H), 1.66–1.61 (m, 1H), 1.60–1.51 (m, 2H), 1.42–0.99 (m, 4H), 0.91–0.65 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.17, 175.08, 174.63, 173.90, 173.71, 173.53, 173.21, 172.96, 171.99, 89.69, 59.98, 58.52, 53.55, 51.51, 50.10, 39.20, 35.97, 35.12, 33.89, 30.33, 29.82, 26.23, 24.49, 21.95, 21.70, 18.31, 17.63, 14.71, 10.01. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₇H₄₇N₆O₁₀⁺, 615.3348; found, 615.3353: C₂₇H₄₆N₆O₁₀ × C₂HF₃O₂ (728.72).

(4S,7S,10S,13S,16S)-10-(4-Aminobutyl)-13-benzyl-7-(carboxymethyl)-16-formyl-4-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (30)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (90.5 mg, 60%): RP-HPLC: >99%, (t_R = 7.72, k = 1.57). ¹H NMR (400 MHz, D₂O): δ 7.28–7.11 (m, 5H), 4.89–4.80 (m, 1H), 4.59–4.42 (m, 2H), 4.14–4.02 (m, 2H), 3.90 (dd, J = 7.0, 1.8 Hz, 1H), 3.11–2.37 (m, 8H), 1.99–1.86 (m, 4H), 1.62–1.40 (m, 4H), 1.20–1.02 (m, 2H), 0.81 (t, J = 1.1 Hz, 6H). ¹³C NMR (101 MHz, D₂O): δ 175.02, 174.58, 174.02, 173.69, 173.28, 173.12, 172.56, 172.48, 172.21, 129.24, 128.80, 128.74, 127.13, 89.65, 59.96, 54.85, 53.80, 51.46, 49.99, 39.13, 36.80, 35.05, 33.85, 30.11, 29.83, 26.17, 21.81, 21.72, 21.67, 18.28, 17.63. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₃₀H₄₅N₆O₁₀⁺, 649.3192; found, 649.3197: C₃₀H₄₄N₆O₁₀ × C₂HF₃O₂ (762.70).

(4S,7S,10S,13S,16S)-10-Benzyl-7-(carboxymethyl)-16-formyl-4,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid (31)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (36.4 mg, 18%): RP-HPLC: 97%, (t_R = 10.89, k = 2.63). ¹H NMR (400 MHz, MeOD): δ 7.32–7.09 (m, 5H), 4.66–4.61 (m, 1H), 4.61–4.50 (m, 2H), 4.36–4.23 (m, 1H), 4.17–4.10 (m, 1H), 4.10–4.02 (m, 1H), 3.18 (dd, J = 14.2, 4.8 Hz, 1H), 3.05–2.93 (m, 1H), 2.87–2.76 (m, 1H), 2.73–2.59 (m, 2H), 2.56–2.40 (m, 1H), 2.13–1.91 (m, 5H), 0.99–0.82 (m, 12H). ¹³C NMR (101 MHz, MeOD): δ 173.68, 173.57, 172.62, 172.57, 172.43, 171.63, 171.32, 137.06, 128.89, 128.11, 126.32, 96.79, 59.45, 55.27, 50.55, 50.12, 47.89, 36.89, 34.90, 33.06, 30.42, 30.17, 21.09, 18.33, 17.38. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₉H₄₂N₅O₁₀⁺, 620.2926; found, 620.2931: C₂₉H₄₁N₅O₁₀ (619.67).

