Optimization and Structure-Activity Relationships of Phosphinic Pseudotripeptide Inhibitors of Aminopeptidases That Generate Antigenic Peptides

Article abstract of DOI:10.1021/acs.jmedchem.6b01031

The oxytocinase subfamily of M1 aminopeptidases, consisting of ER aminopeptidase 1 (ERAP1), ER aminopeptidase 2 (ERAP2), and insulin-regulated aminopeptidase (IRAP), plays critical roles in the generation of antigenic peptides and indirectly regulates human adaptive immune responses. We have previously demonstrated that phosphinic pseudotripeptides can constitute potent inhibitors of this group of enzymes. In this study, we used synthetic methodologies able to furnish a series of stereochemically defined phosphinic pseudotripeptides and demonstrate that side chains at P₁′ and P₂′ positions are critical determinants in driving potency and selectivity. We identified low nanomolar inhibitors of ERAP2 and IRAP that display selectivity of more than 2 and 3 orders of magnitude, respectively. Cellular analysis demonstrated that one of the compounds that is a selective IRAP inhibitor can reduce IRAP-dependent but not ERAP1-dependent cross-presentation by dendritic cells with nanomolar efficacy. Our results encourage further preclinical development of phosphinic pseudotripeptides as regulators of adaptive immune responses.

Full text of DOI:10.1021/acs.jmedchem.6b01031

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Optimization and Structure-Activity Relationships
of Phosphinic Pseudotripeptide Inhibitors of
Aminopeptidases that Generate Antigenic Peptides
Paraskevi Kokkala, Anastasia Mpakali, François-Xavier Mauvais, Athanasios Papakyriakou, Ira
Daskalaki, Ioanna Petropoulou, Sofia Kavvalou, Mirto Papathanasopoulou, Stefanos Agrotis,
Theodora-Markisia Fonsou, Peter van Endert, Efstratios Stratikos, and Dimitris Georgiadis
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01031 • Publication Date (Web): 08 Sep 2016
Downloaded from http://pubs.acs.org on September 11, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Optimization and StructureꢀActivity Relationships
of Phosphinic Pseudotripeptide Inhibitors of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Aminopeptidases that Generate Antigenic Peptides.
†
‡
§
‡
Paraskevi Kokkala, Anastasia Mpakali, FrancoisꢀXavier Mauvais, Athanasios Papakyriakou,
†
†
†
†
Ira Daskalaki, Ioanna Petropoulou, Sofia Kavvalou, Mirto Papathanasopoulou, Stefanos
†
†
§
*,‡
Agrotis, TheodoraꢀMarkisia Fonsou, Peter van Endert, Efstratios Stratikos and Dimitris
*
,†
Georgiadis
†
Department of Chemistry, Laboratory of Organic Chemistry, University of Athens,
‡
Panepistimiopolis, Zografou, 15771, Athens, Greece, National Center for Scientific Research
§
‘
Demokritos’, Agia Paraskevi Attikis, GR 15310, Greece and Institut National de la Santé et de
la Recherche Médicale, Unité1151; Université Paris Descartes, Sorbonne Paris Cité; Centre
National de la Recherche Scientifique, Unité 8253, 75015 Paris, France.
ABSTRACT
The oxytocinase subfamily of M1 aminopeptidases, consisting of ER aminopeptidase 1
(
ERAP1), ER aminopeptidase 2 (ERAP2) and insulinꢀregulated aminopeptidase (IRAP), plays
critical roles in the generation of antigenic peptides and indirectly regulates human adaptive
immune responses. We have previously demonstrated that phosphinic pseudotripeptides can
constitute potent inhibitors of this group of enzymes. In this study, we used synthetic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
methodologies able to furnish a series of stereochemically defined phosphinic pseudotripeptides
and demonstrate that sideꢀchains at P ′ and P ′ positions are critical determinants in driving
1
2
potency and selectivity. We identified low nanomolar inhibitors of ERAP2 and IRAP that
display selectivity of more than 2 and 3 orders of magnitude, respectively. Cellular analysis
demonstrated that one of the compounds that is a selective IRAP inhibitor can reduce IRAPꢀ
dependent but not ERAP1ꢀdependent crossꢀpresentation by dendritic cells with nanomolar
efficacy. Our results encourage further preꢀclinical development of phosphinic pseudotripeptides
as regulators of adaptive immune responses.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Antigen processing aminopeptidases (APAs) trim antigenic peptide precursors and generate
1
peptides for binding onto Major Histocompatibility Complex Class I molecules (MHCI). The
peptideꢀMHCI complex is then transported to the cell surface of somatic cells and presented to
circulating T lymphocytes. Interactions between specialized Tꢀcell receptors and peptideꢀMHCI
complexes are used to determine if the cell is infected or otherwise aberrant, initiating
2
biochemical cascades that lead to apoptosis. Three major APAs have been identified to date: ER
Aminopeptidase 1 (ERAP1), ER Aminopeptidase 2 (ERAP2) and InsulinꢀRegulated
3
Aminopeptidase (IRAP). The first two are localized in the ER and trim antigenic peptide
2
precursors translocated into the ER but generated in the cytosol. IRAP trims antigenic peptide
4
precursor generated by endocytosed proteins in dendritic cells for crossꢀpresentation. These
three enzymes are highly homologous (50% sequence identity on average) and have conserved
