|Home | About | Journals | Submit | Contact Us | Français|
The immunoproteasome, having been linked to neurodegenerative diseases and hematological cancers, has been shown to play an important role in MHC class I antigen presentation. However, its other pathophysiological functions are still not very well understood. This can be attributed mainly to a lack of appropriate molecular probes that can selectively modulate the immunoproteasome catalytic subunits. Herein, we report the development of molecular probes that selectively inhibit the major catalytic subunit, LMP2, of the immunoproteasome. We show that these compounds irreversibly modify the LMP2 subunit with high specificity. Importantly, LMP2-rich cancer cells compared to LMP2-deficient cancer cells are more sensitive to growth inhibition by the LMP2-specific inhibitor, implicating an important role of LMP2 in regulating cell growth of malignant tumors that highly express LMP2.
In the era of proteomics, temporal and spatial control of protein functions, which are often difficult with conventional genetic manipulations, are critical to the understanding of the dynamics of cellular processes. While traditional genetic approaches have provided useful insight into the functions of proteins, they are limited by the possibility that some phenotypes may be due to compensatory responses that occur during development. In addition, the inhibition of the target gene function is often irreparable, and thus the desired protein deficiency cannot be readily regulated, making it difficult to dissect the precise roles of gene products. One way to complement classical genetic approaches is to use small molecules that selectively modulate protein functions. This small-molecule approach has increasingly contributed to further our understanding of biological processes.
The proteasome has emerged as a major player in many important signaling processes, such as cell cycle progression , inflammatory responses , and development . Typically, more than 80%of cellular proteins are degraded by the ubiquitin-proteasome system. The ubiquitin-proteasome pathway is a highly regulated process in which proteins are first targeted for degradation by conjugation to ubiquitin, a 76 amino acid polypeptide. Ubiquitinated proteins are, in turn, recognized by the 19S regulatory domain of the constitutive 26S proteasome. Through a series of ATP hydrolysis-dependent processes, deubiquitinated proteins are threaded into the core proteolytic complex, the 20S proteasome, where they are degraded into small peptides. The 20S core has a four-ring stacked structure with seven different subunits in each ring. The two inner rings each contain three catalytically active β subunits. The noncatalytic outer α rings form a gated channel for unfolded protein entry and a base for the 19S regulatory complexes, which provide the specificity of the polypeptide recognition.
The 20S catalytic core proteasome has been shown to exhibit three major activities: a chymotrypsin-like (CT-L) activity that cleaves after large hydrophobic residues, a trypsin-like (T-L) activity that hydrolyzes after basic amino acids, and a caspase-like (C-L) activity that cleaves after acidic amino acids. Two other less-characterized catalytic activities have also been ascribed to the proteasome: BrAAP, which cleaves after branched-chain amino acids, and SNAAP, which cleaves after small, neutral amino acids. Although most efforts are directed to develop proteasome inhibitors against CT-L activity, a few studies have also been successful in designing compounds that inhibit other proteasomal activities, such as C-L  and T-L activity-specific inhibitors [5–8]. While the CT-L activity of the proteasome has been suggested to be largely responsible for the proteolytic function of the proteasome in vivo andin vitro [9, 10], the contribution of the other major activities remains to be determined. In recent years, researchers have been investigating the functions of the different proteolytic activities in cancer cells by using a variety of proteasome inhibitors [10, 11]. Regarding clinical applications of proteasome inhibitors, bortezomib (VELCADE), a broad-spectrum proteasome inhibitor targeting both the constitutive proteasome and immunoproteasomes, was recently approved by the FDA for the treatment of multiple myeloma (MM) . However, its clinical use is severely limited due to drug-related toxicities .
In higher vertebrates, exposure of cells to stimuli, such as interferon (IFN)-γ or tumor necrosis factor (TNF)-α, induces the synthesis of certain catalytic subunits (LMP7, LMP2, and MECL-1), which replace the constitutive β sub-units X, Y, and Z, respectively, and form an alternative pro-teasome form known as the immunoproteasome . The immunoproteasome, as compared to the constitutive (or regular) proteasome, has an enhanced capacity to generate peptides bearing hydrophobic and basic amino acids at their C termini and a reduced capacity to produce pep-tides bearing acidic residues at their C termini . Consequently, the spectrum of the resultant peptides is shifted toward peptides that associate with MHC class I molecules with increased affinity . While the immunoproteasome is suggested to play a major role in MHC class I antigen presentation, it is believed not to be solely responsible for antigen presentation, as the constitutive pro-teasome also generates immunogenic epitopes .