(S)-3-((S)-2-Acetamido-3-methylbutanamido)-4-(((S)-6-amino-1-((S)-2-(((S)-1-carboxy-3-oxopropan-2-yl)carbamoyl)pyrrolidin-1-yl)-1-oxohexan-2-yl)amino)-4-oxobutanoic Acid Hydrotrifluoroacetate (32)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (25.4 mg, 19%): RP-HPLC: >99%, (t_R = 4.91, k = 0.64). ¹H NMR (400 MHz, D₂O): δ 4.97–4.88 (m, 1H), 4.60 (dd, J = 8.0, 5.8 Hz, 1H), 4.53–4.45 (m, 1H), 4.37–4.24 (m, 1H), 4.24–4.06 (m, 1H), 3.94 (d, J = 7.0 Hz, 1H), 3.75–3.45 (m, 2H), 2.94–2.35 (m, 6H), 2.21–2.07 (m, 1H), 1.98–1.89 (m, 5H), 1.89–1.26 (m, 8H), 0.82 (dd, J = 6.8, 1.9 Hz, 6H). ¹³C NMR (101 MHz, D₂O): δ 175.31, 175.23, 174.52, 173.91, 173.84, 173.65, 171.71, 171.63, 171.57, 89.69, 60.75, 60.70, 60.53, 59.82, 51.50, 51.45, 49.99, 47.93, 39.21, 35.18, 34.10, 29.87, 29.66, 29.48, 26.38, 24.67, 24.53, 21.75, 21.67, 18.30, 17.57. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₃N₆O₁₀⁺, 599.3035; found, 599.3047: C₂₆H₄₂N₆O₁₀ × C₂HF₃O₂ (712.67).

(S)-3-((S)-2-Acetamido-3-methylbutanamido)-4-((S)-2-((S)-2-(((S)-1-carboxy-3-oxopropan-2-yl)carbamoyl)pyrrolidine-1-carbonyl)pyrrolidin-1-yl)-4-oxobutanoic Acid (33)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (28.1 mg, 26%): RP-HPLC: >99%, (t_R = 6.30, k = 1.10). ¹H NMR (400 MHz, D₂O): δ 4.96–4.91 (m, 1H), 4.91–4.84 (m, 1H), 4.61–4.53 (m, 1H), 4.34–3.98 (m, 2H), 3.94 (d, J = 7.3 Hz, 1H), 3.80–3.21 (m, 4H), 2.94–2.37 (m, 4H), 2.28–2.07 (m, 2H), 2.04–1.69 (m, 10H), 0.86–0.76 (m, 6H). ¹³C NMR (101 MHz, D₂O): δ 175.33, 175.27, 174.33, 173.91, 173.79, 173.41, 172.19, 171.93, 169.55, 89.66, 60.65, 60.55, 59.56, 58.95, 51.48, 51.41, 48.38, 47.94, 47.74, 35.09, 34.06, 29.99, 29.30, 28.25, 24.61, 24.51, 21.62, 18.33, 17.53. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₅H₈₇N₅O₁₀⁺, 568.2613; found, 568.2624: C₂₅H₃₇N₅O₁₀ (567.60).

(S)-3-((S)-2-Acetamido-3-methylbutanamido)-4-((S)-2-(((S)-6-amino-1-(((S)-1-carboxy-3-oxopropan-2-yl)amino)-1-oxohexan-2-yl)carbamoyl)pyrrolidin-1-yl)-4-oxobutanoic Acid Hydrotrifluoroacetate (34)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (66.3 mg, 50%): RP-HPLC: 99%, (t_R = 5.43, k = 0.97). ¹H NMR (300 MHz, D₂O): δ 4.91–4.87 (m, 1H), 4.87–4.79 (m, 1H), 4.30–4.20 (m, 1H), 4.16–4.03 (m, 2H), 3.90 (d, J = 7.3 Hz, 1H), 3.78–3.47 (m, 2H), 2.91–2.32 (m, 6H), 2.21–2.05 (m, 1H), 1.97–1.73 (m, 7H), 1.73–1.46 (m, 4H), 1.39–1.19 (m, 2H), 0.84–0.71 (m, 6H). ¹³C NMR (75 MHz, D₂O): δ 175.24, 175.11, 174.38, 174.21, 174.02, 173.96, 173.56, 173.49, 173.43, 170.43, 89.69, 60.72, 60.61, 59.58, 53.72, 53.68, 51.48, 51.44, 48.32, 48.27, 48.09, 39.20, 35.28, 34.13, 30.17, 29.94, 29.44, 26.14, 24.54, 22.08, 21.59, 18.31, 17.58. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₆H₄₂N₆O₁₀⁺, 655.3661; found, 655.3668: C₂₆H₄₂N₆O₁₀ × C₂HF₃O₂ (768.75).