active sites, although key aminoꢀacid differences defining the specificity pockets can account for
5
different substrate preferences. Structural differences between the three enzymes also include a
ACS Paragon Plus Environment

Page 3 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5b
unique orientation of the GAMEN motif in IRAP and an overall conformational change that
reorganizes the structure from an open to a closed form, affecting the folding of the active site
6
and nearby specificity pockets.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The enzymatic activity and expression level of APAs, frequently affected by polymorphic
variation, has been repeatedly associated with predisposition to autoimmunity, infections as well
7
as cancer immune evasion. Functional studies have established that alterations in the enzymatic
activity of APAs can result in changes in the presented antigenic peptides and concomitant
8
cytotoxic responses by Tꢀcells. Furthermore, the activity of ERAP1 has been also linked to the
9
function of innate immune responses. Accumulating evidence has therefore established the
1
0
tractability of these enzymes as pharmaceutical targets for the regulation of immune responses.
We have previously demonstrated that a phosphinic pseudotripeptide, 6e (DG013A), can act as
a highly potent inhibitor of all three APAs and is able to regulate antigen presentation in cellular
1
1
systems (Figure 1). This nanomolar potency inhibitor was shown to be able to enhance antiꢀ
cancer cytotoxic responses and to downꢀregulate Tꢀcell mediated and innate inflammatory
responses, indicating that this class of compounds holds promise for pharmaceutical
1
1ꢀ12
applications.
Low selectivity between the three APAs however, limits options for preꢀclinical
development. However, since this inhibitor utilizes the three first specificity pockets of these
aminopeptidases, as it was revealed by Xꢀray crystallographic analysis of 6e/ERAP2 complex
1
1
(Figure 1), varying the sideꢀchains that occupy these pockets can be a valid approach for
optimizing both selectivity and potency. The ability of phosphinic pseudopeptides to successfully
discriminate active sites with high structural homology and functional similarity has been wellꢀ
demonstrated in several demanding cases in the past, such as the development of domainꢀ
13
selective inhibitors of angiotensinꢀconverting enzyme and a selective subꢀnanomolar inhibitor
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
14
of MMPꢀ12. This property of phosphinic peptides, as compared to other classes of protease
inhibitors, is attributed to the weak zincꢀbinding ability of phosphinic group which allows for
inhibitor affinity to be determined by proper optimization of weaker but more specific enzymeꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15
ligand interactions and not by the dominance of tight chelating interactions with the zinc ion.
Figure 1. Structure, in vitro inhibition data and schematic representation of 6e in the active site
of ERAP2 (PDB ID: 4JBS), illustrating the three interacting subsites that are colorꢀcoded per
residue type (white: non polar, pink: polar, red: acidic, blue: basic). The inhibitor is shown as
sticks with green C, blue N red O, magenta P, and the catalytic Zn(II) as a yellow sphere.
In this study, we focused on the structureꢀactivity relationships (SAR) of P ′ and P ′ sideꢀ
1
2
chains in driving potency and selectivity for APAs, by synthesizing a series of stereochemically
defined phosphinic pseudotripeptide derivatives. In vitro evaluation and molecular modeling
revealed that while aromatic residues at P ′ position are critical for achieving potency for
2
ERAP1, hydrophilic residues at the same site can enhance selectivity for ERAP2. Conversely,
although small P ′ sideꢀchains are important for achieving potency, bulkier groups can be well
1
tolerated only by IRAP resulting in selective inhibitors. We identified several new nanomolarꢀ
ACS Paragon Plus Environment

Page 5 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
affinity selective inhibitors for ERAP2 and IRAP that displayed selectivity for each enzyme of
over 2 and 3 orders of magnitude respectively. Proofꢀofꢀconcept cellular analysis using a crossꢀ
presentation model system indicated that one of the selective IRAP inhibitors can reduce IRAPꢀ
dependent but not ERAP1ꢀdependent crossꢀpresentation by dendritic cells with nanomolar
efficacy. Our results encourage further preꢀclinical development of phosphinic pseudotripeptides
for therapeutic applications.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Inhibitor design: In previous studies, it has been demonstrated that 6e influences with high
efficacy antigen presentation in several cellꢀbased systems. For this reason, we were based on the
core structure of 6e and used it as a template for screening various modifications at P ′ and P ′
1
2
positions, aiming to enhance potency and selectivity. The factor of stereochemical purity was set
as a priority, even though available synthetic possibilities to achieve this goal are extremely
16
limited. This requirement can be critical for the accurate evaluation of structure–activity
relationships and the discovery of improved inhibitors, especially in terms of selectivity.
