2
S. Giovani et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
schizonts or immature merozoites is fatal for the parasite.11
PfSUB1 is released in the PV space just prior to egress where it
mediates the proteolytic maturation of a family of PV proteins
called SERA.10 Moreover, PfSUB1 processes several merozoite sur-
face proteins (MSP1, MSP6, and MSP7) thus priming the merozoite
for the subsequent invasion step.9 Inhibition of PfSUB1 activity or
discharge prevents SERA maturation and blocks egress and the
resulting merozoites are defective in invasion.10,11 These studies
convincingly point to PfSUB1 as a promising drug target to be val-
idated for the development of innovative antimalarial therapies.
To date, few PfSUB1 inhibitors have been described by us and
others.10,12–14 Poor potency, lack of selectivity, or poor cell perme-
ability are the main issues in most cases. Covalent peptidyl
a-
ketoamide inhibitors based on authentic substrates of the protease
have also been recently described.13 With the aim of developing
more potent and selective PfSUB1 inhibitors we started with the
generation of a homology model of the enzyme. Very recently,
during the revision of the manuscript, the PfSUB1-prodomain-
NIMP.M7 Fab complex has been released, PDB ID: 4LVN.15
Originally, to rationally design the novel inhibitors we analyzed
the binding mode of the decapeptide KITAQ;DDEES, derived from
the cleavage sequence of the PfSUB1 substrate SERA4st1 (details
concerning our rational design based on the homology model
was provided in Supplementary data; Figs. S1–S13). Notably, acidic
amino acids are present at the prime-side of all PfSUB1 cleavage
sequences identified so far.13,16 In earlier predictions13,16 and our
model, the P10 aspartate of the decapeptide was predicted to form
an interaction with the PfSUB1 residue K465. Our inhibitor design
approach consisted of maintaining the natural P-side sequence of
the decapeptide, whilst replacing the cleavable peptide bond with
an electrophilic carbonyl group. In order to reproduce in our inhib-
itor the key interaction between the carboxylic function of the P10
residue and K465, a difluorostatone moiety was envisaged as hav-
ing the appropriate length to project an N-terminal carboxylic
group toward K465. Fine-tuning of the distance between the elec-
trophilic carbonyl group and the N-terminal carboxylic moiety was
accomplished through the synthesis of both 1a and 2, differing in
the length of the linker. Once we established the optimal size of
the linker, the molecular weight of inhibitor 1a was progressively
reduced by eliminating amino acid residues at the P-side
(Scheme 1).
Scheme 2. Synthesis of difluorostatones 1a–c and 2. Reagents and conditions: (a)
N,O-dimethylhydroxylamine hydrochloride, EDC, HOBt, DIPEA, dry DCM, 0–25 °C,
17 h; (b) LiAlH4, dry THF, 0 °C, 20 min; (c) ethyl bromodifluoroacetate, zinc dust, dry
THF, from 25 °C to reflux, 30 min; (d) 0.25 N solution of LiOH, MeCN, 25 °C, 2 h; (e)
H-Gly-OBnꢀHCl (for 7a) or H-bAla-OBnꢀHCl (for 7b), HATU, DIPEA, dry THF, from 0 to
25 °C, 14 h; (f) (i) 7a, b, TFA/DCM, 25 °C, 2 h, (ii) 8a–c, EDC, HOBt, TEA, dry DMF, 0–
25 °C, 12 h; (g) Dess–Martin periodinane, anhydrous NMP, from 0 to 25 °C, 1–24 h;
(h) Pd/C 10%, H2, 40 °C, 12–48 h.
of LiOH and the free carboxylic acid was immediately coupled to
glycine or b-alanine benzyl esters in the presence of O-(7-aza-
benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium hexafluorophos-
phate (HATU) and N,N-diisopropylethylamine (DIPEA) in dry
N,N-dimethylformamide (DMF) to afford the desired difluorosta-
tine synthons 7a, b. In the following steps of the synthetic path-
way, statine-derivatives 7a, b were exposed to a 50:50 mixture
of trifluoroacetic acid (TFA) and dichloromethane (DCM) resulting
in Boc-deprotection and formation of the corresponding trifluoro-
acetate salts. These latter compounds were immediately coupled
with peptides 8a–c using EDC and HOBt as coupling agents in
dry DMF. Peptides 8a–c were prepared by means of microwave
assisted solid phase synthesis as described in the Supplementary
data. Oxidation of the resulting alcohols 9a–c and 10 with Dess–
Martin periodinane in anhydrous N-methyl-2-pyrrolidone afforded
difluorostatones 11a–c and 12 in good yields.19 Deprotection of the
benzyl groups was accomplished using Pd/C in a hydrogen atmo-
sphere (1 atm) providing the final compounds 1a–c and 2 in quan-
titative yields.
The synthesis of PfSUB1 inhibitors 1a–c and 2 is described in
Scheme 2. Enantiomerically pure17 aldehyde 4 was obtained from
the commercially available Boc-protected amino acid
LiAlH4-mediated reduction of N,O-dimethylhydroxamide
(Weinreb amide) intermediate. Formation of the Weinreb amide
was accomplished in quantitative yield by coupling and
3 via
a
3
N,O-dimethylhydroxylamine in the presence of 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide (EDC) and hydroxybenzotria-
zole (HOBt). Aldehyde 4 was submitted to the Reformatsky
reaction protocol using ethyl bromodifluoroacetate and activated
zinc in refluxing THF to furnish alcohol 5 as a mixture of diastere-
oisomers.18,19 Ethyl ester 5 was hydrolyzed with a 0.25 N solution
The inhibitory activity of the synthesized inhibitors and
selected intermediates against recombinant PfSUB1 (Table 1) was
assessed using a previously described fluorimetric assay.13 The
design strategy of inhibitors, performed using a homology model
of PfSUB1 built adopting a multiple template-based homology
modeling approach, is reported in the Supporting information
(Figs. S1–S13).13,16,20,21
Since the crystal structure of the PfSUB1 catalytic core (PDB ID:
4LVN15) is now available, we compared it to our previously devel-
oped homology model. The superposition depicted in Figure 1
reveals the good quality of the developed PfSUB1 homology model
and validates our rational design approach. The sole difference
observed in the two models (Fig. 1) is the side-chain orientation
of Y427, which establishes an H-bond with K465 in the experimen-
tal structure.
Scheme 1. Difluorostatone-based inhibitors 1a–c, 2.