(4S,7S,10S,13S,16S)-13-(4-Aminobutyl)-4-((S)-sec-butyl)-7-(carboxymethyl)-16-formyl-10-isopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid Hydrotrifluoroacetate (35)

The title compound was synthesized according to the general procedure B, yielding a fluffy white solid (22 mg, 16%): RP-HPLC: 95%, (t_R = 7.05, k = 1.35). ¹H NMR (400 MHz, D₂O): δ 4.94–4.91 (m, 1H), 4.67–4.60 (m, 1H), 4.37–4.08 (m, 2H), 4.05–3.90 (m, 2H), 2.92–2.81 (m, 3H), 2.78–2.38 (m, 3H), 1.92 (s, 4H), 1.78–1.49 (m, 5H), 1.43–1.21 (m, 3H), 1.16–1.00 (m, 1H), 0.90–0.71 (m, 12H). ¹³C NMR (101 MHz, D₂O): δ 175.30, 175.12, 174.53, 174.51, 174.17, 174.15, 173.80, 173.69, 173.29, 173.22, 173.15, 173.13, 172.16, 172.06, 89.70, 59.81, 58.75, 53.59, 51.50, 50.04, 39.22, 36.10, 35.17, 34.19, 33.90, 30.56, 30.44, 29.93, 26.20, 24.62, 22.01, 21.97, 21.65, 18.46, 18.41, 17.70, 14.75, 10.34. HRMS (ESI-MS) m/z: [M + H⁺] calcd for C₂₇H₄₇N₆O₁₀⁺, 615.3348; found, 615.3360: C₂₇H₄₆N₆O₁₀ × C₂HF₃O₂ (728.72).

(4S,7S,10S,13S,16S)-7-(Carboxymethyl)-16-formyl-4-(4-hydroxybenzyl)-10,13-diisopropyl-2,5,8,11,14-pentaoxo-3,6,9,12,15-pentaazaoctadecan-18-oic Acid (36)

The title compound was synthesized according to the general procedure A, yielding a fluffy white solid (32.1 mg, 27%): RP-HPLC: 99%, (t_R = 8.88, k = 1.96). ¹H NMR (300 MHz, D₂O): δ 7.03–6.91 (m, 2H), 6.74–6.60 (m, 2H), 4.87 (dd, J = 6.8, 4.7 Hz, 1H), 4.48 (t, J = 6.7 Hz, 1H), 4.35 (t, J = 7.5 Hz, 1H), 4.18–4.03 (m, 1H), 3.97–3.86 (m, 2H), 2.96–2.28 (m, 6H), 1.96–1.76 (m, 5H), 0.83–0.67 (m, 12H). ¹³C NMR (75 MHz, D₂O): δ 175.33, 175.16, 174.10, 174.05, 173.39, 173.34, 173.11, 172.88, 172.78, 171.66, 171.59, 154.48, 130.48, 127.95, 115.51, 89.69, 59.81, 59.69, 55.35, 51.51, 49.94, 36.10, 35.16, 33.94, 29.97, 21.62, 18.37, 17.88, 17.75. m/z: [M + H⁺] calcd for C₂₉H₄₂N₅O₁₁⁺, 636.2875; found, 636.2887: C₂₉H₄₁N₅O₁₁ (635.67).

Fluorometric Enzyme Assay

384-Well Protocol

Compound affinity for caspases was measured in fluorometric assays. Casp2, cpCasp2, and Casp3 were produced in house as described below. Human recombinants Casp1, 6, 7, and 9 were purchased from BioVision (Milpitas/CA, USA). AFC fluorogenic substrates and control peptides (AcYVAD-CHO, AcVDVAD-CHO, AcDEVD-CHO, AcVEID-CHO, and AcLEHD-CHO) were purchased from Bachem (Torrance/CA, USA). K_m values were determined experimentally to be the following: Casp1: 5.9 μM; Casp2: 37.1 μM; cpCasp2: 89.2 μM; Casp3: 7.6 μM; Casp6: 43.1 μM; Casp7: 13.8 μM; and Casp9: 149.1 μM. Enzymes were diluted in buffer: 100 mM MES (pH 6.5) for Casp2 and cpCasp2 or 100 mM HEPES (pH 7.0) for all other caspases, plus 150 mM NaCl, 0.1% CHAPS, 1.5% sucrose, 10 mM DTT. Enzyme concentrations were 0.05 U/well for Casp1, 6, and 7; 0.5 U/well for Casp9, 20 nM/well for Casp2 and cpCasp2; and 2 nM/well for Casp3. Enzyme in buffer (19 μL) was added per well in a black 384-well Corning 4514 assay plate. Test compounds were serially diluted in dimethyl sulfoxide (DMSO) and plated in duplicate into a Corning 3656 transfer plate. The test compound was added to assay plates in 0.5 μL aliquots per well and mixed 10 times using a BiomekFX (Beckman Coulter). The compound and enzyme mixture were incubated at 37 °C for 5 min for reversible inhibitors. The BiomekFX was then used to add and mix 0.5 μL of the AFC substrate in DMSO from a Corning 3656 transfer plate (final assay concentrations: 5 μM AcYVAD-AFC for Casp1, 10 μM Z-VDVAD-AFC for Casp2 and cpCasp2, 5 μM AcDEVD-AFC for Casp3, 5 μM Z-VEID-AFC for Casp6, 5 μM AcDEVD-AFC for Casp7, and 34 μM AcLEHD-AFC for Casp9) to the assay plate for a total assay volume of 20 μL. Fluorescence from free AFC was read at 37 °C every 5 min over an hour using a CLARIOstar (BMG Labtech) plate reader (λ_ex = 400 nm, λ_em = 505 nm). The 40 min time point was reported, consistent with reported literature.^23,42−44