Interestingly, inversion of the selectivity exhibited by different diastereoisomers has been
recently highlighted in the case of phosphinic tripeptide inhibitors of neprilysin (NEP) and
1
7
endothelinꢀconverting enzyme (ECEꢀ1). In a noteworthy example from this report, the 20ꢀfold
selectivity for NEP achieved by using a mixture of P ′ꢀepimers of a phosphinic tripeptide,
1
dropped down to a 5ꢀfold selectivity for NEP when the “natural” P ′ꢀepimer was employed and
1
this was inversed to a 500ꢀfold selectivity for ECEꢀ1 when the “unnatural” P ′ꢀepimer was
1
examined. Notably, most of the studies related to phosphinic peptide inhibitors of Znꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
18
aminopeptidases involve evaluation of diastereoisomeric mixtures, with only a few
19
exceptions. To our knowledge, in the only relevant systematic study a large selection of
stereochemically pure phosphinic tripeptides was evaluated but all of them were isolated after
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
19a
separation by RPꢀHPLC of the final diastereoisomeric mixtures of inhibitors. In our case,
considering the 2ꢀ3 orders of magnitude difference in affinity between two different
11
diastereoisomers of 6e (epimers at P ′ position, 6e and 6e′) and with the aim to eliminate
1
possible inaccuracies during SAR analysis, we devised appropriately adjusted synthetic plans to
control the stereochemical purity and identity of final compounds. Most of the inhibitors
presented herein were obtained as single diastereoisomers directly from the synthesis, using lateꢀ
stage diversification approaches on stereochemically pure building blocks. In order to explore
the S ′ cavity of target aminopeptidases, we based our synthesis on a wellꢀstudied postꢀ
1
modification protocol of phosphinic peptides, which has been developed several years ago in our
2
0
laboratory. This strategy affords diverse, linear isoxazoleꢀsubstituted P ′ sideꢀchains and has
1
been successfully applied in the past to the discovery of inhibitors for other families of Znꢀ
1
4, 17a, 21
metalloproteases,
but not for aminopeptidases. Finally, nonꢀlinear P ′ substitutions were
1
pursued in order to estimate the effect of expanding the bulk of sideꢀchains towards more than
one direction. This differentiation in the spatial installation of bulky groups at P ′ position
1
revealed significant variations between inhibitors with linear “extended” and nonꢀlinear
“expanded” P ′ sideꢀchains, as it will be discussed below.
1
Chemistry: For the synthesis of compounds 6a-k, our efforts were focused on the preparation of
parent phosphinic building block 4 in a stereochemically pure form (Scheme 1). Stereochemical
purity was considered essential during our design in order to extract accurate structureꢀactivity
relationships that would eventually reveal the structural determinants responsible for high
ACS Paragon Plus Environment

Page 7 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potency and selectivity. Given the lack of general synthetic methods able to provide
15
stereochemical control during the formation of P ′ stereogenic center of phosphinic peptides,
1
we explored the possibility of resolving diastereoisomeric mixtures by selective crystallization, a
technique that has previously succeeded in several cases of phosphinic dipeptide and tripeptide
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
7a, 20, 22
isosters.
In this regard, the preparation of enantiomerically pure aminophosphinic acid 1
by a scalable method was an absolute requirement for accomplishing our goal. It has been
previously described in the literature that optically pure 1 can be obtained by HPLC enantiomeric
2
3
separation of the corresponding racemic mixture, however this approach is limited to mg
19b
19b
quantities.