In recent years, questions regarding the role of the immunoproteasome in cells from nonimmune systems have arisen due to the findings in which expression levels of individual immunoproteasome subunits are correlated with pathological processes, such as hematological cancers and neurodegenerative diseases [18–20]. For instance, a high level of immunoproteasome catalytic sub-units has been detected in neurodegenerative human brains [21, 22], which is known to be an immunologically privileged organ . Specifically, the LMP2 catalytic sub-unit is more highly expressed in the brains of Alzheimer's disease (AD) patients than in the brains of nondemented elderly, whereas its expression in young brains is negligible or absent . More recently, it has been shown that LMP2 is required for estrogen receptor-mediated gene transcription and for estrogen-stimulated cell cycle progression . Further, some studies indicated that the immunoproteasome and its catalytic subunits may be involved in Huntington's disease (HD) neurodegeneration . MM is also known to express a high level of immunoproteasome subunits due to the bone marrow microenvironment where it replicates [26–29]. Based on these observations, an intriguing hypothesis has been proposed that the specific inhibition of the immunoproteasome sub-units may induce selective apoptosis of MM cells while sparing other cells lacking or minimally expressing immunoproteasome subunits , making them an attractive investigative target for clinical applications.
Despite the potential role of the immunoproteasome catalytic subunits in pathogenesis, their functions are still not fully understood. In addition, the catalytic activities of individual immunoproteasome subunits are not clearly characterized. The major problem that limits further understanding of immunoproteasome biology is the lack of appropriate molecular probes that selectively target the immunoproteasome catalytic subunits. Unfortunately, the proteasome inhibitors developed to date either preferentially target the constitutive proteasome subunits or fail to exhibit appropriate specificity toward the immunoproteasome subunits.
With this in mind, our ongoing efforts are aimed at the design and synthesis of small-molecule probes that selectively target the immunoproteasome catalytic subunits. While the sequence comparison of catalytic subunits from the constitutive proteasome and immunoproteasomes exhibits a high level of homology, structural information about the active sites of the immunoproteasome subunits remains to be elucidated, complicating our efforts toward the design of immunoproteasome sub-unit-specific probes via a rational target-based design strategy.
Two natural product proteasome inhibitors, epoxomicin and eponemycin (Figure 1), are members of the α′,β′-epoxyketone linear peptide family [31, 32]. It has been previously shown that, despite structural similarities, epoxomicin (1) and dihydroeponemycin (2), an active derivative of eponemycin, differ considerably in their proteasome subunit binding specificity [31, 32]. For example, dihydroeponemycin labels the catalytic threonine residues of the immunoproteasome subunits LMP2 and LMP7 and the constitutive proteasome subunit X. On the other hand, epoxomicin covalently modifies the N-terminal catalytic threonine residues of the constitutive proteasome (X and Z) and immunoproteasome (LMP7 and MECL1) subunits. We have previously shown that a relatively higher specificity of dihydroeponemycin toward the immunoproteasome subunits (LMP2 and LMP7), as compared to epoxomicin, is due to a linear hydrocarbon residue at the N terminus (i.e., isooctanoic group) . In addition, it has been suggested that amino acid residues at the P1′-P2′ sites (see Figure 1) of immunoproteasome substrates may also play an important role in immunoproteasome specificity . We have also previously reported that serine at the P2 site of dihydroeponemycin can be replaced with alanine while maintaining the subunit-binding pattern of dihydroeponemycin .
Based on these observations, we would like to develop a new molecular probe that will selectively inactivate the catalytic threonine residue of immunoproteasome subunit LMP2 by derivatizing the P1′-OH group of dihydroeponemycin (Figure 1). We envision that LMP2 inhibitors that covalently modify the catalytic threonine residue of LMP2 will provide a valuable chemical genetic tool in the functional exploration of individual immunoproteasome subunits. Herein, we report the syntheses, through the use of easily available protecting groups, of a variety of P1′-derivatized dihydroeponemycin analogs. We also show that certain P1′ derivatives of dihydroeponemycin irreversibly inactivate the LMP2 subunit with remarkable specificity in living cells. Finally, we show that LMP2-rich cancer cells are more sensitive to growth-inhibitory activity of the LMP2 inhibitor compared to LMP2-deficient cancer cells.
Biotin-tagged epoxomicin and dihydroeponemycin  were used as assay probes with which to perform a screening assay for LMP2-specific compounds. We first corroborated the screening assay by western blot analysis with epoxomicin (1) and dihydroeponemycin (2), proteasome subunit-binding patterns of which have been previously well defined [31, 35]. The EL4 cell system was chosen because these cells express high levels of catalytic subunits of both the constitutive proteasome and immunoproteasomes. Specifically, various concentrations of these compounds were preincubated in EL4 cells at 37°C for 30 min. Assay probes, biotin-tagged dihydroeponemycin or epoxomicin, were then added. Cells were incubated for an additional hour at 37°C before cell lysis. Whole-cell lysates were then analyzed by using 12% SDS-PAGE and were transferred to PVDF membranes. Proteins that were newly biotinylated by assay probes were visualized by using streptavidin-horseradish peroxidase (HRP) and the enhanced chemiluminescence (ECL) detection system.
