B. Wilding et al.
Bioorganic & Medicinal Chemistry Letters 42 (2021) 128050
DG013A 1 (Table 1, Entry 1) is a pseudotripeptide phosphinic acid
focused on the homophenylalanine transition state mimic phosphinic
acid precursor 10. Consistent with previous reported syntheses (Scheme
1, Route A),16,18,23,24 we generated intermediate 10 by phosphine
addition to the prochiral imine precursor 4. Following hydrolysis and re-
protection of the primary amine as a Cbz-group, the racemic phosphinic
acid (rac)-10 was subjected to kinetic resolution via recrystallization
containing 3 stereogenic centers (((S)-2-(((S)-1-amino-3-(1H-indol-3-
yl)-1-oxopropan-2-yl)carbamoyl)-4-methylpentyl)((R)-1-amino-3-phe-
nylpropyl)phosphinic acid, the S,S,R-diastereomer), where only the
tryptophan moiety can be obtained without asymmetric synthesis or
chiral resolution. The R,S,R-diastereomer, which is epimeric at the
central leucine-mimetic position, is described as DG013B 2 (Table 1,
Entry 2) and has been used as a weakly binding negative control.16
Several recent studies have highlighted the importance for ERAP1 af-
finity of the homophenylalanine, leucine and tryptophan mimetic
groups in their interactions with the S1-, S1′- and S2′-pockets, respec-
tively.17,18 Despite the high biochemical affinity of this chemotype re-
ported for both ERAP1 and ERAP2,16 the predicted physicochemical
properties are inconsistent with high cellular passive permeability
(Table 1, Entry 1).19,20 Consequently, it is unclear whether DG013A 1
can be utilized as a chemical probe to investigate the phenotype of
ERAP1 inhibition in cancer cells and immunopeptidomic studies.21
Analysis of a recently published crystal structure of 1 bound to ERAP1 in
the closed conformation (Fig. 1, PDB: 6M8P),22 suggests that although
the ligand is entirely encapsulated by the protein, the tryptophan
moiety-primary amide points towards the large solvent cavity above the
active site and forms no direct interactions. Therefore, we hypothesized
that changes to this region could potentially modulate the physico-
chemical properties of the ligand; whilst maintaining the interactions in
the peptide residue pockets and the polar contacts within the terminal
amine pocket, the zinc ion, the catalytic tyrosine-438 and glycine-317 in
the GAMEN loop. Because of the perceived importance of the stereo-
chemical purity to the binding of 1 to ERAP1, and our desire to make
analogues at the tryptophan primary amide, we re-examined the pub-
lished synthetic route for this complex and challenging target.
from ethanol using S-(-)-α-methylbenzylamine, to generate a diaste-
reomeric salt as had been previously described.18 The enantiopurity of
this compound had only been determined using optical rotation (liter-
ature [
α
]
D
20 = ꢀ 35, c = 1, EtOH),18 which was not suitable to accurately
assess the enantiomeric ratio (er). Therefore, we developed a chiral high
performance liquid chromatography (HPLC) method to measure the er
(see SI for details).25 Following multiple recrystallizations, a stable op-
tical rotation consistent with the literature value was obtained, but our
chiral HPLC analysis demonstrated this kinetic resolution methodology
could only result in an er = 5.8:1 for R-10, consistent with the obser-
vation by Mucha et al. that enantio-enrichment using recrystallization
can be challenging with this intermediate.25b
To improve the enantiomeric purity of the critical precursor R-10, we
developed an enantioselective route using Ellman’s chiral sulfinamide
(Scheme 1, Route B). Condensation between 3-phenylpropanal 6 and
(S)-2-methylpropane-2-sulfinamide 7 gave imine 8, which was then
reacted with (diethoxymethyl)phosphinate, using Rb2CO3 as a base, to
afford diastereomeric phosphinate 9. Following acidic hydrolysis and
subsequent protection of the primary amine as a Cbz-group, the chiral
phosphinic acid R-10 was obtained in high er (up to 12.9:1, 150 mg
scale) as determined by chiral HPLC; although the er obtained for this
transformation was typically dependent on the scale the reaction was
performed, with the er decreasing with larger scale synthesis (8.4:1, 2 g
scale, see SI for details). Recrystallization from ethanol and (S)-
The synthesis of DG013A 1 was first described by Stratikos et al. in
2013,16 although very similar analogues to Cbz-protected primary
amine 3 (Table 1, Entry 3), were reported as inhibitors of matrix met-
alloproteinases in 1999.23 To obtain stereochemically pure 1, we first
(-)-α-methylbenzylamine using this enantioenriched material now gave
the product R-10 as a single observable enantiomer (>99:1 er).
Synthesis of the second chiral center was carried out via a phospha-
Michael addition to acrylate 11. Following a small screen of reaction
Scheme 1. Stereoselective synthesis of the proposed ERAP1 chemical probe DG013A 1. Reagents and conditions: i) H3PO2, EtOH, 80 ◦C, 3 h, 68%; ii) HBr (48%), 100
◦C, 3 h, 72%, iii) a) Cbz-OSu, Na2CO3, THF/H2O (1:1), 20 ◦C, 24 h, 100%; b) (S)-(-)-
-methylbenzylamine, EtOH, multiple recrystallizations 19%, c) 4M HCl aq. 20
◦C, 3 h, 94%, iv) CuSO4, CH2Cl2, 20 ◦C, 87%; v) Rb2CO3, ethyl (diethoxymethyl)phosphinate, CH2Cl2, 20 ◦C, 56%; vi) a) 4M HCl aq., reflux, 93%; b) Cbz-OSu,
α
Na2CO3, THF/H2O, 94%; c) S-(-)-α
-methylbenzylamine, EtOH, multiple recrystallizations; d) 4M HCl aq., 20 ◦C, 33% (over two steps), vii) BSA, 11, 20 ◦C, 16 h;
ix) a) NaOH, EtOH, H2O, 20 ◦C, 24 h, 80% (over 2 steps); xi) a) 33% HBr in AcOH, 88%;b) Boc2O, Et3N, DMF, 78%; xi) c) EDC.HCl, HOBt, DIPEA, CH2Cl2, 20 ◦C, 80%;
d) TFA/CH2Cl2/TIS/H2O, quant.; e) Separation by HPLC, 1 31%, DG013B (2) 24%, viii) a) 12, CH2Cl2, DIPEA, TMSCl, 0 ◦C to RT, 20 h; b) Separation by HPLC, 14
38%; x) H2O2, LiOH aq., 0 ◦C, 87%.
2