In contrast with earlier reports,
we were pleased to find that resolution of
diastereoisomeric salts of racemic 1 with (S)ꢀ(ꢀ)ꢀαꢀmethylbenzylamine by crystallization,
24
following the protocol of Baylis et al., was highly efficient. By this procedure, the
2
3
20
D
laevorotatory (R)ꢀisomer (1) [salt with (S)ꢀ(ꢀ)ꢀamine: [a]
= ꢀ22.3 (c = 1, EtOH), free acid:
2
0
[
a]D = ꢀ35.7 (c = 1, EtOH)] was successfully isolated in a multigram quantities. PꢀMichael
25
addition of 1 to ethyl 4ꢀmethylꢀ2ꢀmethylenepentanoate using silylating conditions afforded 2 as
a mixture of two diastereoisomers in nearly equimolar ratio (Scheme 1). All attempts of
diastereoisomeric separation by crystallization at this stage failed, therefore compound 2 was
subjected to saponification and crystallization efforts were repeated at racemic diacid 3. Two
recrystallizations by ethyl acetate proved sufficient for the gramꢀscale isolation of the less
3
1
soluble 3 (R,S) isomer with dr >95%, as determined by P NMR. Finally, the Cbz group was
exchanged by Boc following a twoꢀstep procedure. This exchange was considered necessary due
to the advantageous behavior of Boc group during the final deprotection of target tripeptides, as
compared to Cbz group.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 1. Synthesis of Phosphinic Building Block 4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
o
o
Reagents and conditions: (a) HMDS, H C=C[CH CH(CH ) ]COOEt, 110 C, 1 h, then 90 C,
2
2
3 2
o
+
3
h, then EtOH, 70 C, 30 min; (b) NaOH, EtOH, H O, rt, 24 h, then H O ; (c) 2 ×
recrystallizations by AcOEt, 46%, three steps; (d) HBr/AcOH 33%, rt, 1 h; (e) Boc O, Et N,
2 3
2
3
DMF, rt, 24 h, 92%, two steps.
With the stereochemically pure building block 4 in hand, the synthesis of inhibitors 6aꢀk,
required a coupling protocol that does not involve prior protection of the hydroxyphosphinyl
moiety. In this regard, we employed a previously described, epimerizationꢀfree protocol based on
2
0, 22a
the EDC/HOBt coupling system (Scheme 2).
During optimization, high concentrations (0.5
mmol substrate/1 mL solvent) were found essential for efficient coupling reactions. Aiming to
reach the final tripeptides 6a-k in a single deprotection step, aminoacid carboxamides 5aꢀk with
acidꢀlabile protecting groups (if necessary) were employed (see Supporting Information for
experimental details). Compound 6e obtained by this procedure was found spectroscopically and
chromatographically identical to previously described inhibitor 6e (whose stereochemistry was
1
1
confirmed by Xꢀray crystallographic analysis), which confirms both the stereochemical purity
and identity of building block 4 used for the synthesis of inhibitors of type 6.
ACS Paragon Plus Environment

Page 9 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 2. Synthesis of P ′ Diversified Phosphinic Pseudotripeptide Inhibitors 6a-k
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) HꢀXaaꢀNH (5a-k), [Xaa for 5a: LꢀAla, 5b: LꢀLeu, 5c: LꢀPhe,
2
5
d: LꢀPro, 5e: LꢀTrp, 5f: LꢀTyr(tBu), 5g: LꢀSer(TBS), 5h: LꢀLys(Boc), 5i: LꢀHis(Boc), 5j: Lꢀ
Glu(tBu), 5k: DꢀPhe], EDCꢁHCl, HOBt, DIPEA, CH Cl , rt, 2 – 4 h; (b) TFA/CH Cl /TIS/H O
48:49:2:1, rt, 2h, [yields for two steps, for 6a (LꢀAla): 65%, 6b (LꢀLeu): 59%, 6c (LꢀPhe): 32%,
d (LꢀPro): 41%, 6e (LꢀTrp): 32%, 6f (LꢀTyr): 55%, 6g (LꢀSer): 43%, 6h (LꢀLys): 56%, 6i (Lꢀ
His): 10%, 6j (LꢀGlu): 52%, 6k (DꢀPhe): 31%].
2
2
2
2
2
6
For the structural diversification of P ′ position of phosphinic tripeptides, we employed a
1
20
postꢀmodification protocol that has been successfully used in the past to the development of
1
4, 17a, 21
important inhibitors of various Znꢀmetalloproteases.
