Synthesis and SAR Study of Novel Peptide Aldehydes as Inhibitors of 20S Proteasome

Based on the analysis of the crystal structure of MG101 (1) and 20S proteasomes, a new series of peptide aldehyde derivatives were designed and synthesized. Their ability to inhibit 20S proteasome was assayed. Among them, Cbz-Glu(OtBu)-Phe-Leucinal (3c), Cbz-Glu(OtBu)-Leu-Leucinal (3d), and Boc-Ser(OBzl)-Leu-Leucinal (3o) exhibited the most activity, which represented an order of magnitude enhancement compared with MG132 (2). The covalent docking protocol was used to explore the binding mode. The structure-activity relationship of the peptide aldehyde inhibitors is discussed.


Introduction
Lysosomes and the ubiquitin-proteasome pathway (UPP) are two major routes for cellular protein degradation [1][2][3]. The UPP is essential for many cellular regulatory mechanisms and plays a crucial role in the regulation of many physiological processes. For example, degradation of the p53 tumor suppressors [4] and inhibition of cyclin-dependent p27 kinases [5] can promote tumorigenesis, disorders of protein degradation that originated from UPP can cause the development of many human diseases, such as cancer, Alzheimer's and Parkinson's diseases, etc. [6][7][8]. Recently, the study of OPEN ACCESS proteasome inhibition has received much attention [9][10][11][12][13][14]. In UPP, proteolysis takes place in the 26S proteasome, which consist of one or two 19S regulatory particles (RP) [15] and a central catalytic particle known as the 20S proteasome (CP). The 20S proteasome is a large cylindrically-shaped complex composed of two copies of seven distinct α-and seven distinct β-type subunits [16,17]. It possesses three protease activities, namely the post-glutamyl-peptide hydrolyzing (PGPH), the trypsinlike (T-L), and the chymotrypsin-like (ChT-L) activity, which are assigned as the active subunits β1, β2, and β5, respectively [18,19].
Small molecules, have been developed to inhibit the proteasome such as cyclic peptides [10,[20][21][22], peptide boric acids [23], peptide epoxides [24], peptide vinyl sulfones [25], and nonpeptidic molecules [26][27][28][29]. Among all the proteasome inhibitors ever studied, peptide aldehydes were the first developed and are still the most widely used in in vitro and in vivo studies [30]. MG101 (1, Ac-Leu-Leu-nLeu-al, Figure 1), one of calpain inhibitors, is the first well-known 20S proteasome inhibitor [16,17,31]. The crystal structure of the 20S proteasome in complex with MG101 confirms that the hydroxyl group of the N-terminal threonine of the β5 subunit reacted with the aldehyde group and formed a reversible hemiacetal. MG132 (2, Cbz-Leu-Leu-Leu-Al, Figure 1), a more potent and selective analog of MG101, which bears a benzyloxycarbonyl group instead of an acetyl group, is one of the most commonly used synthetic proteasome inhibitors [32,33]. Up to now, many peptide aldehydes have been designed and synthesized [34,35].  Previous studies have demonstrated that hydrophobic groups around the P1 and P3 positions are beneficial to enhance the activity of peptide aldehydes 3 [36][37][38]. Bulky substituents at the P2 position and aromatic groups at P4 position also contribute to enhance the inhibitory activity [37]. According to the crystal structure of complexed MG101 and 20S proteasome, the leucine side chain of P3 projects into the S3 pocket of β5 subunit, which is an open space in the vicinity of the isopropyl groups. Since the P3-leucine moiety only partially fills the S3 pocket, we supposed that introducing a large group at P3 to fill the open space might enhance inhibitory activity. Thus, in this study, we mainly focus on the variation of P3 position to reveal the structure-activity relationships. A series of peptide aldehyde derivatives are designed which have a bulky P3 moiety aiming to increase the hydrophobic interactions with S3.

Synthesis of Peptide Aldehydes 3a-r
The synthesis of the peptide aldehydes is shown in Scheme 1. L-Leucine (4) was treated with NaBH 4 and I 2 under argon to give L-leucinol (5) in 89% yield [39], which was then coupled with Boc-protected amino acids to form the dipeptide alcohols 6 in 71%-80% yield. The dipeptide alcohold were deprotected with 20% trifluoroacetic acid in dichloromethane, followed by reaction with t-butoxy-carbonyl (Boc)-or benzyloxycarbonyl (Cbz)-protected amino acids to give 8a-r (crude products were used in the next step without further purification). After Swern oxidation [40], compounds 3a-r were obtained in 49%-59% yields.

