Synthesis of Novel Caspase Inhibitors for Characterization of the Active Caspase Proteome in Vitro and in Vivo

Abstract

Caspases are cysteine proteases that are essential for cytokine maturation and apoptosis. To facilitate the dissection of caspase function in vitro and in vivo, we have synthesized irreversible caspase inhibitors with biotin attached via linker arms of various lengths (12a–d) and a 2,4-dinitrophenyl labeled inhibitor (13). Affinity labeling of apoptotic extracts followed by blotting reveals that these affinity probes detect active caspases. Using the strong affinity of avidin for biotin, we have isolated affinity-labeled caspase-6 from apoptotic cytosolic extracts of cells overexpressing procaspase 6 by treatment with 12c, which contains biotin attached to the N^ε-lysine of the inhibitor by a 22.5 Å linker arm, followed by affinity purification on monomeric avidin-Sepharose beads. 13 has proven sufficiently cell permeable to rescue cells from apoptotic execution. These novel caspase inhibitors should provide powerful probes for the study of the active caspase proteome during apoptosis both in vitro and in vivo.

INTRODUCTION

Apoptosis is a highly conserved process by which eukaryotic cells commit suicide.^1,2 This form of programmed cell death involves a reproducible series of cellular changes that include cell shrinkage, chromatin condensation, membrane blebbing and, in most cases, DNA fragmentation. Studies performed over the past decade have revealed that many of the changes observed in apoptotic cells result from the action of a family of cysteine-dependent aspartate-directed proteases termed caspases.^3–5 The caspase family can be divided into two distinct subfamilies, the cytokine activators (caspases 1, 4, 5, 11 and 12) and those involved in apoptosis. Apoptotic caspases can be further subdivided into initiator caspases (caspases 2, 8, 9 and 10), which transduce an apoptotic signal into proteolytic activity, and effector caspases (caspases 3, 6 and 7), which are activated by initiator caspases and are in turn responsible for the cleavage of the majority of the substrates leading to the demise of the cell.^3–5

Previous studies have demonstrated that the various caspases differ in their substrate specificies.^4,6 While all cleave on the carboxyl side of aspartate residues, caspases 3 and 7 prefer acidic residues in the P₄ position of the substrate, whereas caspases 1, 4, 5, 6 and 8–10 prefer aromatic or hydrophobic residues in this position.⁷ These substrate specificities have been explained, at least in part, by differences in the corresponding S₄ substrate binding pockets of the various caspase family members as crystal structures have become available.^8–13

Earlier progress in understanding the role of caspases in apoptosis was critically dependent on the availability of caspase inhibitors and affinity labels. After the demonstration that apoptotic cleavage of poly(ADP-ribose) polymerase (PARP1)^14,15 occurs at the sequence asp-glu-val-asp↓gly,¹⁶ an observation that implicated caspases in apoptotic cleavage, the availability of the peptide caspase 1 inhibitor tyr-val-ala-asp-chloromethylketone (YVAD-cmk)^a allowed demonstration of the critical role of this class of enzymes in apoptotic events.¹⁶ Moreover, biotinylated derivatives of YVAD-cmk and of benzyloxycarbonyl-glu-lys-asp-acyloxymethylketone (Z-EKD-aomk), an inhibitor designed with apoptotic effector cleavage sequences in mind,¹⁷ were successfully utilized to identify multiple species of caspases 3 and 6 as the predominant active caspases in apoptotic cells^17,18 despite the fact that the YVAD and EVD peptide sequences were subsequently shown to bind caspase 1 more avidly than caspases 3, 6 and 7.^19,20 This ability of covalent caspase inhibitors to derivatize apoptotic caspases despite less than optimal affinity of their peptide moieties for the enzyme active sites has been attributed to the high concentrations and prolonged incubation times often utilized in biological experiments.¹⁹

Although the previous chloromethylketone (cmk) and acyloxymethylketone (aomk) affinity labels have proven useful for affinity labeling of cell-free extracts prepared from apoptotic cells, these reagents also had several limitations. First, because of the short arm linking biotin to the label, they were difficult to use for affinity purification of caspases. Instead, because the biotin did not extend far beyond the active site cleft of the enzymes, the earlier affinity labels were most useful for detection on blots after polypeptides were denatured. Second, because biotin was used as a tag, endogenous biotinylated polypeptides were also detected on blots. Third, the previous affinity labels penetrated cells poorly and were generally used to label caspases under cell-free conditions,^{17,18,21–23} although it has been reported that caspases in intact cells can sometimes be labeled when cells are subjected to prolonged incubations with these compounds prior to introduction of an apoptotic stimulus.^24,25

In view of the large number of caspase species that can be detected when samples of affinity labeled caspases are separated by charge and size on two-dimensional polyacrylamide gels,^17,18,21 it is clear that further work is required to fully characterize the posttranslational modifications that regulate caspase activation and activity. Here we describe the synthesis and initial characterization of several novel irreversible acyloxymethylketone caspase inhibitors labeled with biotin 12a–d or 2,4-dinitrophenyl (DNP) 13 that overcome some of the limitations of earlier inhibitors. The construction of affinity probes in which the biotin was coupled via an extended linker arm enabled the affinity purification of nondenatured enzyme from cell-free extracts prepared from apoptotic cells overexpressing caspase-6 fused to enhanced green fluorescent protein (EGFP). The probe 13 labeled with DNP, rather than biotin, not only displayed much lower background labeling of polypeptides in cell-free extracts, but also penetrated cells and efficiently inhibited apoptosis in situ.

CHEMISTRY

To provide an affinity label that could be readily tagged with various moieties, the peptidyl acyloxymethylketone inhibitor Z-EKD-aomk (11) was synthesized as shown in Scheme 1. D-glyceraldehyde acetonide 1 was prepared from D-mannitol.²⁶ Wittig reaction of the chiral acetonide with Ph₃P=CHCO₂CH₂CH₃ in methanol at 0°C gave a separable isomeric mixture (Z/E, 8:1) of the α,β-unsaturated esters 2.²⁷ Reaction of the mixture of esters with benzylamine in the absence of solvent at −50°C for 2 days afforded the (3R)-benzylamino ester 3. Removal of the Z protecting group by hydrogenation furnished the amine 4, which was coupled with N-α-benzyloxycarbonyl-N-ε-butoxycarbonyl-lysine-succinyl ester to give the dipeptide 5. Hydrogenation of 5 afforded amine 6, which was coupled with N-α-benzyloxycarbonyl-γ-O-tert-butyl-glutamyl-succinyl ester to give the protected tripeptide analog 7. Selective acid hydrolysis of the acetonide group of 7 to give the diol 8, without partial removal of the ^tbutyl protecting groups, could not be achieved even under relatively mild solution conditions²⁸ but was effected in good yield by treatment with Montmorillonite K10 in aqueous ethanol. Esterification of the primary hydroxyl of 8 with 2,6-dimethylbenzoyl chloride in pyridine/dimethylaminopyridine/DMPU at room temperature produced the ester 9. Attempts to effect oxidation of the sterically-crowded secondary alcohol group of 9 with a variety of mild oxidizing reagents proved unsuccessful. However, oxidation under Dess-Martin conditions²⁹ afforded the ketone 10 in virtually quantitative yield. Finally, cleavage of the protecting groups of the glutamyl, lysyl and ‘aspartyl analog’ moieties of 10 was achieved with TFA containing a catalytic amount of water to give the desired Z-EKD-aomk 11. A series of biotin-labelled derivatives of 11 (compounds 12a–12d) was then prepared by reaction of the free ε-amino group of lysine with the N-hydroxysuccinamides of biotin, iminobiotin, 6-[biotinamido]hexanoate (LC biotin) and 6-[biotinamido]-6-hexanamidohexanoate (LCLC biotin). These have ‘spacer arms’ between the N^ε-lysine of the inhibitor and the avidin-binding ureido ring nucleus of biotin of ~13.5, 13.5, 22.5 and 30.0 Å, respectively. The N-ε-2,4-dinitrophenyl affinity probe 13 was prepared by reaction of 11 with 2,4-dinitrofluorobenzene.