This protocol involves the
application of a 1,3ꢀdipolar cycloaddition reaction (1,3ꢀDCR) between readily accessible nitrile
oxides and an appropriate dipolarophilic precursor, such as tripeptide 9, bearing a propargyl
chain in P ′ position (Scheme 3). The synthesis of stereochemically pure precursors of similar
1
2
0
type has been described by our group in the past, therefore we decided to apply this chemistry
20
to the preparation of compound 9. Phosphinic acid 7 [[a]D = ꢀ41.3 (c = 1, EtOH)] was prepared
2
6
based on literature protocols and subjected to a PꢀMichael addition with ethyl 2ꢀ
1
7a, 20
methylenepentꢀ4ꢀynoate as the electrophile. In accordance to previous observations,
the
31
reaction proceeded with a significant degree of diastereoselectivity (d.r. ~2:1, based on
P
NMR). Saponification and coupling of the resulting diacid 8 with phenylalanine carboxamide 5c
afforded 9 which allowed the efficient isolation of the major isomer 9 (R,S,S) after 2
recrystallizations by ethyl acetate in satisfactory overall yield. This stereochemically pure
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
20
building block afforded inhibitors 13a-i after application of the oneꢀpot Huisgen protocol,
which involves the use of Nꢀchlorosuccinimide (NCS) as an oxidant for the in situ generation of
2
7
necessary nitrile oxides from the corresponding oximes, and final acidic deprotection.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 13a was reꢀsynthesized as a mixture of two epimers at P ′, 13a (R,S,S) and 13a′
1
(R,R,S), by using the corresponding diastereoisomeric mixture of 9. Inhibitor 13a obtained from
the stereocontrolled synthesis corresponds to the first RPꢀHPLC eluted isomer of the
diastereoisomeric mixture obtained from the nonꢀstereocontrolled synthesis. Based on previous
studies which correlate the stereochemical configuration of P ′ position of phosphinic peptides
1
1
7a, 19a, 20
with their elution order in RPꢀHPLC,
this is a strong indication of the R,S,S
stereochemical assignment for propargylic precursor 9. This was further supported after isolation
of the two isomers by semiꢀpreparative RPꢀHPLC and evaluation of the inhibitory potency of
individual isomers. Inhibition values showed that the firstꢀeluted RPꢀHPLCꢀeluted isomer (13a)
was 1ꢀ2 orders of magnitude more potent than its R,R,Sꢀepimer (13a′) towards all
aminopeptidases, a profile that is consistent with the higher potency observed for 6e (possessing
the stereochemical configuration of a natural substrate) as compared to its R,R,Sꢀepimer 6e′.
Finally, inhibitor 14 was obtained by cleavage of the Boc group from 9, using standard acidic
deprotection conditions (Scheme 4).
ACS Paragon Plus Environment

Page 11 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 3. Synthesis of P ′ Isoxazoleꢀdiversified Phosphinic Pseudotripeptide Inhibitors 13a-i
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) 33% HBr/AcOH, rt, 1 h; (b) Boc O, Et N, DMF, rt, 24 h, 89%,
2
3
o
o
two steps; (c) HMDS, H C=C(CH CH≡CH)COOEt, 110 C, 1 h, then 90 C, 3 h, then EtOH, 70
C, 30 min; (d) NaOH, MeOH, H O, rt, 24 h, then H O , 99%, two steps; (e) HBr·HꢀPheꢀNH
(5c), EDCꢁHCl, HOBt, DIPEA, CH Cl , rt, 2 h; (f) recrystallization by AcOEt, 41%, two steps;
(g) 10a-i (X for 10a: H, 10b: oꢀOMe, 10c: mꢀOMe, 10d: pꢀOMe, 10e: oꢀCl, 10f: mꢀCl, 10g: pꢀCl,
0h: oꢀOTBS, 10i: pꢀOTBS), HClꢁH NOH, CH COONa, MeOH, H O, rt, 24 h, 85ꢀ95%; (h) 11,
NCS, pyridine (cat.), CHCl , 45 C, 3ꢀ4 h, then 9, Et N, 45 C, 3 d, 1ꢀ6 repetitions; (i) Cs CO ,
DMF/H O, rt, 2ꢀ3 d; (j) TFA/CH Cl /TIS/H O 48:49:2:1, rt, 2h, two steps, for 13a: 80%, 13b:
5%, 13c: 75%, 13d: 76%, 13e: 73%, 13f: 67%, 13g: 66%, 13h (X = oꢀOH, three steps): 51%,
13i (X = pꢀOH, three steps): 61%.
2
2
o
+
2
3
2
2
2
1
2
3
2
o
o
3
3
2
3
2
2
2
2
8
a
Scheme 4. Synthesis of Phosphinic Inhibitor 14
a
Reagents and conditions: (a) TFA/CH Cl /TIS/H O 48:49:2:1, rt, 2h, 87%;
2
2
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Next, we proceeded to the synthesis of compounds 22a-f, according to the general synthetic
strategy outlined in Scheme 6. In particular, the synthesis of acrylic derivatives 15 and 19a,b was
initially performed, as it is described in Scheme 5 (adamantylꢀsubstituted acrylic ester 20 was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
28
prepared according to literature procedures ). Benzhydryl derivative 15 was prepared in good
yield according to the protocol of Hin et al. which involves DCCꢀmediated coupling of 2,2ꢀ
diphenylacetic acid with Meldrum’s acid, reduction of the resulting enol and final Manich
2
9
reaction using Eschenmosher’s salt and EtOH as a scavenger of the ketene intermediate. For
the synthesis of terphenyl derivatives 19a and 19b, aldehydes of type 17 were readily
synthesized via Suzuki coupling of 3,5ꢀdibromobenzaldehyde with the appropriate boronic
3
0
acids. The resulting aldehydes were converted to the corresponding bromides of type 18 after
reduction with NaBH and subsequent treatment with PBr . Substitution of the latter
4
3
intermediates by the anion of triethyl methanetricarboxylate, followed by
saponification/decarboxylation, Knoevenagel condensation of the crude substituted malonic
acids with paraformaldehyde and esterification of the resulting acrylic acids with EtOH
25, 31
successfully afforded the target acrylates 19.