Assays for Proteasome Activities and SAR
Inhibitory activities of peptide aldehydes on the 20S proteasome are assayed in vitro [41]. MG132 was used as the positive control ( Table 1). The results indicate that most of the peptide aldehydes exhibited inhibitory activities against ChT-L, which is closely associated with the substituted amino acids at P3. Out of 17 synthesized compounds, nine exhibited inhibitory activities with IC 50 in the nM range, and three compounds in particular (3c, 3d, and 3o) demonstrated much higher activities than the control MG132.
These inhibitors can be classified into Cbz and Boc series, based on the moiety at the P4 position (R 4 ). Among the Cbz series, a P3 residue with a bulky hydrophobic branch (compounds 3a-3e, 3h) affords a highly active inhibitor, whereas, the electropositive branches (compounds 3f and 3g) show dramatically decreased activity. Compounds 3c and 3d exhibited about 10-fold higher activity than the others, indicating that a relatively long bulky side chain may favor the increase of activity. For the Boc series, the P3 residue with a hydrophobic bulky side chain generally affords an active ChT-L inhibitor. The activities of Boc series compounds with different P3 residues go in an order of 3o > 3p > 3i and 3k > 3j > 3l, also showing that a bulky side chain is too long to give higher activity. Both 3m and 3n show poor activities with IC 50 > 50 μM, and the reason for this deceased potency might be derived from the presence of a proline pyrrolidine moiety at the P3 position, which is consistent with reported results [37]. Furthermore, in contrast to the Cbz series, in which the activity does not vary obviously with the side chain of P2, in the Boc series, a benzyl group at P2 (3i and 3k) affords much higher activity than the corresponding isobutyl branched compound (3j and 3l). Table 1. Inhibition of peptide aldehydes to ChT-L activity of 20S proteasome.
Cbz To fully understand the SAR of inhibitors, we constructed a binding mode of the peptide aldehydes with the β5 subunit of the 20S proteasome based on the crystal structure of 20S proteasome complexed with MG101. Though docking and biochemical data are often not easily comparable, the insights gained into the binding behavior by molecular modeling is meaningful. Given that covalent binding is a unique feature of peptide aldehyde inhibitors, we adopted a covalent docking approach and then developed a protocol to investigate the binding mode of peptide aldehydes with the 20S proteasome. The binding mode of the control MG132 is similar to that of MG101 observed in the crystal structure (Figure 2a). MG132 adopts a β-conformation and fills the gap between strands S2 and S4 by forming hydrogen bonds with residues Thr21, Gly47, and Ala49 and generating an anti-parallel β-sheet structure (Figure 2b) [10]. The P1-leucine side chain of MG132 projects into the S1 pocket and the P2-leucine side chain is towards outside. The P3-leucine side chain stretches out into the subunit-specific S3 pocket and is in close contact with residues of the adjacent β6 subunit.
Other peptide aldehydes are also docked into the 20S proteasome using the same protocol. Similar orientations of P1-P4 residues are found in the docked conformations. For example, the P1-P4 residues of 3c are towards the S1-S4 pockets, respectively, like in MG132 (Figure 2a), and so do the hydrogen bonds (Figure 2b). Biochemical investigation shows that the size and length of the P3 side chain is crucial to activity [42,43]. Among the Cbz series compounds, Glu(O t Bu) residues at P3 (3c and 3d) give the most active inhibition. The structure of the 20S proteasome shows that the β5 and β6 subunits constitute the binding cleft of the S3 pocket, which is able to accommodate long and linear side chains. The docking results show that the tert-butyl glutamic ester (3c) fits this site better than tert-butyl aspartic ester (3a) and provides a strong hydrophobic interaction with the β5/β6 interface (Figure 2c). Among the Boc series of compounds, when a phenyl ester was used to replace a tert-butyl ester at P3, the Asp(OBzl) residue (3j) exhibited more active inhibition than Glu(OBzl) (3l). According to the docking analysis (Figure 2d), although the phenyl ester of both Asp(OBzl) and Glu(OBzl) project into the S3 pocket, the large sized benzene ring makes the conformation rigid and pushes the backbone slightly out of the original orientation, so the shorter side chain of Asp(OBzl) is more suitable for the cleft. The most suitable length of side chain in this Boc-series is Ser(OBzl), and it gives most active inhibition to ChT-L. When the residue at the P3 position is changed to proline (3m and 3n), the results show that the pyrrolidine moiety projects into the S3 pocket (Figure 2e), which makes the binding model of the main chain different from that of MG132 (Figure 2f), and resulting in the disappearance of activity.