Scheme 1. Synthesis of caspase inhibitors.

a) Ph₃P=CHCO₂CH₂CH₃, MeOH, 0°C. b) Benzylamine, −50°C, 48 h. c) H₂, Pd/C, EtOH. d) Z-Lys(Boc) N-hydroxysuccinimide, CH₂Cl₂, 12 h. e) H₂, Pd/C, EtOH. f) Z-Glu(OtBu) N-hydroxysuccinimide, CH₂Cl₂, 12 h. g) Montmorillonite, EtOH/H₂O (5:1), 75°C, 3 h. h) 2,6-dimethylbenzoylchloride, pyridine, DMAP, DMPU, 48 h. i) Dess-Martin periodate, 0°C to rt., 4 h. j) 95% TFA/H₂O, 12 h. k) labelling reagent, water, 12 h.

RESULTS

Synthesis of novel caspase inhibitors

Peptidyl halo- and acyloxymethylketone inhibitors of caspases are notoriously difficult to prepare; and a number of commercially available preparations of these compounds exhibited variable chemical composition and biological activity upon testing in our laboratory. This variability appears to reflect, at least in part, instability of the 3-aminoacyl-4-oxopentanoic acid (‘aspartyl’) unit substituted with a good leaving group in the 5 position, which undergoes a number of reactions, including epimerization and facile cyclization to the corresponding γ-lactone, during chemical syntheses (Fig. 1). Accordingly, initial attempts to employ a linear synthetic strategy to the target compound 11 via the protected acyloxymethylketone 14 proved unsuccessful. While this intermediate was accessible, albeit in modest yield, by a classical diazomethylketone route from Z-aspartic acid,³⁰ the inherent instability of the acyloxymethylketone group gave rise to significant problems during the subsequent coupling and deprotection steps in the synthesis. As a result, we elected to use a new synthetic approach in which the ‘aspartyl’ acyloxymethylketone subunit of the molecule is created only after the peptide bond forming steps of the synthesis are complete. Using this strategy, the tagged inhibitors 12a–d and 13 could all be prepared in >95% purity.

Figure 1. Epimerization and cyclization of ‘aspartyl’ halo- and acyloxy- methylketone analogs.

R = H or acyl, X = halogen or acyloxy.

The novel probes inhibit caspases 1–8

Synthetic colorimetric substrates containing a cleavage sequence similar or identical to the preferred tetrapeptide recognition sequence of each caspase are commercially available to measure enzyme activity. Using standard spectrophotometric methods, kinetic parameters for the cleavage of these substrates catalyzed by caspases 1–8 were determined (Table 1). The v_max and K_m values for the p-nitroaniline (pNA) substrates are in good agreement with those reported previously.^19,31,32 Assays of caspases 9 and 10 yielded K_m values that were outside the range of substrate concentrations tested and are not further considered here.

Table 1.

Kinetic parameters determined for human caspases 1–8 reacted with LC biotin-derivatized inhibitor 12c

Caspase	Substrate	Vmax (pmol/min)	Km (μM)	Ki (μM)
1	YVAD-pNA	110	100	0.0063
2	LEHD-pNA	400	560	29
3	DEVD-pNA	110	74	0.84
4	LEHD-pNA	510	1100	NT
5	LEHD-pNA	610	2500	NT
6	VEID-pNA	230	220	550
7	DEVD-pNA	530	230	1.2
8	IETD-pNA	130	140	0.54

The inhibition constants (K_i) for 12c when added to caspases 1–8 were evaluated using steady state kinetic methods under the assumption of a slow, tight binding inhibitor (Table 1). 12c inhibited all tested caspases (Table 1 and Fig. 2A). Consistent with a previous observation showing that removal of the N-terminal aspartate from the DEVD sequence diminishes the affinity for caspases 3 and 7 much more dramatically than for caspases 1 and 4,¹⁹ 12c had a lower K_i for caspase 1 than for caspases 3, 7 and 8. Nonetheless, submicromolar to low micromolar values for were calculated for the K_i for these apoptotic caspases (Table 1 and Fig. 2A). As was the case with covalent inhibitors based on YVAD or Z-VAD,²⁰ 12c was not particularly potent at inhibiting caspase 6, with a calculated K_i of 550 μM. 12a, 12b, 12d and 13 were confirmed to have similar potencies by steady state kinetics with caspase-3 (Table 2 and Fig. 2B). The similar potencies of these various derivatives are consistent with previous reports that the P₂ sidechain of substrate analogues points away from the caspase active site, permitting extensive derivatization at this site without altering affinity for the enzyme.^10,33

Figure 2. Michaelis-Menton graphs of caspase 3-catalyzed Ac-gDEVD-pNA cleavage in the absence and presence of peptidyl aomk inhibitors.

A) Biotinylated inhibitor 12a, B) DNP labelled inhibitor 13. Inhibitor concentrations: closed circle, 0 μM; open circle, 0.1 μM; inverted closed triangle, 1.0 μM; inverted open triangle, 10 μM; closed square, 100 μM.

Table 2.

Inhibition of Caspase 3 by Substituted Peptidyl aomk Inhibitors

Label on Nε-lysine of 11	Vmax (pmol/min)	Km (μM)	Ki (μM)
βbiotin, 12a	140	72	1.4
−iminobiotin, 12b	140	70	1.5
6-[biotinamido]hexanoate (LC-biotin), 12c	140	52	0.79
6-[biotinamido]-6-hexanamidohexanoate (LCLC-biotin), 12d	150	79	1.1
DNP, 13	150	75	1.0

The novel inhibitors detect active caspases on blots

Blotting with horseradish peroxidase (HRP) –coupled streptavidin was used to evaluate the ability of the synthesized inhibitors to affinity label active caspases in extracts from apoptotic Jurkat cells. In this way it was shown that all derivatized inhibitors are able to bind covalently to active caspases in vitro (Fig. 3A, B). Pretreatment of extracts with the commercially available caspase inhibitor DEVD-aomk as a control abolished the labelling by the inhibitors, thus confirming the detected bands to be active caspases labelled with the various affinity probes.

Figure 3. Use of new affinity labels for blotting.

After preincubation with diluent or DEVD-aomk, apoptotic cell lysates were reacted with the DNP labeled caspase probe 13 (A) or biotinylated probe 12c (B), subjected to SDS-polyacrylamide gel electrophoresis, transferred to PVDF membrane and probed with anti-DNP antibody (A) or HRP-coupled streptavidin (B).