PꢀMichael addition of phosphinic acid 7 to
acrylates 15, 19a,b and 20 afforded dipeptides of type 21 as mixtures of 2 diastereoisomers with
d.r. values ranging from 1:1 (21a) to 4:1 (21d). Coupling of 5b or 5c to diacids of type 21
followed by TFAꢀmediated deprotection afforded target compounds 22aꢀf. For 22a,b and 22d,
two isomers were obtained which were separated by semiꢀpreparative HPLC and evaluated
separately (annotated as 22a′, 22b′ and 22d′ in Tables 1 and 2) whereas for 22c,e and 22f single
isomers were obtained due to efficient separation during column purification of coupling
products.
ACS Paragon Plus Environment

Page 13 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 5. Synthesis of acrylic esters 15 and 19a,b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) DCC, DMAP, CH Cl , rt, 3 h; (b) NaBH , AcOH, CH Cl , rt, 24
2
2
4
2
2
h, 49%, two steps; (c) [(CH ) N=CH ]I, THF, EtOH, 65 ºC, 24 h, 93%; (d) 16a/16b Pd(PPh )
Na CO DME, H O, 24 h, for 17a: 98%, 17b: 82%; (e) NaBH , EtOH, rt, 1.5 h; (f) PBr ,
CH Cl , rt 1 h, two steps, for 18a: 91%, 18b: 81%; (g) HC(CO Et) , K CO , DMF/toluene,
reflux, 1.5 h; (h) KOH, EtOH, rt 24 h; (i) (HCHO) , Et NH, AcOEt, reflux, 4 h; (j) EtOH,
3
2
2
3 4,
2
3,
2
4
3
2
2
2
3
2
3
n
2
EDCꢁHCl, DMAP, DIPEA, CH Cl , rt, 24 h, four steps, for 19a: 56%, 19b: 25%.
2
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 6. Synthesis of P ′ Diversified Phosphinic Pseudotripeptide Inhibitors 22a-f
1
R
O
O
O
H
H
N
PH
OH
N
P
OH
a, b
Boc
Boc
OEt
+
OH
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R
7
21a-d
15, 19a,b, 20
R
O
H N
P
Xaa-NH2
2
c, d
OH
O
22a-f
Cl
R:
,
R1
,
,
R2
Cl
R3
R4
a
o
Reagents and conditions: (a) HMDS, acrylic ester 15 or 19a,b or 20 (R = 1ꢀAd), 110 C, 1 h,
o
o
+
then 90 C, 3 h, then EtOH, 70 C, 30 min; (b) NaOH, EtOH, H O, rt, 24 h, then H O , two steps,
2
3
for 21a (R=R ): 79%, 21b (R=R ): 61%, 21c (R=R ): 88%, 21d (R=R ): 87%; (c) HBr·Hꢀ
L)LeuꢀNH (5b) or HBr·Hꢀ(L)PheꢀNH (5c), EDCꢁHCl, HOBt, DIPEA, CH Cl , rt, 2ꢀ4 h; (b)
TFA/CH Cl /TIS/H O 48:49:2:1, rt, 2 h, two steps, for 22a (R=R , Xaa=LꢀPhe): 79%, 22b
R=R , Xaa=LꢀPhe): 80%, 22c (R=R , Xaa=LꢀLeu): 75%, 22d (R=R , Xaa=LꢀPhe): 78%, 22e
1
2
3
4
(
2 2 2 2
2
2
2
1
(
2 2 3
(
R=R , Xaa=LꢀLeu): 65%, 22f (R=R , Xaa=LꢀLeu): 40%.
3
4
Structure-activity relationships - P ′ position: Our first goal was to evaluate the effect of
2
various proteinogenic aminoacid sideꢀchains at P ′ position (with the only exception of 6k) on
2
the overall inhibition profile of APAs (Table 1). This sideꢀchain has been found to occupy the S2′
specificity pocket of ERAP2 in a previously determined crystal structure of ERAP2 with the
11
inhibitor 6e. In vitro evaluation demonstrated that having aromatic residues at that position (6c,
6
e,f) is a prerequisite for retaining sufficient potency for ERAP1, although a small hydrophobic
residue (6b) can be also well accommodated. Charged or hydrophilic residues (6g-j) greatly
ACS Paragon Plus Environment

Page 15 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reduced the affinity for ERAP1 and so did the unnatural stereochemistry of P ′ in 6k. In contrast,
2
virtually all tested sideꢀchains could be well tolerated by ERAP2, leading to nanomolar
inhibition for all compounds of type 6, even for 6k that possesses (R)ꢀstereochemistry at P2′.