As shown in Figure 2, most of the proteasome subunit bands, which were covalently modified and visualized by using assay probes, were easily competed away with excess dihydroeponemycin, its P2 analog, or epoxomicin. This result confirms that neither epoxomicin nor dihydroeponemycin have specificity toward subunits of either the constitutive proteasome (X) or the immunoproteasome (LMP2 and LMP7). Dihydroeponemycin [Ser → Ala], in which the P2 serine of the dihydroeponemycin analogisreplaced with alanine, displayed a similar subunit-binding pattern to that of dihydroeponemycin. Usingasimilar competition assay, we screened P1′ derivatives of dihydroeponemycin for LMP2-specific compounds. Our expectation was that preincubation of a LMP2-specific inhibitor in EL4 cells will result in the modification of the threonine catalytic residue of the LMP2 subunit, preventing further modification of the occupied LMP2 subunit by assay probes. Therefore, the preoccupied LMP2 will no longer be visualizedon western blot. However, proteasome catalytic subunits that are not targeted by the LMP2 inhibitor will be covalently labeled by the assay probes and visualized by western blotting by using the HRP-ECL system.
Since a linear hydrocarbon group at the P3 position is shown to provide high specificity toward the LMP2 sub-unit , the heptanoic group was positioned at the P3 site of dihydroeponemycin (Figure 1). We then focused on the derivatization at the P1′-OH group (Figure 3). First, we added an easily available methoxymethyl ether (MOM) group, preparing compounds 11 and 14. This replacement caused a dramatic loss in the potency and specificity compared to dihydroeponemycin (Figure 4A). Similarly, compounds with a bulky tert-butyldiphenylsilyl (TBDPS) group (10) or tetrahydropyranyl (THP) group (13) also resulted in loss of subunit-binding activity against the immunoproteasome (Figure 4A).
Strikingly, when the MOM group was replaced with a methoxyethoxymethyl (MEM) ether group (12), high specificity toward LMP2 was observed (Figure 4B). When a tert-butyldimethylsilyl (TBDMS) group was attached at the C-terminal hydroxyl group (15), an even higher specificity toward the LMP2 subunit was obtained, as shown in Figure 4B. Preincubation of EL4 cells with 1 μM compound 15 was sufficient to covalently modify all of the LMP2 subunit in EL4 cells, preventing further modification of the LMP2 subunit by assay probe. This resulted in selective attenuation of the LMP2 protein band on the western blot. Experiments with another assay probe (biotin-epoxomicin) (Figure 4C), which covalently labels protea-some subunits LMP7, X, MECL-1, and Z, exhibit no competition, further supporting the conclusion that both compounds 12 and 15 selectively target the LMP2 sub-unit, but not other proteasome subunits.
To demonstrate that compound 15 covalently and selectively modifies the LMP2 subunit, but not other subunits, we investigated the mobility shift of the LMP2-15 adduct by using EL4 cells (Figure 4D). After EL4 cells were incubated with compound 15 or assay probes (bioti-nylated epoxomicin [Biotin-EPX] and Biotin-EPN) for 1.5 hr at 37°C, cells were lysed. Whole-cell proteins were then analyzed by western blot with anti-LMP2, anti-LMP7, anti-X, and anti-Y antibodies. In this experiment, biotiny-lated epoxomicin (Biotin-EPX) and biotinylated dihydroeponemycin (Biotin-EPN) were used as mobility shift controls since they have been shown to cause mobility shift via covalent modification of LMP2 [31, 32, 36].
Due to increased molecular weights (by 828.08 for biotin-epoxomicin and 1078.45 for biotin-dihydroeponemycin), assay probe-LMP2 adducts displayed a slower mobility shift when compared to free LMP2 (lanes 1–3, Figure 4D). While a mobility shift for the LMP2-15 adduct on SDS-PAGE was clearly shown in comparison to free LMP2 (lanes 4–6, Figure 4D), it was observed to be slower than that of assay probe-LMP2 adducts (lanes 2–3 versus lanes 4–6). This can be explained with the lower molecular weight of compound 15 (484.76) as compared to the assay probes. While 1 μM compound 15, but not the assay probes, was sufficient to modify most of the LMP2 sub-units in cells (lanes 2–3 versus 4–6), no mobility shift was observed in other proteasome subunits, indicating that compound 15 selectively inactivates the LMP2 subunit with high efficiency.