Affinity purification of native caspase 6 using the novel inhibitors

One potential use of inhibitors 12c, 12d and 13 is as affinity tags to allow purification from cell extracts of sufficient quantities of caspases for further studies (e.g., the analysis of posttranslational modifications). This use is illustrated for 12c, which was employed to purify caspase 6-EGFP from apoptotic DT40 chicken lymphoma cells that were engineered as previously described³⁴ to overexpress a chicken procaspase 6-EGFP fusion protein (Fig. 4). To avoid the copurification of endogenous biotinylated proteins, apoptotic extracts were first passed over an avidin column to remove the bulk of endogenous biotinylated polypeptides. Extracts were then reacted with the biotinylated probe 12c, dialyzed to remove excess probe, and passed over monomeric avidin, which bound the affinity-labeled caspases. After a series of increasingly stringent washes, caspases were eluted with D-biotin as a highly enriched fraction, subjected to sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAG)E and stained with Coomassie blue (Fig. 4A). This test-of-concept experiment was performed on cells that overexpressed caspase 6 with the idea that successful recovery of EGFP-caspase 6, which has a somewhat low affinity for the inhibitor (Table 1), would indicate its potential usefulness with other caspases in the future.

Figure 4. Affinity purification of caspase 6-EGFP fusion protein from cell extracts.

(A) SDS-PAGE gel stained with Colloidal Coomassie Blue showing the fraction eluted with biotin. (B) Immunoblot characterization of the biotin eluted fraction. After SDS-PAGE, proteins were transferred to a nitrocellulose membrane and reacted with: lane 1, streptavidin-HRP; lane 2, anti-chicken caspase-6; and lane 3, anti-EGFP antibody. The blot was subjected to a stripping procedure⁴⁴ between detection reagents.

The affinity purified caspase fraction contained four major bands with estimated molecular weights of 18, 28, 42.5 and 44 kDa (Fig. 4A). In order to identify these, a duplicate SDS-polyacrylamide gel was transferred to a nitrocellulose membrane, probed with HRP-coupled streptavidin, stripped, probed with rabbit anti-chicken caspase-6 antiserum, stripped again and probed with anti-EGFP antiserum (Fig. 4B). Band 1 bound HRP-streptavidin very strongly. Bands 1, 3 and 4 were also recognized by anti-chicken caspase-6, whereas bands 2 and 4 were recognized by anti-EGFP. These results suggested that band 1 corresponded to the large subunit of caspase 6, band 2 to EGFP, and bands 3 and 4 to the EGFP-caspase 6 fusion species. To verify this identification, bands were excised, digested with trypsin and subjected to Matrix Assisted Laser Desorption/Ionization- Time of Flight (MALD-ToF) mass spectrometry. Results of this analysis confirmed the band identifications deduced by immunoblotting (Table 3). Thus, this experiment demonstrates the utility of this affinity label for the isolation of caspases from cell extracts under nondenaturing conditions.

Table 3.

Identification of polypeptides in biotin eluted fraction by MALDI-ToF mass spectrometry

Band #	App. Mwa	Protein ID	GenBank #	Mw	Mw total	MOWSEb	Cover(%)
1	18 kDa	Caspase 6	AF469049	18167		4.11E+04	46
2	28 kDa	GFP	P42212	26887		5.38E+08	49
3	38 kDa	Caspase 6	AF469049	18167	45053	1.79E+03	17
		GFP	P42212	26887		1.52E+07	45
4	40 kDa	Caspase 6	AF469049	18167	45053	2.54E+04	22
		GFP	P42212	26887		1.52E+07	45

^aApparent masses based on the SDS preparative gel from which the band was excised.

^bMolecular weight search s(MOWSE) scores calculated according to the method of Pappin et al. ⁴⁸ using the MS-Fit tool.

A DNP-derivatized probe is a potent inhibitor of apoptosis in intact cells

Although Z-VAD(OMe)-fmk and Z-EK(biotin)D-aomk 12a have a similar mode of action, i.e., reaction with the sulfhydryl group of the active site cysteine, Z-VAD(OMe)-fmk is sufficiently lipophilic to cross the plasma membrane and inhibit caspases in situ, whereas the more polar Z-EK(biotin)D-aomk is not cell-permeable. We hypothesized that the substitution of an aromatic DNP group for biotin to yield 13 might increase the lipophilicity of this class of inhibitors, thus facilitating passage through the plasma membrane and inhibition of apoptosis in intact cells. To address this possibility, we treated Jurkat cells with etoposide (10 μM) for 5 h in the presence or absence of varying concentrations of 13. Apoptotic cells were detected by TUNEL labeling and analyzed by flow cytometry (Fig. 5). As a control, the level of apoptosis inhibition was compared to that achieved with cell permeable Z-VAD(OMe)-fmk (10 μM). 13 at 1 μM and 10 μM was found to essentially abolish apoptosis in etoposide-treated Jurkat cells. These data demonstrate that 13 is a cell permeable inhibitor of caspases and apoptosis in human cells.

Figure 5. 13 rescues Jurkat cells from apoptosis induced by etoposide.

Jurkat cells were treated with etoposide with or without 13 as indicated for 5 h. Cells were then stained with TUNEL reagent and analyzed by flow cytometry to assess the level of cell death. A total of 20,000 events were analyzed for each curve.

DISCUSSION

The human genome encodes at least eleven different caspase gene products, but many more active caspase species are detected when apoptotic cell lysates are incubated with affinity probes such as 12a or biofinylated YVAD-cmk and fractionated by two-dimensional gel electrophoresis.^17,18 In fact, this active caspase proteome may comprise 30 or more species in certain cell types.³⁵ In two studies where caspases were labeled, it was shown that many of the spots in the two-dimensional gels were derived from caspases-3 and −6, but other spots remained unidentified.^17,18 Many of these different active caspase species are likely to result from differential processing of the proenzymes, phosphorylation,^35–37 nitrosylation^38–41 and possibly other post-translational modifications. In order to characterize these modifications, it will be important to be able to purify caspases away from more abundant cellular polypeptides.

One approach to this is the isolation of caspases tagged at their C-termini with peptide tags. This has two potential drawbacks, however. First, unlike affinity labeling, such methods do not distinguish between active caspases and their procaspase precursors. Second, there is the possibility that tagging will alter the posttranslational modifications. We have observed, for example, that tagged caspase-7 no longer exhibits a prominent posttranslational modification that is present in the wild-type enzyme (N. Korfali, X.W. Meng, WCE and SHK, unpublished). These considerations prompted development of additional caspase affinity labeling reagents.

In developing these new agents, we have attempted to address some of the deficiencies of previous reagents. Several of the previously available caspase affinity labeling reagents have proven difficult to use during affinity purification. For example, caspases affinity labeled with 12a quantitatively flow through avidin- and streptavidin-agarose columns under nondenaturating conditions (F. Durrieu and WCE, unpublished observations), apparently because the covalently bound inhibitor occupies a deep pocket shielded by a “flap” of protein that blocks access to the biotin residue.^10,33 In fact, affinity purification of the large subunits of active caspases was possible with 12a if the labeled enzymes were first denatured in hot SDS prior to passage over the affinity column (F. Durrieu and WCE, unpublished observations), suggesting that efficient purification of native caspases might be possible with a new generation of inhibitor in which the biotin was linked to the P₂ lysine residue by an extended linker arm.

We were also aware that the repertoire of caspase affinity labeling reagents was limited, at least in part, by fact that synthesis of peptide acyloxymethyl ketone inhibitors is challenging. The synthetic route described here provides a convenient approach for the preparation of a new generation of irreversible caspase inhibitors to which a variety of affinity tags can be attached. This route not only results in a relatively high yield, but also avoids the cyclization reaction between the aspartate sidechain and the acyloxymethylketone moiety that have plagued previous syntheses. As a result, purity of the inhibitors prepared by this route is >95%.