96-Well Protocol (Casp2/3)

Compound affinity for caspases was measured in fluorometric assays. Casp2, cpCasp2, and Casp3 were produced in house as described below. AFC fluorogenic substrates Z-VDVAD-AFC and AcDEVD-AFC and control peptides AcVDVAD-CHO and AcDEVD-CHO were purchased from Bachem (Torrance/CA, USA). The enzyme was diluted in buffer: 100 mM MES (pH 6.5) for Casp2 and cpCasp2 or 100 mM HEPES (pH 7.0) for Casp3, plus 150 mM NaCl, 0.1% CHAPS, 1.5% sucrose, and 10 mM DTT. Enzyme concentrations were 5 nM/well for Casp2 and cpCasp2 and 2 nM/well for Casp3. Enzyme in buffer (96.5 μL) was added per well in a black Corning 3356 96-well assay plate. Test compounds were serially diluted in DMSO and plated in triplicate in a Corning 3357 transfer plate. The test compound was added to assay plates in 1 μL aliquots per well and mixed 10 times using a BiomekFX (Beckman Coulter). The compound and enzyme mixture was incubated at 37 °C for 5 min. The BiomekFX was then used to add and mix 2.5 μL of the AFC substrate in DMSO from a transfer plate (final assay concentrations: 25 μM Z-VDVAD-AFC for Casp2 and cpCasp2, 10 μM AcDEVD-AFC for Casp3) to the assay plate for a total assay volume of 100 μL in the assay plate. Fluorescence from free AFC was read at 37 °C every 5 min over an hour using a CLARIOstar (BMG Labtech) plate reader (λ_ex = 400 nm, λ_em = 505 nm). The 40 min time point was reported, consistent with the reported literature.^23,42−44