The elucidation of protein function, especially with regard to multiprotein complexes, is clearly one of the important goals of the small-molecule approach to cell biology. With respect to the proteasomal complex, researchers have long been interested in assigning specific catalytic activities to each β subunit of the proteasome. Based on data obtained from X-ray analysis and direct inhibition of proteasome activities with a variety of proteasome inhibitors, all three β subunits of the constitutive proteasome have been assigned to three different activities [4, 9, 37, 38]. Whereas LMP7 and MECL1 of the immunoproteasome are shown to possess CT-L and T-L activities, respectively, a clear assignment for the LMP2 subunit of the immunoproteasome has not yet been made. LMP2, which replaces the constitutive proteasome catalytic subunit Y that is responsible for C-L activity, has been suggested to cleave substrates after hydrophobic amino acid residues to generate peptides favored for MHC class I presentation [39, 40], but experimental data to support this assumption are lacking. Therefore, with our LMP2-specific inhibitor in hand, we wanted to determine whether LMP2 is responsible for CT-L activity. To test this, we used the natural product lactacystin, which primarily binds the LMP7 (and MECL1) subunit  and inactivates its catalytic activity [42, 43], and then determined whether free LMP2 has CT-L activity. It should be noted that MECL1 is reported to be responsible for T-L activity [44, 45]. As shown in Figure 5A, when the immunoproteasome was preincubated with lactacystin at the concentration at which 95% of the CT-L activity of the constitutive proteasome is inhibited, ~20% of the CT-L activity of the immunoproteasome was blocked compared to control. Similarly, when the immunoproteasome was preincubated with compound 15 at the concentration at which only LMP2 is inactivated (see Figures 4D and and6D),6D), ~20% of the CT-L activity of the immunoproteasome was also inhibited (Figure 5A). In the presence of both lactacystin and compound 15, ~45% of the CT-L activity was blocked, whereas the CT-L activity of the immunoproteasome was completely blocked when the concentration of either lactacystin or compound 15 was increased (data not shown). This finding indicates that these compounds have an additive inhibitory action on the CT-L enzymatic function of the immunoproteasome. Collectively, these results suggest that LMP2 is, at least in part, responsible for the CT-L activity of the immunoproteasome.
Despite the fact that compound 15 selectively modifies the catalytic LMP2 subunit and inhibits the CT-L activity of the immunoproteasome, it is still unclear whether compound 15 also inhibits essential CT-L activity of the constitutive proteasome in living cells. Thus, we next tested whether compound 15 can disrupt cellular events that are regulated by the constitutive proteasome. Angiogenic growth of blood vessels requires normal activities of the constitutive proteasome and is highly sensitive to proteasome inhibitors  such as epoxomicin and dihydroepo-nemycin. Conversely, endothelial cells do not normally express the immunoproteasome (data not shown), and, thus, we expect that angiogenic sprouting in an in vitro model will be highly sensitive to regular proteasome inhibitors, but not to LMP2-specific inhibitors.
The three-dimensional endothelial cell sprouting assay (3D-ECSA) is an in vitro experimental system that closely mimics the in vivo angiogenesis processes, featuring the differentiation of endothelial cells into sprouting structures within a 3D matrix of fibrin or collagen I . We employed the 3D-ECSA assay to investigate whether compound 15 and proteasome inhibitors differ in their inhibitory activity on growth factor-induced sprouting angiogenesis. As anticipated, epoxomicin  and dihydroeponemycin potently inhibited sprouting morphogenesis due to inhibition of the essential constitutive proteasome activity (Figure 5B). On the other hand, compound 15 did not inhibit endothelial sprouting at doses much higher than those of dihydroeponemycin or epoxomicin, even though compound 15 is more efficient at inactivating the LMP2 subunit (see Figure 4D). Quantitative measurements of endothelial sprouting as described earlier [46, 48] revealed that compound 15 with concentration as high as 10 μM only marginally disrupts sprouting (~80-fold higher concentration than dihydroeponemycin) (data not shown). Similarly, compound 12 did not inhibit endothelial sprouting. These angiogenesis assay results indicate that the LMP2 inhibitors do not perturb the functions of the constitutive proteasome in living cells.
Given that one of our goals is to explore the pathophysiological functions of LMP2 in relevant disease models, we questioned whether LMP2 inhibitors developed here have any impact on normal biological processes of cancer cells that predominantly express LMP2. First, we investigated which prostate cancer cells constitutively express the LMP2 subunit. Surprisingly, among the prostate cancer cell lines examined, only androgen-independent PC3 prostate cancer cells, but not androgen-dependent LNCaP or LN3 cells, constitutively expressed the immunoproteasome catalytic subunit LMP2 (Figure 6A). Remarkably, the specificity of compounds 15 and 12 toward LMP2 is even more evident in PC3 cells compared to EL4 cells. As shown in Figure 6B, the LMP2 protein band in PC3 cells was selectively attenuated by the LMP2 inactivators 15 and 12, indicating their high specificity toward the LMP2 subunit. Unlike epoxomicin or dihydroeponemycin, compounds 12 and 15 did not compete with the biotinylated probes for binding of LMP7, X, Z, and MECL1 (Figure 6C). Mobility shift assays with western blot, which is analyzed with anti-LMP2, -LMP7, -X, or -Y antibodies, clearly showed that compound 15 covalently inactivates LMP2, but not the other catalytic subunits of proteasomes (Figure 6D). Similar to the results obtained from the mobility shift experiments with EL4 cells (Figure 4D), compound 15 was a more potent LMP2 inhibitor than the biotinylated probes. Probe-LMP2 conjugates at 1 μM concentration were not even observedinthe western blot experiments (lanes 2 and 3 in Figure 6D). LMP7, X, and Y subunits were not covalently modified by compound 15 even at 10 μM concentration. These results clearly indicate that compound 15 selectively modifies LMP2 constitutively expressed in PC3 prostate cancer cells.