This synthetic route provided a suitable opportunity to ask whether biotinylated inhibitors with longer aliphatic arms linking the biotin could be useful for the affinity purification of caspases under nondenaturing conditions. Accordingly, we synthesized the acyloxymethylketone inhibitor Z-EKD-aomk in which the ε-amino group of the lysine residue was linked to biotin by spacers of 13.5, 22.5 and 30 Å (12a–d). As predicted from the fact that the P₂ residue extends away from the enzyme active site, the recognition of these affinity probes by caspases did not appear to be influenced by the length of the linker arm (Table 2).

The present inhibitors were based on the Z-EKD peptide sequence previously shown to allow derivatization of caspases 3 and 6 in apoptotic extracts.¹⁷ These new inhibitors displayed a range of affinity constants (Table 1) similar to that previously reported for EVD-based inhibitors,¹⁹ with high affinity for caspase 1, intermediate affinity for caspases 3, 7 and 8, and somewhat lower affinity for caspase 6. These differences reflect differences in geometry and chemical composition at the S₃ and S₄ subsites of the various enzymes.^8–12 The present compounds were not designed to alter this spectrum of affinities, but to instead test the hypothesis that extended linkers between the lysine moiety and biotin would allow affinity purification under nondenaturing conditions and that substitution of the hydrophobic DNP moiety would allow penetration of intact cells in sufficient concentrations to inhibit apoptosis.

As expected based on our prior unpublished results, when we synthesized Z-EKD-aomk and linked biotin directly to the ε-amino group of the lysine, the resulting 12a efficiently labeled active caspases as detected in immunoblots, but these enzymes failed to bind avidin under native conditions. In contrast, caspases bound to inhibitor 12c, in which the biotin was linked to the ε-amino group of the lysine by a longer linker arm, could bind to avidin under native conditions. As proof of principle, we successfully used 12c to purify nondenatured EGFP-caspase-6 from cell-free apoptotic extracts on monomeric avidin. This protocol allowed us to wash the beads relatively stringently but elute the bound enzyme under mild conditions with D-biotin. Characterization of the purified fraction by both immunoblotting and MALDI-ToF mass spectrometry confirmed the presence of active caspase-6 as a major constituent of the fraction. Even though the affinity of the peptide moiety for caspase 6 was less then optimal (Table 1), prolonged incubation at micromolar concentrations was able to overcome this obstacle, as had been previously reported for other covalent caspase inhibitors.¹⁹ Accordingly, we expect that 12c will also be useful for purification of caspases for which it has higher affinity. Thus, this reagent should prove useful in future studies aimed at characterization of the active caspase proteome by affinity purification of labeled caspases from apoptotic cytosolic extracts and identification of the relevant polypeptides and their post-translational modifications by mass-spectrometric analysis. In addition, because denaturation is not required, this reagent should be useful for purifying polypeptides that are associated with active caspases.

Another problem with characterization of the active caspase proteome is that the previously described affinity-labeling agents were poorly cell permeable and were generally utilized in cell-free extracts.^17,18,21,23 Because effector caspases are activated both by themselves and by initiator caspases, this raises the possibility that some of the active species observed by this technique might have been activated only following cell homogenization, when inactive caspases sequestered in various subcellular compartments would be exposed to active caspases from other compartments. In Jurkat cells that had been exposed to etoposide, striking inhibition of the onset of apoptosis was observed with 13. Therefore, substitution of the DNP moiety for biotin (12a–d) appears to enable 13 to cross the plasma membrane, where it can then inhibit cytosolic caspases and blunt the apoptotic response. This means that 13 can be used to label active caspases in situ, thereby avoiding concerns about caspase activation during cell lysis. Compared to biotinylated VAD-fmk, which has also recently been utilized for this purpose,^24,25 13 has the advantage of not requiring preloading prior to administration of an apoptotic stimulus. Moreover, 13 lacks a biotin moiety, thereby diminishing contamination of any pull-downs by endogenous biotinylated proteins,^17,25 and instead contains the widely studied hapten DNP, for which immunological reagents are commercially available.

In conclusion, the reagents described here comprise a new generation of small-molecule caspase inhibitors that can be used for experiments aimed to characterize the active caspase proteome in apoptotic cells. Identification of the members of this proteome will be an important step toward understanding the global regulation of the apoptotic response by protein kinases, phosphatases and other enzyme modification systems in different normal and transformed cell populations, in response to different stimuli, and when the apoptotic cascade is modulated by the simultaneous up- or down-regulation of other cytosolic signaling cascades.

EXPERIMENTAL SECTION

Chemistry

Reagents were obtained from the following suppliers: D-mannitol, Ph₃P=CHCO₂CH₂CH₃, benzylamine, palladium on charcoal, montmorillonite K10, 2,6-dimethylbenzoyl chloride, pyridine, dimethylaminopyridine, 1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone, Dess-Martin periodinane, trifluoroacetic acid (TFA) and 2,4-dinitrofluorobenzene from Sigma-Aldrich (Poole, UK); N-α-benzyloxycarbonyl-N-ε-butoxycarbonyl-lysine-succinyl ester and N-α-benzyloxycarbonyl-γ-O-tert-butyl-glutamyl-succinyl ester from Bachem AG (Bubendorf, Switzerland); biotin-NHS, iminobiotin-NHS, LC-biotin-NHS and LC-LC-biotin-NHS from Pierce (Rockford, IL). In synthetic procedures, flash column chromatography was carried out over silica 60 (230–400 mesh), analytical thin layer chromatography was carried out on 0.25 mm silica gel plates (Fisher, Loughborough, UK) and organic extracts were routinely dried over MgSO₄, filtered and concentrated in vacuo. Purifications were carried out using a Waters 600 HPLC (Milford, MA) coupled to a Waters 486 detector. A Lichrosorb 100 RP-8 column (7 x 250 mm, 5 μm) was used for preparative HPLC (Jones Chromatography, Hengoed, UK).

Nuclear magnetic resonance (NMR) spectra were recorded on Varian Gemini 200 (Palo Alto, CA) and Bruker AC 250, 360 or 600 spectrometers (Karlsruhe, Germany). Fast Atom Bombardment (FAB) mass spectra were recorded on a Kratos MS 50 TC (Manchester, UK) instrument. HPLC-MS analysis was accomplished by electrospray ionization (EI) and analysis on a MicroMass platform II spectrometer (Manchester, UK) after separation on a Waters Alliance 2795 HPLC using a 10 μm C-18 column (1 x 150 mm) eluted isocratically with AcCN:H₂O:HCOOH (2:98:0.1) for 2 min followed by a gradient to 60% AcCN:H₂O:HCOOH (97:3:0.05) over 20 min (flow rate of 1 ml/min). [(1R)-1-[(4S)-2,2-dimethyl-1,3-dioxolan-4-yl]-2-carboxyethyl]ethylamine (4). (4R) -(2) (400.5 mg, 2.0 mmole), prepared from D-glyceraldehyde acetonide by the procedure of Matsunaga et al.,²⁷ was stirred with benzylamine (437 μl, 429 mg, 4.0 mmole) at −50ºC for 48 h in the absence of solvent. The reaction mixture was dissolved in EtOAc (25 ml), washed with water, dried, and concentrated. Flash column chromatography eluting with EtOAc:hexane (20:1) afforded N-Benzyl-[(1R)-1-[(4S)-2,2-dimethyl-1,3-dioxolan-4-yl]-2-carboxyethyl]ethylamine (3, 460 mg, 75%) as a colorless oil. R_f (5% MeOH/CH₂Cl₂) 0.70; MS (FAB, thioglycerol) m/z 308.1859 (MH+, C₁₇H₂₆NO₄ requires 308.1862); ¹H-NMR (CDCl₃): 1.24 (t, 3H, CH₂CH₃), 1.35 (s, 3H, CH₃), 1.36 (s, 3H, CH₃), 1.82 (br. s, 1H, NH), 2.46 (m, 2H, CH₂CO₂Et, 3.13 (q, 1H NCH), 3.81 (t, 2H, OCH₂CH₃), 4.00 (m, 2H, CH₂Ar), 4.12 (m, 3H, CHO + CH₂O) and 7.30 (m, 5H, Ar); ¹³C-NMR (CDCl₃): 13.4 (OCH₂CH₃), 24.4 (CH₃), 25.7 (CH₃), 35.4 (CH₂CO₂Et), 50.7 (NCH), 55.1 (CH₂Ar), 59.8 (OCH₂CH₃), 65.5 (CH₂OC), 75.8 (CHOC), 108.5 (-OCO-), 126.3, 127.5, 127.7 and 139.7 (ArC) and 171.4 (CO₂Et).