Hydrophobic sideꢀchains were also preferred by IRAP, in contrast to polar and charged residues
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
that resulted in significantly reduced affinities. Overall, our analysis suggests that the P ′ position
2
is more important for controlling inhibitor potency for ERAP1, less so for IRAP and highly
permissive for ERAP2. This comes largely as a surprise since the crystal structure of ERAP2
with 6e showed that the indole moiety of Trp at P ′ position is sandwiched between two Tyr
2
residues that form the opposing sides of the S ′ pocket. One of these Tyr residues (Y438) is
2
32
conserved in both ERAP1 and IRAP and is actually important for catalysis. This apparent
paradox can be rationalized by taking into account the conformational changes that ERAP1 (and
6
b
probably ERAP2 and IRAP) undergo during catalytic cycle. ERAP1 has been crystallized in
two distinct conformational states, a “closed” and an “open” state, as defined by the interactions
between domains II and IV. The “open” state is considered to be less active and is characterized
by disorder at the S specificity pocket and a distinct conformation for the catalytic Y438. The
1
“closed” state is correspondingly characterized by the lack of access of the catalytic site to the
external solvent. Since any encounter complex between the inhibitor and ERAP1 would have to
be with the “open” state, interactions that may facilitate the transition to the “closed” state which
promotes the formation of the S pocket would favor inhibitor affinity. It is conceivable therefore
1
that aromatic sideꢀchains at P ′ may affect inhibitor affinity differently depending on the
2
conformation of the catalytic Tyr residue. Indeed, the orientation of the catalytic Tyr is different
amongst the three enzymes (Figure 2). The importance for aromatic residues at P ′ appears to
2
correlate well with the known orientation of this Tyr (most important for ERAP1, less so for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
IRAP and not important for ERAP2). These observations suggest that SAR for inhibitors that
occupy the S ′ pocket of APAs these enzymes may be complicated by the conformational states
2
available to each enzyme.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. In vitro evaluation of phosphinic pseudotripeptides 6aꢀk varying at position P2′.
IC (nM)
5
0
ID
Xaa
ERAP1
ERAP2
96±6
IRAP
102±9
32±2
41±6
14±1
57±35
29±9
6
a
b
c
LꢀAla 2518±390
6
6
6
LꢀLeu
LꢀPhe
LꢀPro
LꢀTrp
LꢀTyr
682±40
155±38
949±236
48±16
118±40
109±8
144±57
80±10
55±7
d
6e
6f
340±56
6g
LꢀSer 7473±1566 129±11 1754±631
LꢀLys 3093±968 271±55 1800±411
6
6
6
6
h
i
LꢀHis 1969±548 128±23
381±43
j
LꢀGlu 2660±174 243±17 4024±273
k
DꢀPhe 2650±212 211±21
185±5
ACS Paragon Plus Environment

Page 17 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: The three key residues that comprise the S ′ subsite of APAs. The Xꢀray crystal
2
structure of ERAP2 complex with 6e (green C, PDB ID: 4JBS) is superimposed with ERAP1
complexes with bestatin (PDB IDs: 2YD0, 3MDJ) in the closed and open states (dark and light
cyan C, respectively), and IRAP in its ligandꢀfree form (orange C, PDB ID: 5C97). Note the
different orientation of the catalytic Y438, the conserved D434 and the nonꢀconserved S869,
especially in the open and closed states of ERAP1.
Structure-activity relationships - P ′ position: Evaluation of the P ′ position was performed
1
1
along two separate routes involving (a) isoxazoleꢀbased, linearly extended sideꢀchains (13aꢀi) to
explore the depth of S ′ pocket and (b) “expanded” bulky sideꢀchains (22aꢀf) to explore the base
1
of the same pocket. Our results showed that the bulky sideꢀchains of inhibitors 22aꢀf were poorly
tolerated by ERAP1 whereas the extended isoxazole sideꢀchains of 13hꢀi were well tolerated
(Table 2). A similar but less striking phenomenon was evident for ERAP2, albeit some notable
exceptions, such as 22d. In contrast, IRAP was able to easily accommodate the P ′ sideꢀchains of
1
both groups of compounds (13hꢀi and 22aꢀf), resulting in most cases to low nanomolar
inhibition. This was more evident in the case of linear isoxazole sideꢀchains, considering that all
inhibitors of type 13 were able to potently inhibit IRAP with IC values ranging between 4 and
5
0
41 nM. The ability of IRAP to accommodate bulky sideꢀchains at P ′ position may be explained
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
by the different orientation that the conserved GAMEN motif adopts in this enzyme, an
orientation that allows for a more open catalytic site and has been hypothesized to allow
processing of cyclic peptides that may also require additional space around the base of S1′
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
c
pocket. Interestingly, inversion of the stereochemical configuration at P ′ position caused a
1
dramatic drop in the potency of all examples presented in Table 2 (13a′, 22a′, 22b′ and 22d′),
11
which is consistent with previously reported observations for the pair of P ′ꢀepimers 6e and 6e′
1
and emphasizes the importance of stereochemical control in the development of optimized
inhibitors of APAs, both in terms of potency and selectivity.