Expression and Purification of Recombinant Caspase-2 and Tau

The pET23b vector-encoding human caspase-2 was a gift from Prof. Dr. Michelle Arkin at University of California, San Francisco.²⁰ The htau0N4R-pET28a plasmid⁸ was used as the initial template that encodes wild-type human microtubule-associated protein tau N4R isoform (containing four microtubule-binding domains but no amino-terminal inserts) ORF. Primers used to generate tau mutations are listed in Table S2. The DNA sequences of tau mutants were verified by classic Sanger sequencing analyses. Briefly, the Casp2 plasmid was transformed into Escherichia coli Lemo21 DE3 cells with 2 mM rhamnose that were cultured at a 1 L scale in shake flasks at 37 °C. When the culture reached an OD₆₀₀ = 0.3, incubation temperature was reduced to 28 °C. When the OD₆₀₀ reached 0.6–0.8, protein expression was induced with 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and allowed to shake at 250 RPM and 28 °C overnight. 16 h after induction, cells were harvested through centrifugation (5000 rpm for 20 min at 4 °C), lysed by French press (2 passes at 1000 psi), and centrifuged again for 30 min at 20,000 rpm at 4 °C. The supernatant was resuspended in buffer (100 mM Tris pH 8.0, 100 mM NaCl, 20 mM imidazole). Purification was performed on a Cytiva ÅKTA Pure system (Marlborough, MA) by nickel-affinity chromatography on a HisTrap FF crude column (Cytiva) followed by ion-exchange chromatography on a HiTrap Q FF column (Cytiva). Protein was diluted to 0.5 mg/mL and stored at −80 °C in 100 mM MES (pH 6.5), 150 mM NaCl, 0.1% CHAPS, 1.5% sucrose, and 10 mM DTT. Typical yield for the expression, isolation, and purification was approximately 0.1 mg/L. Expression of recombinant Casp2 and tau proteins was induced using 150 μM IPTG (Promega, Madison, WI) in the BL21(DE3) E. coli strain (MilliporeSigma) at 16 °C for 14 h, while shaking at 250 rpm. Cells were then harvested via centrifugation at 6000g, 4 °C for 15 min, followed by lysis via sonication. Proteins were initially purified using HisPur Ni-NTA resin (Thermo Fisher Scientific, Waltham, MA) followed by metal chelating chromatography [5 mL HiTrap Chelating HP columns (GE Healthcare Life Sciences, Piscataway, NJ)]. Further purification was carried out using either anion exchange chromatography [5 mL HiTrap Q HP columns (GE Healthcare Life Sciences)] for Casp2 or cation exchange chromatography [5 mL HiTrap SP HP columns (GE Healthcare Life Sciences)] for tau. Proteins were stored in 1X phosphate-buffered saline (pH 7.4) at 10 mg/mL [concentration determined using a BCA assay (Thermo Fisher Scientific)], −80 °C until further use.

Expression and Purification of Circularly Permutated Caspase-2 (cpCasp2)

A circularly permuted caspase-2 (cpCasp2) based on modifications to Casp2 (UniProt P42575) suggested by Cserjan-Puschmann et al. (2020)⁴⁵ was expressed. The gene encoding 6H-Casp2_334–452-GS-Casp2_170–333 with a D₃₄₇ → A₃₄₇ mutation was inserted into a pET30a(+) vector (GenScript, Piscataway, NJ, USA) using Ndel and EcoRV restriction sites. The transformation into Lemo21 cells was performed as recommended by the vendor (New England Biolabs, Ipswich, MA, USA), and the plates were incubated overnight at 37 °C. A single colony was picked for primary culture using a sterile tip, added to sterile Luria broth (LB) media containing antibiotics, and incubated overnight at 37 °C and 270 rpm. The primary culture was then diluted 1:100 into 1 L shake flasks containing sterile LB and the proper antibiotics. The liter cultures were incubated at 37 °C and 270 rpm, and the OD was monitored until it reached between 0.6 and 0.8. Once the desired OD was reached, the cells were induced overnight with 0.1 mM IPTG at 28 °C and 270 rpm. After induction, the cells were harvested through centrifugation at 5000g for 20 min. Typical yield was about a 3.5 g cell pellet per liter. The pellets were stored at −20 °C until purification.

The pellets were resuspended in lysis buffer (buffer A with 2.5 μM Leupeptin, 5 μM Pepstatin, 0.5 μM DNase, 1 mM PMSF, and 10 mM MgCl₂) and lysed through sonification. The lysate was clarified through centrifugation at 20,000g for 30 min. The supernatant was filtered through a 0.45 μm filter before it was purified on a BioRad Biologic DuoFlow Chromatography System. For nickel affinity chromatography a 5 mL HisTrap FF column was equilibrated with 100% buffer A (100 mM Tris, pH 8.0, 100 mM NaCl, 20 mM imidazole). A gradient of buffer B (100 mM Tris, pH 8.0, 100 mM NaCl, 400 mM imidazole) was then applied from 0 to 50% to wash and elute the protein. cpCasp2 was eluted at about 20% buffer B. Final purification was achieved on a Hi-Trap Q anion exchange column in 25 mM Tris, pH 8.0 utilizing a gradient of 0–1 M NaCl. cpCasp2 was eluted at about 150 mM NaCl. The final yield was 8–10 mg of cpCasp2 per liter. After adding 10% glycerol, the protein was stored at −80 °C.