Hydrogenation of 3 (460 mg, 1.50 mmole) over 10% Pd/C in anhydrous EtOH (25 ml) under atmospheric pressure for 12 h, removal of the catalyst by filtration and purification by flash column chromatography eluting with 1–10% acetone/CH₂Cl₂ gradient afforded 4 (279 mg, 86%) as a colorless oil. R_f (5% MeOH/CH₂Cl₂) 0.25; Found: C 54.79%, H 8.94%, N 6.34% (calc for C₁₀H₁₉NO₄: C 55.28%, H 8.81%, N 6.45%); MS (FAB, NOBA) m/z 218.1393 (MH⁺, C₁₀H₂₀NO₄ requires 218.1392); ¹H-NMR (CDCl₃): 1.19 (t, 3H, OCH₂CH₃), 1.27 (s, 3H, CH₃), 1.35 (s, 3H, CH₃), 2.02 (s, 2H, NH₂, 2.33 (m, 2H, CH₂CO₂Et), 3.13 (m, 1H, NCH), 3.68 (m, 1H, CHO), 3.99 (m, 2H, OCH₂CH₃) and 4.08 (m, 2H, CH₂O);¹³C-NMR (CDCl₃): 12.9 (OCH₂CH₃), 24.3, 25.5 (2xCH₃), 37.6 (CH₂CO₂Et), 49.2 (CHN), 59.3 (OCH₂CH₃), 65.0 (CH₂OC), 75.6 (CHOC), 108.0 (-OCO-), 170.6 (CO₂)

N-[N-((R)-Nα-Benzyloxycarbonyl-Oγ-tert-butyl-glutamyl)-(R)-Nε-tert-butoxycarbonyllysyl]-[(1R)-1-[(4S)-2,2-dimethyl-1,3-dioxolan-4-yl]-2-carboxyethyl]ethylamine (7)

A solution of 4 (279 mg, 1.28 mmole) in 2 ml dry CH₂Cl₂ was added to a solution of (R) N-α-benzyloxycarbonyl-N-ε-tert-butoxycarbonyl-lysine N-hydroxysuccinimide ester (673 mg, 1.41 mmol) in dry CH₂Cl₂ (10 ml) and the mixture stirred overnight at room temperature. The reaction mixture was diluted with CH₂Cl₂ (10 ml), washed with H₂O (10 ml), dried and concentrated. Column chromatography using 0–10% acetone in CH₂Cl₂ as eluent afforded N-[(R)-N^α-benzyloxycarbonyl-N^ε-tert-butoxycarbonyl-lysyl]-[(1R)-1-[(4S)-2,2-dimethyl-1,3-dioxolan-4-yl]-2-carboxyethyl]ethylamine (5, 632 mg, 85%) as a colorless oil. R_f (5% MeOH/CH₂Cl₂) 0.45; MS (FAB, NOBA) m/z 580.3234 (MH⁺, C₂₉H₄₆N₃O₉ requires 580.3235); ¹³C-NMR (CDCl₃): 13.3 (OCH₂CH₃), 21.7 (Kγ-CH₂), 24.0 (Kβ-CH₂), 25.4 (2 x CH₃), 27.7 (^tBu-CH₃), 28.9 (Kδ-CH₂), 36.1 (CH₂CO₂Et), 35.9 (Kε-CH₂), 45.9, 54.5 (2xCHN), 60.1 (OCH₂CH₃), 65.2 (CH₂OC), 66.2 (C(CH₂)₃), 75.5 (CHOC), 78.2 (CH₂Ph), 108.6 (-OCO-), 127.4, 127.8, 135.7 (ArC), 155.6 (CONH), 170.3 and 171.4 (2xCO₂). Hydrogenation of 5 (630 mg, 1.09 mmole) with H₂ over 10% Pd/C in anhydrous EtOH (25 ml) under atmospheric pressure for 3–4 h gave the crude amine after filtration and concentration. Column chromatography eluting with a 1–10% acetone/CH₂Cl₂ gradient gave N-[(R)-Nε-tert-butoxycarbonyl-lysyl]-[(1R)-1-[(4S)-2,2-dimethyl-1,3-dioxolan-4-yl]-2-carboxyethyl]ethylamine (6, 446 mg, 92%) as a colorless oil. R_f (10% MeOH/CH₂Cl₂) 0.55; MS (FAB, NOBA) m/z 446.2866 (MH⁺, C₂₁H₄₀N₃O₇ requires 446.2866); ¹³C-NMR (CDCl₃): 13.3 (OCH₂CH₃), 21.7 (Kγ-C H₂), 24.0 (Kβ-CH₂), 25.4 (2 x CH₃), 27.6 (^tBu-CH₃), 28.8 (Kδ-CH₂), 36.1 (CH₂CO₂Et), 39.2 (Kε-CH₂), 45.9 (CHN), 53.9 (CHN), 59.8 (OCH₂CH₃), 65.2 (CH₂OC), 75.4 (CHOC), 77.8 (CH₂Ph), 108.5 (-OCO-), 155.4 (CONH), 170.2 and 173.3 (2 x C O₂). To a solution of N-α-benzyloxycarbonyl-γ-O-tert-butyl-glutamyl N-hydroxysuccinimide ester (478 g, 1.1 mmol) in dry CH₂Cl₂ (10 ml), 6 (440 mg, 0.99 mmole) was added and the mixture stirred overnight under dry N₂ at room temperature. The reaction mixture was diluted with CH₂Cl₂ (10 ml) washed with H₂O, dried, and concentrated. Column chromatography using a 0–20% acetone/CH₂Cl₂ gradient afforded 7 (620 mg, 81%) as a colorless amorphous solid (mp 62–63°C). R_f (10% MeOH/CH₂Cl₂) 0.60; Found: C 59.28%, H 7.93%, N 7.32% (calc for C₃₈H₆₀N₄O₁₂: C 59.67%, H 7.91%, N 7.09%); MS (FAB, NOBA) m/z 765.4286 (MH⁺, C₃₈H₆₁N₄O₁₂ requires 765.4285); ¹³C-NMR (CDCl₃): 13.9 (OCH₂CH₃), 22.3 (Kγ-CH₂), 24,2, 25.8 (2 x CH₃), 27.3 (^tBu-CH₃), 27.9 (^tBu-CH₃), 28.3, 29.1, 29.2(3 x CH₂), 36.5 (CH₂CO₂Et), 39.9 (Kε-CH₂), 48.2, 53.1, 54.6 (3 x CHN), 60.6 (OCH₂CH₃), 65.9 (CH₂OC), 66.9 (OC(CH₃)₃), 75.7 (CHOC), 80.9 (CH₂Ph), 107.9 (-OCO-), 120.5-136.0 (ArC), 155.9, 156.0 (2 x CONH), 170.7 , 170.9, 171.2 and 172.6 (4 x CO₂)