General
Unless specified otherwise, all starting materials and reagents were used as obtained from commercial suppliers without further purification. Thin layer chromatography was performed using silica gel GF-254 (Qing-Dao Chemical Company, China) plated with detection by UV, and column chromatography was performed on silica gel (200-300 mesh, Qing-Dao Chemical Company, China). Optical rotations were recorded on a Perkin-Elmer 243B polarimeter. 1 H-NMR (300 or 500 MHz) spectra was recorded on Varian VXR-300 and Varian Inova VXR-500 spectrometer. Mass spectra (ESI-TOF + MS) was obtained on a MDS SCLEX QSTAR instrument and only the most representative peaks were reported (m/z).

Synthesis
L-Leucinol (5) [39]. Sodium borohydride (1.42 g, 37 mmol, 2.4 eq.) was dissolved in anhydrous THF (40 mL) and L-leucine (2.00 g, 15 mmol, 1 eq.) was added in one portion. The solution was cooled to −5 °C in an ice-salt bath, and a solution of iodine (3.87 g, 15 mmol, 1 eq.) in anhydrous THF (10 mL) was added dropwise over 40 min. After the gas evolution was ceased, the reaction solution was refluxed for 16 h. The solution was cooled to room temperature and methanol was added cautiously until the mixture became clear. After stirring 30 min, the solution was evaporated and the residue was dissolved by addition of 45 mL aqueous NaOH. The solution was stirred for 2.5 h and extracted with methylene chloride (30 mL × 4). The combined organic extracts were dried over Na 2 SO 4 and concentrated, affording crude product which was distilled under reduced pressure to yield colorless oil Boc-L-Phenylalanine-L-Leucinol (6a). Boc-L-phenylalanine (0.25 g, 0.95 mmol, 1.0 eq.), L-leucinol (0.20 g, 1.04 mmol, 1.0 eq.), and HOBt (0.14 g, 1.0 mmol, 1.1 eq.) were mixed in anhydrous THF (2 mL). N,N`-Dicyclohexylcarbodiimide (DCC, 0.22 g, 1.0 mmol, 1.1 eq.) and 5 (0.11 g, 0.95 mmol, 1 eq.) was added at 0 °C, and the mixture was warmed to room temperature and stirred for 16 h. After filtration to remove dicyclohexylurea, the solvent was removed and the residue was partitioned between EtOAc (20 mL) and H 2 O (10 mL). The organic phase was washed with 10% citric acid (10 mL × 3), saturated NaHCO 3 (10 mL × 3), and then brine (10 mL × 2). The solution was dried over Na 2 SO 4 and evaporated to an amorphous solid. The crude product was purified by flash chromatography on silica gel to give compound 6a as white solid (0. TFA L-Phenylalanine-L-Leucinol (7a). To a suspension of 6a (0.50 g, 1.27 mmol) in CH 2 Cl 2 (3 mL) was added TFA (1 mL) at 0 °C. After stirred at room temperature for 2 h, the solution was evaporated and the crude product was used in the next step without purification. TFA L-Leucine-L-Leucinol (7b) was prepared by a similar procedure.

Molecular Docking
The covalent docking method with Gold 4.0: A radius of 20 Å from the β5-catalytic N-terminal threonine was used to direct site location. For each of the genetic algorithm runs, a maximum number of 100,000 operations were performed on a population of 100 individuals with a selection pressure of 1.1. Operator weights for crossover, mutation, and migration were set to 95, 95, and 10, respectively, as recommended by the authors of the software. 50 GA runs were performed in each docking experiment as done in the software validation procedure. The default GOLD fitness function was used to identify the better binding mode. The distance for hydrogen bonding was set to 2.5 Å and the cut-off value for van der Waals calculation was set to 4 Å. Covalent docking was applied and the terminal carbonyl carbon of all the ligands have been bonded to the hydroxyl oxygen of Thr1.

Conclusions
Based on the binding analysis of proteasome and its inhibitor, a new series of peptide aldehydes was designed and synthesized. Their abilities to inhibit the 20S proteasome were assayed and the results show that some compounds have more potency than the positive control MG132. Covalent docking was used to simulate the binding of the peptide aldehyde compounds with 20S, and the docking mode is similar to that of the observed crystal complex and that the P3-postion substitutes are crucial for inhibitor potency. The suggested binding mode provides a potential way to design more potent inhibitors of the 20S proteasome.