A solid-phase approach for the synthesis of α-aminoboronic acid peptides

A solid-phase synthesis of α-aminoboronic acid peptides using a 1-glycerol polystyrene resin is described. Standard Fmoc solid-phase peptide chemistry is carried out to construct bortezomib and ixazomib. This approach eliminates the need for liquid–liquid extractions, silica gel column chromatography, and HPLC purifications, as products are isolated in high purity after direct cleavage from the resin.

a-Aminoboronic acids are currently being investigated for their utility as reversible covalent inhibitors in a diverse range of therapeutic applications (Fig. 1). 1 These compounds' Lewis acidity enables the formation of stable tetrahedral adducts with nucleophilic residues in biological targets (Fig. 2). In 2003, the rst boronic acid drug, bortezomib, was approved for the treatment of multiple myeloma. 2 Ixazomib, a related a-aminoboronic acid inhibitor, was later approved in 2015 for the same indication. 3 Peptidic a-aminoboronic acids, such as bortezomib and ixazomib, have traditionally been assembled using standard peptide coupling techniques, 4 wherein an a-aminoboronic ester is introduced onto a pre-constructed peptide and is subsequently deprotected to unmask the boronic acid. Metalcatalyzed decarboxylative borylation strategies have also been reported for the preparation of a-aminoboronic acid peptides. 5 This approach provides direct access to these compounds from their parent peptide constructs but sacrices stereochemical integrity.
Regardless of the method, a-aminoboronic acid/ester peptides are difficult to prepare for a number of reasons. 6 First, the C-B bond can be oxidatively labile. 7 Second, a-aminoboronic acids and esters containing an unsubstituted aamino group can undergo a spontaneous 1,3-rearrangement (Scheme 1, A); this process can be minimized or suppressed entirely if the amino group is rapidly acylated or protonated. 8 Third, boronic esters can be hydrolytically labile, especially at low pH (Scheme 1, B). 6,9 Therefore, any multistep approach must entail careful extractive workups and purications to ensure that the ester remains intact.  While solid-phase peptide synthesis (SPPS) has become a standard method for the construction of peptides, 10 this technology has remained underexplored for the preparation of a-aminoboronic acid peptides. 11 An approach of this type could eliminate liquid-liquid extractions and HPLC purications and could enable high-throughput access to this class of compounds. To the best of our knowledge, there has only been one report of C-terminal SPPS to generate a-aminoboronic acid peptides (Scheme 2). 12 Although this study provides a critical conceptual foundation, the approach it describes lacks the simplicity of a traditional solid-phase approach, requiring a complex 8-step synthesis to prepare resin-bound a-aminoboronic ester 6 for SPPS. This limitation may preclude its use as a general strategy for the preparation of a-aminoboronic acid peptides.
We sought to identify an approach that could enable access to resin-bound a-aminoboronic acids for SPPS in a limited number of steps using the emerging supply of commercially available a-aminoboronic acid building blocks. The Klein group recently described the use of a 1-glycerol polystyrene resin that could be used for Fmoc SPPS to construct boronic acid-containing peptides. [13][14][15] These results prompted us to explore the use of this resin for preparation of a-aminoboronic acid peptides, specically bortezomib and ixazomib.
Considering the unique reactivity of a-aminoboronic acids, we needed to devise a concise loading strategy that would suppress the potential for C to N boron migration. This required the amine to remain protonated or acylated throughout the loading process. These considerations lead to the design of a two-step loading protocol (Scheme 3). Commercially available boroleucine pinanediol ester 7 was hydrolysed with aqueous HCl. The boroleucine salt (8) was isolated in quantitative yield, free of pinanediol impurities, aer a simple liquid-liquid extraction. The crude boroleucine salt was then shaken with the 1-glycerol polystyrene resin (loading capacity 0.6 mmol g À1 ), 13 Fmoc chloride, and N,N-diisopropylethylamine to provide resinbound Fmoc-protected boroleucine 9.

Conclusions
This proof-of-concept study details the use of the 1-glycerol polystyrene resin for C-terminal SPPS of a-aminoboronic acid peptides. The a-aminoboronic acid used in this study was loaded onto the resin in 2 steps from a commercially available building block. This loading protocol represents a dramatic improvement to what has been reported previously. While the efficiency of the solid-phase synthesis of bortezomib (7 steps from 7, 54% yield) and ixazomib (7 steps from 7, 49% yield) is comparable to other standard synthetic approaches, 4,5 the use of a solid support enables a-aminoboronic acid peptides to be constructed rapidly in high purity and eliminates the need for workup and purication.

General
All commercial reagents and anhydrous solvents were used without additional purication. Boroleucine pinanediol ester 7 was purchased from Ark Pharm (Cat. # AK-44948, CAS # 179324-87-9). 1-Glycerol polystyrene resin was purchased from Iris Biotech (Cat. # BR-5206.0025). Nuclear magnetic resonance (NMR) spectra were acquired on a Bruker Avance DPX400 operating at 400 and 100 MHz for 1 H and 13 C, respectively, and are referenced internally according to residual solvent signals. NMR data were processed using MNova soware and recorded as follows: 1 H-NMRchemical shi (d, ppm), multiplicity (s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet), coupling constant (Hz), and integration; 13 C-NMRchemical shi (d, ppm). High-resolution mass spectra (HRMS) were recorded on a Thermo Scientic Orbitrap Q Exact mass spectrometer. Reactions were monitored by a Shimadzu LCMS/UV system with LC-30AD solvent pump, 2020 MS, Sil-30AC auto sampler, SPD-M30A UV detector, CTO-20A column oven, using a 2 -98% acetonitrile/0.1% formic acid (or 0.001% ammonia) gradient over 2.5 minutes. Purity was determined by LCMS analysis with an Agilent 1290 UHPLC coupled with Agilent MSD (6140) mass spectrometer using ESI as ionization source. The LC separation used a Phenomenex XB-C18, 1.7 mm, 50 Â 2.1 mm column with a 0.4 mL min À1 ow rate. Solvent A was water with 0.1% FA and solvent B was acetonitrile with 0.1% FA. The gradient consisted of 2 -98% solvent B over 7 min and hold 98% B for 1.5 min following equilibration for 1.5 min. The LC column temperature was 40 C. UV absorbance was collected at 220 nm and 254 nm.
Step 2. Fmoc deprotectionpiperidine (10% in DMF, 20 mL, 21.8 mmol) was added to the resin and the suspension was shaken for 20 minutes. The resin was dried by vacuum ltration and washed with DMF (2 Â 10 mL). This procedure was repeated to ensure complete deprotection.
Step 4. Cleavage from resinthe resin was dried by vacuum ltration, washed with DMF, CH 2 Cl 2 , and Et 2 O (3 Â 10 mL for each), and dried under vacuum. The dried resin was treated with 9 : 1 THF/water (40 mL) and stirred overnight at room temperature. The resin was ltered and the ltrate concentrated under vacuum to afford the pure a-aminoboronic acid.

Conflicts of interest
There are no conicts to declare.