N-[N-((R)-Nα-Benzyloxycarbonyl-Oγ-tert-butyl-glutamyl)-(R)-Nε-tert-butoxycarbonyl-lysyl]-(1R,2R)-1-(carboxyethylmethyl)-2,3-dihydroxypropylamine (8)

A solution of 7 (100 mg, 0.13 mmole) in EtOH:H₂O (5:1, 10 ml) was stirred with Montmorillonite K10 at 75ºC for 3 h. The suspension was diluted with EtOAc (20 ml), filtered through Celite, washed with H₂O, dried, and concentrated. Flash column chromatography eluting with 0–50% acetone/CH₂Cl₂ yielded 8 (66 mg, 70%) as a colorless oil. R_f (10% MeOH/CH₂Cl₂) 0.55; Found: C 55.38%, H 7.79%, N 7.30% (calc for C₃₅H₅₆N₄O₁₂.2H₂O: C 55.25%, H 7.95%, N 7.36%); MS (FAB, NOBA) m/z 725.3987 (MH⁺, C₃₅H₅₇N₄O₁₂ requires 725.3974); ¹³C-NMR (CDCl₃): 13.4 (OCH₂CH₃), 21.9 (KγCH₂), 27.3 (^tBu-CH₃), 27.7 (^tBu-CH₃), 28.6 (2 x CH₂, Eβ and Kβ), 30.9 (Eγ-CH₂), 33.9 (Kδ-CH₂), 35.9 (CH₂CO₂Et), 39.7 (Kε-CH₂), 47.1, 52.8, 53.9 (3xCHN), 60.6 (OCH₂CH₃), 62.5 (CH₂OH), 66.4 (OC(CH₃)₃), 72.2 (CHOH), 80.4 (CH₂Ph), 127.4, 127.8, 135.6 (ArC), 155.9 (2 x CONH), 170.9, 171.8, 172.1 and 172.5 (4 x CO₂)

3-[N-((R)-Nα-Benzyloxycarbonyl-glutamyl)-(R)-lysyl]amino-(3R)-5-(2,6-dimethylbenzoyl)-4-oxopentanoic acid (Z-EKD-aomk,11)

To a stirring solution of DMAP (22 mg, 0.18 mmole), DMPU (23 mg, 21 μl, 0.18 mmole) and 2,6-dimethylbenzoyl chloride (17 mg, 0.10 mmole) in dry pyridine (5 ml), 8 (66 mg, 91 μmole) was added and the solution was stirred under dry Ar at room temperature for 3 d. The reaction mixture was diluted with CH₂Cl₂ (10 ml), washed with H₂O, dried, and concentrated. Column chromatography eluting with a 0–20% acetone/CH₂Cl₂ gradient gave N-[N-((R)-Nα-Benzyloxycarbonyl-Oγ-tert-butyl-glutamyl)-(R)-Nε-tert-butoxycarbonyl-lysyl]-(1R,2R)-1-(carboxyethylmethyl)-2-hydroxy-3-(2,6-dimethylbenzoyl)propylamine (9, 56 mg, 72%) as a colorless oil. R_f (10% MeOH/CH₂Cl₂) 0.65; MS (FAB, thioglycerol) m/z 857.4547 (MH⁺, C₄₄H₆₅N₄O₁₃ requires 857.4548); ¹³C-NMR (CDCl₃): 13.9 (OCH₂CH₃), 19.2 (2 x ArCH₃), 22.0 (Kγ-CH₂), 27.3 (^tBu), 27.7 (^tBu), 28.7 (2 x CH₂, Eβ and Kβ), 30.5 (Eγ-CH₂), 31.1 (Kδ-CH₂), 34.1 (CH₂CO₂Et), 39.9 (Kε-CH₂), 51.6, 52.6, 54.3 (3 x CHN), 60.6 (OC H₂CH ₃), 66.3(CH₂OCOAr), 66.5 (OC(CH₃)₃), 75.8 (CHOH), 80.6 (CH₂Ph), 127.0–135.6 (ArC), 155.9, 156.0 (2 x CONH), 168.4, 171.3, 171.4, and 172.4 (4 x CO₂). Dess-Martin periodinane (0.92 ml, 138 mg, 0.325 mmol) in CH₂Cl₂ was added drop-wise to a stirred solution of 9 (56 mg, 65 μmole) in dry CH₂Cl₂ under Ar and the reaction was allowed to stir overnight. The reaction mixture was diluted with CH₂Cl₂ (10 ml), washed with H₂O, dried, and concentrated. Column chromatography eluting with a 0–10% acetone/CH₂Cl₂ gradient gave N-[N-((R)-Nα-Benzyloxycarbonyl-Oγ-tert-butyl-glutamyl)-(R)-Nε-tert-butoxycarbonyl-lysyl]-(1R,2R)-1-(carboxyethylmethyl)-3-(2,6-dimethylbenzoyl)-2-oxopropylamine (10, 44 mg, 80%) as a colorless glass. R_f (10% MeOH/CH₂Cl₂) 0.65; MS (FAB, thioglycerol) m/z 855.4392 (MH⁺, C₄₄H₆₇N₄O₁₃ requires 855.4393); ¹³C-NMR (CDCl₃): 13.9 (OCH₂CH₃), 19.2 (2 x ArCH₃), 22.0 (Kγ-CH₂), 27.3 (^tBu), 27.7 (^tBu), 28.7 (2 x CH₂, Eβ and Kβ), 30.5 (Eγ-CH₂), 31.1 (Kδ-CH₂), 34.1 (CH₂CO₂Et), 39.9 (Kε-CH₂), 51.6, 52.6, 54.3 (3 x CHN), 60.6 (OCH₂CH₃), 66.3 (OC(CH₃)₃), 66.5 (OC(CH₃)₃), 75.8 (COCH₂O), 80.6 (CH₂Ph), 127.0–135.6 (ArC), 155.9, 156.0 (2 x CONH), 168.4, 171.3, 171.4, 172.4 (4 x CO₂) and 199.9 (CO). TFA/water (95:5, 300 μl) was added drop-wise to a stirred solution of 10 (5.0 mg, 5.8 μmole) in CH₂Cl₂ (2 ml) at 0°C and the mixture allowed to come to room temperature. After 2 h the reaction mixture was filtered and concentrated in vacuo, affording essentially homogeneous 11 (3.1 mg, 80%) as a colorless residue. R_f (10% MeOH/CH₂Cl₂) 0.05; HPLC-MS R_t (MH⁺) 24.14 min (694); MS (FAB, NOBA) m/z 694.3189 (MHNa⁺, C₃₃H₄₃N₄O₁₁Na requires 694.2826); ¹³C-NMR (CDCl₃): 17.3 (2 x ArCH₃), 21.2 (Kγ-CH₂), 25.4 (2 x CH₂, Eβ & Kβ) 26.9 (Eγ-CH₂), 31.1 (Kδ-CH₂), 32.9 (CH₂CO ₂), 37.9 (Kε-CH₂), 51.6, 52.6, 54.3 (3 x CHN), 78.9 (COCH₂O), 80.6 (CH₂Ph), 126.1–135.4 (ArC), 156.1, 160.2 (2 x CONH), 168.5, 169.6, 172.0, 174.2 (4 x CO₂) and 199.5 (CO)

N-Biotinylated Z-EKD-aomk derivatives (12a–d)

These compounds were prepared on a 4–5 mmolar scale in 70–80% yields by treatment of 11 with the N-hydroxysuccinamides of biotin, iminobiotin, LC biotin and LCLC biotin. Typically, to a stirred solution of 11 (3.1 mg, 4.6 μmole) in dry DMF (500 μl), biotin N-hydroxysuccinamide (1.0 mg, 5.1 μmole) was added. After 3 h the reaction mixture was diluted with CH₂Cl₂ (20 ml) and washed with H₂O to remove unreacted reagent and by-products. Concentration in vacuo afforded 12a (2.9 mg, 75%) as a colorless glass that was further purified by HPLC on a RP-8 silica column using MeOH:H₂O:TFA (66:33:0.1) as eluant.