Next, we wished to determine whether the LMP2-specific inhibitor has effects on the proliferation of PC3 cancer cells that highly express LMP2. Given that LMP2 is a major catalytic subunit of the immunoproteasome and the fact that proteasomes play an important role in cell growth, we hypothesized that PC3 cells may be more sensitive to the LMP2 inhibitor 15 compared to LMP2-deficient LN3 prostate cancer cells. In contrast, we expect that the broadly acting proteasome inhibitors epoxomicin and dihydroeponemycin will not display differential activity toward the prostate cancer cells regardless of their expression level of LMP2.We tested this hypothesis by measuring IC50 values for compound 15 in both PC3 (LMP2-positive) and LN3 (LMP2-deficient) cells. Remarkably, PC3 cells are about 7-fold more sensitive to the LMP2 inhibitor 15 than LN3 prostate cancer cells (Table 1). In contrast, both PC3 and LN3 cells were similarly sensitive to the broad-spectrum proteasome inhibitors epoxomicin and dihydroeponemycin.
Although it is currently not clear why PC3 cancer cells, but not other prostate cancer cell lines, constitutively express LMP2, these results indicate that LMP2 may play an important role in proliferation of cancer cells that constitutively express LMP2. It is also not known whether LMP2 within the immunoproteasome or monomeric LMP2 is a pharmacological target of compound 15. Although it is presumed that newly synthesized LMP2 (pre-LMP2) is catalytically inactive until it has matured into catalytically active LMP2 by LMP7 and has been assembled into the immunoproteasome [49–51], the possibility of alternative maturation into a catalytically active LMP2 monomer that may be functionally important in cell proliferation cannot be ruled out. Regardless, the high sensitivity of LMP2-rich PC3 cancer cells to compound 15 suggests that LMP2 may be a target for therapeutic intervention in cancers that constitutively express this protein. Additionally, these findings have allowed us to conclude that compound 15 may be a molecular probe of LMP2 function that can be exploited to illuminate the biological roles of its substrates in immunoproteasome-dependent processes.
In conclusion, we have developed an epoxyketone-pharmacophore-based LMP2-specific probe with which the physiological roles of LMP2 in cells can be investigated. As the LMP2 inhibitor blocks proliferation of LMP2-rich cancer cells with high specificity, it can now be utilized to determine whether the LMP2 subunit is a valid target for therapeutic intervention in animal models of cancer. Further, compound 15 can also be used as a chemical knockout reagent of LMP2 to screen for immu-noproteasome substrates that are distinct from regular proteasome substrates, which will ultimately enhance our understanding of the role of LMP2 in pathogenic diseases.
Unless otherwise stated, all reactions were carried out under nitrogen with dry, freshly distilled solvents, oven-dried glassware, and magnetic stirring. All solvents were reagent grade. Tetrahydrofuran (THF) was distilled from sodium/benzophenone. Methylene chloride (CH2Cl2) was distilled from calcium hydride. Diethyl ether anhydrous was purchased from EMD Chemicals and was used without further purification. All reagents were purchased from Sigma-Aldrich and were used without further purification. All reactions were monitored by thin-layer chromatography (TLC) by using E. Merk 60F254 precoated silica gel plates. Flash column chromatography was performed by using E. Merk silica gel 60 (particle size 0.040–0.063 mm) and was performed with the indicated solvents. 1H was recorded in CDCl3 by using a Varian 300MHz spectrometer at ambient temperature with an internal deuterium lock unless stated otherwise. Chemical shifts are referenced to residual chloroform (δ = 7.27 ppm for 1H). High- and low-resolution mass spectra were carried out by the University of Kentucky Mass Spectrometry Facility.
Synthesis of 12 is described here as the representative synthetic procedure for dihydroeponemycin analogs.
Synthetic procedures were performed as previously reported .
Methoxyethoxymethyl chloride (0.24 ml, 2.1 mmol) and diisopropyle-thylamine (0.37 ml, 2.1 mmol) were added to a solution of 3 (114 mg, 0.42 mmol) in CH2Cl2 (5 ml) at 0°C. After stirring at room temperature for 3 hr, the resulting mixture was poured into ice water (20 ml) and extracted with CH2Cl2 (3 × 20 ml). The organic layers were combined, washed with brine (20 ml), dried with Na2SO4, filtered, and concentrated under reduced pressure. The product was then subjected to flash column chromatography (5:1 hexane:EtOAc), yielding 4 (101 mg, 67%) as a yellowish oil. 1H NMR: δ = 6.20 (d, 2J = 31.8 Hz, 2H, 1-H), 5.12 (d, 2J = 9.0 Hz, 1H, NH), 5.03 (m, 1H, 4-H), 4.75 (s, 2H, 2-OCH2O), 4.28 (s, 2H, 2-CH2), 3.69 (m, 2H, 2-OCH2CH2O), 3.55 (m, 2H, 2-OCH2CH2O), 3.38 (s, 3H, 2-OCH3), 1.74 (m, 1H, 6-H), 1.50 (m, 1H, 5-Ha), 1.41 (s, 9H, HBoc), 1.31 (m, 1H, 5-Hb), 0.99 (d, 3J = 6.6 Hz, 3H, CH3CHCH3), 0.90 (d, 3J = 6.6 Hz, 3H, CH3CHCH3) ppm.