Expression and Purification of Recombinant Caspase-3

The pET23b vector-encoding human Casp3 was a gift from Prof. Dr. Michelle Arkin at University of California, San Francisco. Expression and purification of Casp3 was guided by previously described protocols.⁴⁶ Briefly, the plasmid was transformed into E. coli BL21 pLyss DE3 cells that were cultured at a 1 L scale in shake flasks at 37 °C. When the culture reached an OD₆₀₀ = 0.3, incubation temperature was reduced to 30 C. When the OD₆₀₀ reached 0.6–0.8, protein expression was induced with 0.2 mM IPTG. Three hours post induction, cells were harvested through centrifugation (5000 rpm for 30 min at 4 °C), lysed by sonication, and centrifuged again for 30 min at 20,000 rpm at 4 °C. The supernatant was resuspended in buffer (100 mM Tris pH 8.0, 100 mM NaCl, 20 mM imidazole). Purification was performed on a Cytiva ÅKTA Pure system (Marlborough, MA) by nickel-affinity chromatography on a HisTrap FF crude column (Cytiva) followed by ion-exchange chromatography on a HiTrap Q FF column (Cytiva). Protein was diluted to 0.5 mg/mL and stored at −80 °C in 100 mM HEPES (pH 7.0), 150 mM NaCl, 0.1% CHAPS, 1.5% sucrose, and 10 mM DTT. Typical yield for the expression, isolation, and purification was approximately 3 mg/L.

Crystallography

Casp3 Crystallography

The preparation of Casp3 co-crystal structures with covalently bound inhibitors followed previously utilized protocols.^47,48 500 μL of Casp3 at 0.5 mg/mL was incubated with the 500 μM inhibitor for 30 min on ice before being concentrated to 4 mg/mL for crystallization. Crystals were grown by hanging-drop vapor diffusion in which 1 μL(protein) + 1 μL (well solution) drops were suspended over either 15% PEG 6000, 5% glycerol (v/v), 100 mM sodium citrate pH 5.3, 10 mM DTT, and 30 mM NaN₃ or 16% PEG 6000, 5% glycerol, 100 mM sodium citrate pH 6.5, and 10 mM DTT. Plate-like crystals grew within 24–48 h and were cryoprotected with well solution supplemented with 10% PEG 6000 prior to flash freezing.

Diffraction data were collected at IMCA-CAT beamline 17-ID at the Advanced Photon Source (APS), Argonne, Illinois, USA. Collection was completed at 100 K using radiation of wavelength 1.00 Å and a Dectris Eiger2 9M detector. Data were processed using autoProc and scaled using aP_scale.⁴⁹ All structures were solved using molecular replacement as implemented in Phenix.⁵⁰ The Casp3 structure with bound AcVDVAD-CHO, PDBid: 2h65, served as a search model. Iterative rounds of refinement and model building were conducted using Phenix and Coot.⁵¹ Summary data collection and refinement statistics for each of the eight structures are given in Table S1 (Supporting Information).

In Vitro Caspase-2-Catalyzed Tau Cleavage Assay

Procedure

Purified recombinant Casp2 (final concentration: 1 nM) was incubated with purified recombinant tau (molar ratio = 1:1) in a water bath at 37 °C in 1X reaction buffer (25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 0.1% (w/v) 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 10 mM dithiothreitol (DTT), pH 7.5). 20 μL of the sample was transferred at designated times (i.e., 0, 5, 30, 60, 120, and 240 min after the reaction started), and the reaction was immediately stopped by adding Laemmli sample buffer (Bio-Rad, Hercules, CA), followed by boiling at 95 °C for 5 min.

Western Blotting

The reaction samples were size-fractionated on 10% Criterion Tris–HCl Precast gels (Bio-Rad, Hercules, CA) and electrophoretically transferred onto 0.2 μm nitrocellulose membranes at a constant current of 0.4 A for 4 h at 4 °C. Membranes were first blocked using 5% (w/v) bovine serum albumin (BSA) blocking buffer [BSA dissolved in 1× phosphate-buffered saline, 0.1% (v/v) polyoxyethylene (20) sorbitan monolaurate (Tween 20), pH 7.4] at room temperature for 1 h and then incubated with mouse monoclonal antibody tau-13 (directed against amino acids 15–25 of human tau; 1:30,000; BioLegend, San Diego, CA; Cat #835201; RRID: AB_2565341) at 4 °C overnight. Following 5 min washes with wash buffer [10 mM Tris–HCl, pH 7.4; 200 mM NaCl; 0.1% (v/v) Tween 20] at room temperature five times, membranes were incubated at room temperature in goat-anti-mouse immunoglobulin G-conjugated horseradish peroxidase (HRP) (Thermo Fisher Scientific; diluted in wash buffer) for 1 h. Membranes were then washed again as described above. Western blots were developed using the West Pico electrochemiluminescence detection system (Thermo Fisher Scientific). Densitometry-based quantification of the ∼35 kDa cleavage product was performed using Optiquant (Packard Cyclone, PerkinElmer Life Sciences Inc., Boston, MA). The levels of proteins were determined from two independent experiments.