12a

¹³C-NMR (CD₃CN-H₂O): 19.8 (2 x ArCH₃), 23.6 (Kγ-CH₂), 26.3 (biotin γ-CH₂), 27.6 (Kβ-CH₂), 28.9 (biotin β-CH₂), 29.2 (biotin δ-CH₂), 30.2 (Eβ-CH₂), 31.0 (Kδ-CH₂), 31.3 (CH₂CO₂), 35.4 (Eγ-CH₂), 36.4 (Kε-CH₂), 39.7 (biotin α-CH₂), 40.8 (biotin C-5), 53.9, 54.5, 55.5 (3 x CHN), 56.2 (biotin C-2), 61.1 (biotin C-3), 62.3 (COCH₂O), 62.9 (biotin C-4), 67.5 (COCH₂O), 68.0 (CH₂Ph), 128.6, 129.1, 129.5, 130.9, 133.5, 136.3, 137.5 (12 x ArC), 158.0 (CO), 165.5 (biotin ureido-C), 170.3, 172.2, 172.3, 174.0, 176.0, 176.8 (6 x CO) and 202.1 (ketone-CO).

12b

¹³C-NMR (CD₃CN-H₂O): 20.3 (2 x ArCH₃), 23.7 (Kγ-CH₂), 26.7 (biotin γ-CH₂), 28.4 (Kβ-CH₂), 29.1 (biotin β-CH₂), 29.2 (biotin δ-CH₂), 30.7 (Eβ-CH₂), 31.9 (Kδ-CH₂), 32.6 (CH₂CO₂), 35.9 (Eγ-CH₂), 36.9 (Kε-CH₂), 40.0 (biotin α-CH₂), 40.7 (biotin C-5), 54.2, 54.4, 56.1 (3 x CHN), 56.6 (biotin C-2), 62.7 (biotin C-3), 64.0 (COCH₂O), 64.8 (biotin C-4), 68.1 (COCH₂O), 68.5 (CH₂Ph), 129.1, 129.5, 129.9, 131.3, 134.2, 136.8, 138.0 (12 x ArC), 156.1, 158.0, 158.6, 170.3, 172.2, 172.3, 174.0, 176.1 (CO & C=N) and 202.1 (ketone CO).

12c

¹³C-NMR (CD₃CN-H₂O): 19.3 (2 x ArCH₃), 23.0 (Kγ-CH₂), 25.6, 25.8 (2 x linker CH₂), 26.3 (biotin γ-CH₂), 27.6 (Kβ-CH₂), 28.3 (biotin β-CH₂), 28.6 (biotin δCH₂), 28.8 (linker CH₂), 29.7 (Eβ-CH₂), 30.8 (Kδ CH₂), 31.6 (CH₂CO₂), 34.9 (Eγ-CH₂), 36.0 (linker CH₂CO), 36.1 (kε-CH₂), 39.3 (linker CH₂N), 39.9 (biotin α-CH₂), 40.3 (biotin C-5), 53.3, 53.9, 55.3 (3 x CHN), 55.8 (biotin C-2), 60.5 (biotin C-3), 61.8 (COCH₂O), 62.2 (biotin C-4), 67.1 (COCH₂O), 67.5 (CH₂PH), 128.1, 128.5, 130.0, 133.0, 135.7, 135.7, 137.0, 139.0 (12 x ArC), 157.5 (CO), 164.8 (biotin ureido-C), 169.7, 171.5, 171.6, 173.6, 173.7, 175.3, 175.4 (7 x CO) and 201.8 (ketone-CO).

12d

¹³C-NMR (CD₃CN-H₂O): 19.3 (2 x ArCH₃), 23.0 (Kγ-CH₂), 25.6, 25.8 (4 x linker CH₂), 26.3 (biotin γ-CH₂), 27.5 (Kβ-CH₂), 28.3 (biotin β-CH₂), 28.7 (biotin δ-CH₂), 29.8 (2 x linker CH₂), 29.6 (Eβ-CH₂), 31.0 (Kδ-CH₂), 31.4 (CH₂CO₂), 34.9 (Eγ-CH₂), 36.0 (2 x linker CH₂CO), 36.2 (Kε-CH₂), 39.1, 39.3 (2 x linker CH2N), 39.9 (biotin α-CH₂), 40.3 (biotin C-5), 53.3, 53.9, 55.3 (3 x CHN), 55.8 (biotin C-2), 60.5 (biotin C-3), 61.8 (COCH₂O), 62.2 (biotin C-4), 67.1 (COCH₂O), 67.5 (CH₂Ph), 128.1, 128.5, 130.0, 130.3, 133.0, 135.8, 137.0, 139.0 (12 x ArC), 157.6 (CO), 164.9 (biotin ureido-C), 169.7, 171.6, 171.7, 173.6, 173.7, 175.4, 175.5 (7 x CO) and 201.8 (ketone-CO).

HPLC-MS (Rt, MH⁺): 12a (18.50 min, 897), 12b (18.34 min, 896), 12c (18.28 min, 1010), 12d (18.11 min, 1123). Because exact mass determinations could not easily be obtained for these compounds directly, they were derivatized as their ethyl esters by treatment with 3% HCl in EtOH for 1 h at 10°C. MS (FAB, thioglycerol) of the monoethyl esters of 12a–d gave exact masses within a deviation of 0.8 ppm of the calculated MH⁺ values (See Supporting Information).

N-(2,4-Dinitrophenyl) Z-EKD-aomk (13)

To a stirred solution of 11 (3.1 mg, 4.6 μmole) in dry DMF (500 μl), N-methylmorpholine (50 μg, 5 μmole) and 2,4-dinitrofluorobenzene (950 μg, 5.1 μmole) were added. After 12 h the reaction mixture was diluted with CH₂Cl₂ (20 ml), washed with H₂O, dried, and concentrated to afford 13 (2.9 mg, 75%) as a colorless glass. R_f (10% MeOH/CH₂Cl₂) 0.55; HPLC-MS R_t (MH⁺) 20.41 min (837); MS (FAB, NOBA) m/z MH⁺ 837.3028 (MH⁺, C₃₉H₄₅N₆O₁₅ requires 837.2943); ¹³C-NMR (CDCl₃): 19.1 (2 x ArCH₃), 21.5 (Kγ-CH₂), 22.4 (Kβ-CH₂), 27.4 (Eβ-CH₂), 29.0 (Kδ-CH₂), 30.8 (Eγ-CH₂), 35.8 (CH₂CO₂), 42.5 (Kε-CH₂), 50.5, 51.6, 52.7 (3 x CHN), 78.9 (COCH₂O), 79.2 (CH₂Ph), 113.4 (CNO ₂), 123.7 (CNO₂), 126.1–135.4 (15 x ArC), 147.7 (ArC) 156.1, 160.2 (2 x CONH), 170.6, 171.1, 171.3, 173.6 (4 x CO₂) and 199.5 (ketone-CO).