Benzonitrile (0.29 ml, 2.8 mmol), H2O2 (0.40 ml, 50% solution in H2O, 7.0 mmol), and diisopropylethylamine (0.5 ml, 2.8 mmol) were added to a solution of 4 (100 mg, 0.28 mmol) in MeOH (5 ml) at 0°C. The reaction was stirred at 0°C for 3 hr. The resulting mixture was then concentrated under reduced pressured and was subjected to flash column chromatography (10:1 hexane:EtOAc) to yield 5 and 6 at a ratio of 3:1 (60 mg, 60%). 5: 1H NMR: δ = 4.82 (d, 2J = 8.4 Hz, 1H, NH), 4.71 (s, 2H, 2-OCH2O), 4.39 (d, 2J = 11.4 Hz, 1H, 2-CHa2), 4.32 (m, 1H, 4-H), 3.68 (m, 2H, 2-OCH2CH2O), 3.57 (m, 2H, 2-OCH2CH2O), 3.49 (d, 2J = 11.4 Hz, 1H, 2-CHb2), 3.40 (s, 3H, 2-OCH3), 3.27 (d, 2J = 4.8 Hz, 1H, 1-Ha), 3.03 (d, 2J = 4.8 Hz, 1H, 1-Hb), 1.75 (m, 1H, 6-H), 1.58 (m, 1H, 5-Ha), 1.41 (s, 9H, HBoc), 1.13 (m, 1H, 5-Hb), 0.97 (d, 3J = 6.6 Hz, 3H, CH3CHCH3), 0.94 (d, 3J = 6.6 Hz, 3H, CH3CHCH3) ppm.
Lithium hydroxide (91 mg, 3.8 mmol) was added to a solution of (S)-O-tert-butyldiphenyl-siloxymethyl-N-heptanoyl-seryl methyl ester (890 mg, 1.8 mmol) in a methanol:water (3:1) solution. The reaction was stirred at 5°C for 15 hr. The resulting mixture was poured into H2O with cold 1 N HCl and was extracted with CH2Cl2. The organic layers were combined, washed with brine, dried under Na2SO4, filtered, concentrated, and dried under high vacuum. The product obtained yielded 7 as a yellowish oil. 1H NMR: δ = 7.61 (m, 4H, Ar-H), 7.41 (m, 6H, Ar-H), 6.24 (d, 2J = 7.5 Hz, 1H, NH), 4.69 (m, 1H, 2-H), 4.17 (dd, 2J = 10.4 Hz, 2J = 3.6 Hz, 1H, 3-Ha), 3.89 (dd, 2J = 10.4 Hz, 2J = 3.6 Hz, 1H, 3-Hb), 2.20 (t, 3J = 7.5 Hz, 2H, 2′-H), 1.60 (m, 2H, HHep), 1.29 (m, 6H, HHep), 1.05 (s, 9H, Hbutyl), 0.88 (t, 3J = 6.9 Hz, 3H, 7′-CH3) ppm.
tert-butyldiphenylsilyl chloride (1.95 ml, 7.6 mmol) and imidazole (519 mg, 7.6 mmol) were added to a solution of (S)-N-heptanoyl-serine methyl ester (588.6 mg, 2.5 mmol) in CH2Cl2 (20 ml), and the solution was stirred overnight at room temperature. The resulting mixture was concentrated under reduced pressure and was subjected to column chromatography (5:1 hexane:EtOAc), yielding (S)-O-tert-butyldiphe-nylsiloxymethyl-N-heptanoyl-seryl methyl ester (890 mg, 74%) as a colorless oil. 1H NMR: δ = 7.59 (m, 4H, Ar-H), 7.41 (m, 6H, Ar-H), 6.28 (d, 2J = 8.4 Hz, 1H, NH), 4.70 (m, 1H, 2-H), 4.12 (dd, 2J = 10.1 Hz, 2J = 3.0 Hz, 1H, 3- Ha), 3.89 (dd, 2J = 10.1 Hz, 2J = 3.0 Hz, 1H, 3-Hb), 3.74 (s, 3H, 1-OCH3), 2.11 (t, 3J = 7.7 Hz, 2H, 2′-H), 1.57(m, 2H, HHep), 1.30 (m, 6H, HHep), 1.04 (s, 9H, Hbutyl), 0.88 (t, 3J = 6.7 Hz, 3H, 7′-CH3) ppm.