Data Analysis

Statistical analyses were performed using GraphPad Prism version 8 (GraphPad Software, La Jolla, CA). GraphPad Prism version 9 was used to calculate the IC₅₀ by fitting the dose–response data with four-parameter variable slope nonlinear regression. These were transformed into pK_i values using the Cheng–Prusoff equation.⁵² Because our compounds are covalent-reversible inhibitors, they were characterized using pK_i values and not “k_inact/K_i”, as would be necessary for covalent irreversible inhibitors. P < 0.05 was considered statistically significant. See legends of Tables 2, 3, and Figure 7 for detailed statistical methods.

Glossary

Abbreviations

AFC

7-amino-4-trifluoromethyl coumarin

APS

Advanced Photon Source

AU

absorption units

BCA

bicinchoninic acid

Casp1

caspase-1

Casp2

caspase-2

Casp3

caspase-3

Casp4

caspase-4

Casp5

caspase-5

Casp6

caspase-6

Casp7

caspase-7

Casp8

caspase-8

Casp9

caspase-9

Casp10

caspase-10

Casp11

caspase-11

CDI

1,1′-carbonyldiimidazol

CHAPS

3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate

Comp.

compound

cpCasp2

circularly permuted caspase-2

D₂O

deuterium oxide

Dab

diaminobutyric acid

Dap

diaminopropionic acid

DIC

N,N′-diisopropylcarbodiimide

DIPEA

N,N-diisopropylethylamine

EDCI

1-ethyl-3-(3-dimethylaminopropyl)carbodiimide

EtOAc

ethyl acetate

HATU

1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate

HCl

hydrochloric acid

HD

Huntington’s disease

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HOBt

1-hydroxybenzotriazole hydrate

HRP

horseradish peroxidase

IPTG

isopropyl-β-d-thiogalactopyranoside

k

retention/capacity factor

LB

Luria broth

MES

2-(N-morpholino)ethanesulfonic acid

MeCN

acetonitrile

MeOD

deuterated methanol

MeOH

methanol

NAFLD

nonalcoholic fatty liver disease

NASH

nonalcoholic steatohepatitis

OD

optical density

ORF

open reading frame

Orn

ornithine

Pd/C

palladium on carbon

PE

petroleum ether

PMSF

phenylmethylsulfonylfluorid

PTFE

polytetrafluoroethylene

PTMs

post-translational modifications

R²

coefficient of determination

RFU

relative fluorescence units

RP-HPLC

reversed phase high performance liquid chromatography

SD

standard deviation

S1P

site 1 protease

t₀

dead time

SPPS

solid-phase peptide synthesis

UHD

ultrahigh definition

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.1c00251.

• Crystallography, preparation of aspartic acid-loaded semicarbazide amino-merrifield resin, structures of synthesized pentapeptides 1 and 3–36, NMR spectra of pentapeptides 1 and 3–36, RP-HPLC chromatograms of pentapeptides 1 and 3–36, chemical stability of pentapeptides 3, 15, 22, and 32, saturation binding experiments in the fluorometric enzyme assay, competition binding experiments of 24 in the fluorometric enzyme assay, mutageneses of tau protein, and physicochemical properties and bioavailability radars of 1–2, 4–5, 24, and 35 (PDF)

• SMILES_MFS (ZIP)

Accession Codes

Atomic coordinates and reflection data for crystallographic complexes of caspase-3 with peptides 1, 3, 4, 18, 21, 22, 24, and 31 (Table S1, Supporting Information) have been deposited into the Protein Data Bank with accession codes: 7rnb, 7rnd, 7rne, 7rnf, 7rn7, 7rn8, 7rn9, and 7seo.

The authors declare no competing financial interest.