Spectrophotometric assays of caspase inhibition

Recombinant human caspases 1–10 were purchased from Biomol (Plymouth Meeting, PA). Peptidyl caspase substrates conjugated to p-nitroaniline (pNA), including Ac-YVAD-pNA, Ac-LEHD-pNA, Ac-DEVD-pNA, Ac-VEID-pNA and Ac-IETD-pNA, from Biomol were confirmed to be >95% pure by HPLC, as were the peptidyl inhibitors prepared in this study. Assays were conducted in 96-well microtiter plates and contained: 150 μl of assay buffer [10 mM Tris-HCl, 1 mM DTT, 10% (v/v) glycerol, 0.1% (w/v) CHAPS, pH 7.5], 10 μl of enzyme in assay buffer (final concentration 1–100 nM), 20 μl of substrate solution (dissolved in DMSO and diluted in assay buffer to 3.9-1000 μM) and 20 μl of inhibitor solution (dissolved in DMSO and diluted in assay buffer to 1-1000 μM). Assays were conducted in triplicate with the assay buffer, substrate and inhibitor preincubated in the microtiter plate at 37°C. Reactions were started by the addition of enzyme warmed to 37°C. Production of pNA was measured by following the absorbance at 405 nm minus that at 650 nm using an EL_x808 Ultra microplate reader from Bio-Tek Instruments Inc (Winooski, Vermont).

Progress curves of product formation versus time were fitted using the program KinetiCalc for Windows (Bio-Tek Instruments, Winooski, Vermont) to measure the initial velocity for reactions over the range of 3.9 – 1000 μM substrate and 0.1 – 100 μM inhibitor. Michaelis-Menton curves of the initial velocity data were fitted by non-linear least squares regression analysis using SigmaPlot version 9.01 with the Enzyme Kinetics Module (Systat Software, Inc., Richmond, CA, USA). The substituted Z-EKD-amok caspase inhibitors were modelled as tight binding inhibitors.⁴²

Tissue culture, cytosolic extract preparation and affinity labelling

The human Jurkat T cell lymphoma cell line was cultured in RPMI 1640 supplemented with 10% fetal bovine serum. The chicken lymphoma B-cell line DT40 expressing chicken procaspase 6 fused to EGFP³⁴ was cultured as previously described.⁴³ To induce apoptosis, cells were treated with 10 μM etoposide for 5. Cytosolic extracts were prepared from untreated and apoptotic cells according to the procedure described by Ruchaud et al.³⁴

Aliquots of apoptotic cytosol were incubated for 30 min at 37°C with 2 μM inhibitor. At the completion of the incubation, samples were diluted with 0.5 volume of 3X SDS sample buffer, heated to 95°C for 5 min, subjected to SDS-PAGE on 16% (w/v) acrylamide gels, transferred to nitrocellulose or polyvinylidene fluoride membranes (Amersham Biosciences, Uppsala, Sweden), probed with HRP-coupled streptavidin (Sigma, St. Louis, MO). Alternatively, blots were probed with antibodies as described below.

Immunoblotting

Samples containing 30 μg of total cellular protein were subjected to electrophoresis for 75 min at 200 V on SDS polyacrylamide minigels containing 16% (w/v) acrylamide and electrophoretically transferred to nitrocellulose or polyvinylidene fluoride membranes (Amersham Biosciences) at ambient temperature for 90 min at 40 mA per minigel. Immunoblotting followed by enhanced chemiluminescent detection was performed as described³⁴ using mouse monoclonal anti-DNP from Molecular Probes (Leiden, NL), rabbit anti-EGFP antiserum from Ilan Davis (Univ. of Edinburgh) or rabbit polyclonal anti-chicken caspase-6 (R549) raised against the large subunit of chicken caspase-6³⁴ as primary antibodies and HRP-conjugated anti-mouse IgG or anti-rabbit IgG (Amersham Biosciences) as secondary antibody.

Affinity purification of derivatized caspase-6

Avidin-Sepharose, monomeric avidin-Sepharose beads, and Slide-a-lyze dialysis slides were purchased from Pierce (Rockford, IL). Endogenous biotinylated proteins were removed from apoptotic cytosolic extracts of DT40 cells expressing EGFP:caspase-6³⁴ by incubation with avidin-Sepharose beads for 30 min at 4 ºC. Unbound polypeptides were eluted with calcium- and magnesium-free Dulbecco’s phosphate buffered saline (PBS) and labelled with Z-EK(LC-biotin)D-aomk (12c, 2 μM) by incubation for 30 min at 37°C. Excess inhibitor was removed by dialysis against PBS overnight at 4 ºC. After the labelled fraction was incubated with monomeric avidin-Sepharose beads for 30 min at ambient temperature, the column was washed (three washes of 3X bed volume) with PBS. The non-specifically bound polypeptides were removed with increasingly stringent washes (4 times 3X bed volume of 0.5 M NaCl, 4 times 3X bed volume of 1.0 M NaCl). Bound polypeptides were then eluted with 2 mM biotin in PBS (4 times 3X bed volume). Eluted fractions were concentrated using 5K spin columns (Millipore, Bedford, MA).

Each fraction was loaded onto two SDS-polyacrylamide gels containing 16% (w/v) acrylamide. One gel was stained with Colloidal Coomassie Blue (Genomic Solutions, Huntington, UK) and the other was transferred to a nitrocellulose, probed with HRP-coupled streptavidin, stripped (62.5 mM Tris pH 6.8, 2% SDS, 100 mM β-ME),⁴⁴ reprobed with anti-chicken caspase-6, stripped again and reprobed with anti-EGFP antibody.

MALDI-ToF mass spectrometry

Polypeptide identification was accomplished by peptide mass fingerprinting and sequence database searching.⁴⁵ After polypeptide bands were excised from Coomassie blue-stained gels, in-gel digestion with trypsin was performed as previously described,⁴⁶ followed by sample preparation using miniaturized sample concentration/desalting techniques.⁴⁷ For the mass spectrometric analysis, a PerSeptive Biosystems Voyager DE™STR MALDI-ToF mass spectrometer (Applied Biosystems, Foster City, CA) was used. Peptide ion signals were assigned with a mass error less than 50 ppm. Lists of tryptic peptide masses were used to search protein sequence databases using the MS-Fit tool (http://prospector.ucsf.edu/ucsfhtml4.0u/msfit.htm). For positive identification a molecular weight search (MOWSE) score⁴⁸ of 10⁴ and at least 20% coverage of the protein by the peptide fragments was required.

Flow cytometry

Jurkat cells were treated with 10 μM etoposide (Calbiochem, San Diego, CA) plus 10 μM Z-VAD(OMe)-fmk (Calbiochem) or ZEK(DNP)D-aomk (13) at 1 or 10 μM for 5 h, harvested, washed in PBS, fixed in 4% (w/v) paraformaldehyde in PBS, permeabilized in 0.1 % (w/v) Triton X-100 containing 0.1% (w/v) sodium citrate in PBS for 2 min on ice and labeled using the TUNEL (terminal deoxyribonucleotidyl transferase nick end labeling) reaction according to the manufacturer’s recommendations (In situ Cell Death Detection Kit, TMR red, Roche, Indianapolis, IN). The cells were then analyzed on a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) using Becton Dickinson CellQuest software.

Supplementary Material

si20060809_080. Supporting Information Available.

Further documentation of the characterization of target compounds 12a–d and 13 is available as supporting infromation. This material is available free of charge via the Internet at http://pubs.acs.org.