HBTU (1.83 g, 4.8 mmol), HOBt (0.74 g, 4.8 mmol), and lastly diisopropylethylamine (2.8 ml, 16 mmol) were added to a solution of heptanoic acid (0.46 ml, 3.2 mmol) and H-Ser-OCH3 (0.5 g, 3.2 mmol) in CH2Cl2 (15 ml). The reaction was stirred overnight at room temperature. The resulting mixture was subjected to flash column chromatography (1:2 hexane:EtOAc), yielding (S)-N-heptanoyl-serine methyl ester (588.6 mg, 79%) as a yellowish oil. 1H NMR: δ = 6.47 (b, 1H, NH), 4.69 (m, 1H, 2-H), 3.94 (m, 2H, 3-H), 3.79 (s, 3H, 1-OCH3), 2.27 (t, 3J = 7.6 Hz, 2H, 2′-H), 1.63 (m, 2H, HHep), 1.29 (m, 6H, HHep), 0.88 (m, 3H, 7′- CH3) ppm.
Trifluoroacetic acid (100 μl, 0.87 mmol) was added to a solution of 5 (45 mg, 0.12 mmol) in CH2Cl2 (0.5 ml) at room temperature for 30 min. Subsequently, the concentrated mixture was dried under high vacuum to remove trifluoroacetic acid. The resulting crude product, 6 (33 mg, ca. 100%), was then used in the following coupling reaction without further purification. HBTU (68 mg, 0.17 mmol), HOBt (27 mg, 0.17 mmol), and lastly diisopropylethylamine (104 μl, 0.59 mmol) were added to a solution of 6 (33 mg, 0.12 mmol) and 7 (65 mg, 0.14 mmol) in CH2Cl2 (5 ml). The reaction was stirred overnight at room temperature. The resulting mixture was subjected to flash column chromatography (3:1 hexane:EtOAc) to give (4S)-2-methox-yethoxy-methoxymethyl-4-([S]-O-tert-butyldiphenylsiloxy-methyl-N-heptanoylseryl-amino)-6-methyl-1,2-oxiranyl-heptane (36 mg, 42%). 1H NMR: δ = 7.71 (m, 4H, Ar-H), 7.44 (m, 6H, Ar-H), 7.02 (d, 2J = 8.4 Hz, 1H, 4-NH), 6.17 (d, 2J = 6.6 Hz, 1H, 2′-NH), 4.72 (s, 2H, 2-OCH2O), 4.60 (m, 2H, 4-H, 2′-H), 4.42 (d, 2J = 11.4 Hz, 1H, 2-CHa 2), 4.03 (m, 1H, 3′-Ha), 3.70 (m, 3H, 3′-CHb2, 2-OCH2CH2O), 3.55 (m, 2H, 2-OCH2CH2O), 3.52 (d, 2J = 11.4 Hz, 1H, 2-CHb2), 3.40 (s, 3H, 2-OCH3), 3.29 (d, 2J = 5.4 Hz, 1H, 1-Ha), 3.04 (d, 2J = 4.8 Hz, 1H, 1-Hb), 2.13 (t, 3J = 7.6 Hz, 2H, 2″-H), 1.63 (m, 4H, 6-H, 5-Ha, HHep), 1.26 (m, 6H, HHep), 1.07 (s, 9H, 3′-tBu), 0.96 (d, 3J = 6.3 Hz, 3H, CH3CHCH3), 0.91 (d, 3J = 6.3 Hz, 3H, CH3CHCH3), 0.86 (t, 3J = 7.6 Hz, 3H, 7″- CH3) ppm.
Tetrabutylammonium fluoride (50 μl, 1 M in THF, 0.05 mmol) was added to a solution of (4S)-2-methoxyethoxy-methoxymethyl-4-([S]-O-tert-butyldiphenylsiloxy-methyl-N-heptanoylseryl-amino)-6-methyl-1,2-oxiranyl-heptane (30 mg, 0.042 mmol) in THF (1 ml). The reaction was stirred at room temperature for 1 hr, followed by flash column chromatography (1:2 hexane:EtOAc), yielding 12 (16mg, 80%) as a yellowish oil. 1H NMR: δ = 6.83 (d, 2J = 7.5 Hz, 1H, 4-NH), 6.44 (d, 2J = 7.5 Hz, 1H, 2′-NH), 4.71 (s, 2H, 2-OCH2O), 4.50 (m, 2H, 4-H, 2′-H), 4.41 (d, 2J = 11.7 Hz, 1H, 2-CHa2), 4.08 (m, 1H, 3′-Ha2), 3.68 (m, 2H, 2-OCH2CH2O), 3.55 (m, 3H, 2-OCH2CH2O, 3′-Hb2), 3.46 (d, 2J = 11.7 Hz, 1H, 2-CHb2), 3.40 (s, 3H,2-OCH3), 3.27 (d,2J = 5.1 Hz, 1H, 1-Ha), 3.05 (d,2J = 4.8 Hz, 1H, 1-Hb), 2.22 (m, 2H, 2″-H), 1.60 (m, 4H, 6-H, 5-Ha, HHep), 1.28 (m, 6H, HHep), 0.96 (d, 3J = 3.9 Hz, 3H, CH3CHCH3), 0.94 (d, 3J = 3.9 Hz, 3H, CH3CHCH3), 0.88 (t, 3J = 6.7 Hz, 3H, 7″- CH3) ppm. MS (ESI): m/z = 475, calcd. for C23H42N2O8: m/z = 474.59.
Murine lymphoma EL4 cells and prostate cancer PC3 cells (ATCC) were grown in RPMI medium (GIBCO-BRL) supplemented with 10% fetal bovine serum and 1% penicillin and streptomycin at 37°C in a 5% CO2 incubator. Cells were pretreated with 1 μM biotinylated compounds 30 min prior to the addition of increasing concentrations of either dihydroeponemycin, epoxomicin, or dihydroeponemycin analogs as indicated. The cells were then incubated for an additional hour. Cell lysates were analyzed by 12% SDS-PAGE and were transferred to PVDF membranes. Proteins that were covalently modified by biotinylated compounds were visualized with enhanced chemilumi-nescence by using streptavidin-conjugated horseradish peroxidase (Sigma-Aldrich) or anti-LMP2 (Affinity BioReagents) and Biomax X-ray film (Kodak).
k association values were determined as follows. Inhibitors were mixed with a fluorogenic peptide substrate and assay buffer (20 mM Tris [pH 8.0], 0.5 mM EDTA, and 0.035% SDS) in a 96-well plate. The chy-motrypsin-like activity was assayed by using the fluorogenic peptide substrates Suc-Leu-Leu-Val-Tyr-AMC (Sigma-Aldrich). Hydrolysis was initiated by the addition of bovine 20S proteasome or immunoproteasome (Biomol International), and the reaction was followed by fluorescence (360 nm excitation/460 nm detection) by using a Microplate Fluorescence Reader (FL600; Bio-Tek Instruments, Inc., Winooski, VT) employing the software KC4 v.2.5 (Bio-Tek Instruments, Inc.). Reactions were allowed to proceed for 60–90 min, and fluorescence data were collected every 1 min. Fluorescence was quantified as arbitrary units, and progression curves were plotted for each reaction as a function of time. k observed/[I] values were obtained by using the PRISM program by nonlinear least-squares fit of the data to the following equation: fluorescence =vst+([v0 — vs]kobserved)(1 – exp[–kobserved t]), where v0 and vs are the initial and final velocities, respectively, and kobserved is the reaction rate constant. The range of inhibitor concentrations tested was chosen so that several half-lives could be observed during the course of the measurement. Reactions were performed with inhibitor concentrations that were < 100-fold of those of the proteasome assayed.
Endothelial cell spheroids were generated from human umbilical vein endothelial cells (HUVECs; Cascade Biologicals, Portland, OR) as described . The spheroids (4–6/well) were distributed in 96-well plates in collagen I matrix for the 3D-ECSA. Cell culture medium was added to each well along with 20 ng/ml VEGF in the presence and absence of the individual inhibitor. The 3D cultures were incubated in tissue culture chambers at 37°C in 5% CO2 for 24 hr. Photographic images of spheroids were obtained with the 10× objective of a Nikon TE2000 microscope. Sprouting was quantified from digital images according to our previously published method .
Prostate cancer PC3 and LN3 cells were seeded in 12-well plates. They were then incubated at 37°C until they were 70% confluent before the appropriate inhibitors were added as indicated in increasing concentrations. Cells were treated for 48 hr. Viable cells were counted on a hemocytometerin 0.2% trypan blue solution (Sigma-Aldrich). IC50 values were calculated from sigmoid dose-response curves by the method of nonlinear regression to a logarithmic function. These data represent the average of three or more experiments.
The immunoproteasome catalytic subunits have been implicated in a number of disease states. For example, they have been suggested as potential new drug discovery targets for the treatment of multiple myeloma. However, there are currently no immunoproteasome catalytic subunit-specific inhibitors that can be used to validate these subunits as therapeutic targets. Furthermore, the exact role of the immunoproteasome catalytic subunits in pathogenesis is not clearly understood. Thus, the development of the immunoprotea-some catalytic subunit LMP2-specific inhibitors described in this report may not only hold great potential as a therapeutic agent for certain diseases, but can also provide a valuable chemical genetic probe to investigate immunoproteasome biology.
We are grateful to the Kentucky Lung Cancer Research Program (K.-B.K), for Kentucky Tobacco Research and Development Center grants, for a Research to Prevent Blindness Challenge grant (R.M.), and for the financial support of a University of Kentucky Graduate School Fellowship (Y.K.H.).