Table 2. In vitro evaluation of phosphinic pseudotripeptides 13aꢀi, 14 and 22aꢀf varying at
position P1′.
IC (nM)
5
0
ID
R
Xaa
ERAP1
126±31
ERAP2
IRAP
18±3
1
3a
LꢀPhe
LꢀPhe
190±24
13a′
2087±347
178±67
878±334
130±15
434±50
13±2
1
1
3b
3c
LꢀPhe
LꢀPhe
LꢀPhe
LꢀPhe
98±11
48±9
217±76
84±21
20±4
10±2
13±4
1
1
3d
3e
102±11
123±14
1
3f
LꢀPhe
143±14
225±67
41±5
ACS Paragon Plus Environment

Page 19 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
3g
LꢀPhe
LꢀPhe
71±8
345±138
83±11
34±9
9±2
13h
13i
65±38
LꢀPhe
LꢀPhe
LꢀPhe
33±5
43±4
56±9
37±4
4±1
2±1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4
22a
726±157
412±89
46±17
2
2
2a′
LꢀPhe 27008±7715 2271±486 561±66
2b
LꢀPhe
LꢀPhe
3694±908
>100000
740±129
32±9
2
2
2b′
1922±199 221±24
2c
LꢀLeu 31212±6939
246±7
16±2
21±3
22d
LꢀPhe
LꢀPhe
>35000
105±15
2
2
2d′
>100000
4622±2517 281±100
2e
LꢀLeu 21229±907
2565±47
317±20
22f
LꢀLeu 25111±6363 13814±566 577±21
Critical parameters for maximizing potency: As it is demonstrated in Tables 1 and 2, the
potency of phosphinic tripeptide inhibitors against ERAP1 can be greatly affected by both P1′
and P ′ sideꢀchain composition. Aromatic residues are preferred at P ′ position presumably due
2
2
to the importance of interaction with Y438 that facilitates the conformational change to the
closed form of the enzyme. Long sideꢀchains at P ′ position, such as the arylꢀsubstituted
1
isoxazole derivatives 13aꢀi, are well accommodated but do not confer any additional affinity
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compared to a small hydrophobic residue (such as Leu, 6b) at the same position. In contrast,
bulky substituents are highly disfavored at P ′ position.
1
The effectiveness of phosphinic tripeptides against ERAP2 was found to be highly permissive
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
with respect to position P ′ since all compounds exhibited activity in the nanomolar range (with
2
the exception of 22e, 22f and all primed diastereoisomers listed in Table 2). Similar to ERAP1,
bulky residues at P ′ position may not be optimal but are well tolerated and in some cases can
1
result to highly potent inhibitors (such as 22d).
P2′ position is also important for potency against IRAP: polar and charged residues lead to
reduction of the affinity, presumably reporting a hydrophobic environment for subsite S2′.
Inhibitor potency was also affected by the nature of the P ′ sideꢀchain, especially for bulky
1
substituents, but in general IRAP was able to accommodate bulky sideꢀchains while retaining
nanomolar potency.
The affinity displayed by inhibitor 14 that bears a propargyl group at P ′ position is also highly
1
notable since it appears to outperform leucine (6c) for all enzymes and especially IRAP,
resulting in an IC value of 2 nM. This affinity is in the same range as for several macrocyclic
5
0
angiotensin IV analogues developed by Hallberg and coworkers which are among the most
33
potent known IRAP inhibitors. Moreover, to the best of our knowledge, 14 is the most potent
inhibitor of ERAP2 described to date (IC 37 nM) and parallels 6e for potency against ERAP1
50
(
IC 43 nM for 14 and 48 nM for 6e). Although the origin of this outstanding behavior is not
50
clear, putative interactions between the electronꢀrich triple bond of the propargyl group and the
3
4
catalytic Zn(II) atom at certain orientations may provide a reasonable explanation. Given the
absence of any literature precedent concerning propargyl substituted phosphinic inhibitors of Znꢀ
ACS Paragon Plus Environment

Page 21 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
metallopeptidases, further investigations including crystallographic analysis will be necessary to
clarify this effect.
Critical parameters for maximizing selectivity: Our screening results revealed several
compounds to be selective inhibitors of ERAP2 and IRAP with respect to ERAP1 but afforded
only few compounds displaying selectivity for ERAP1, consistent with previous reports (Figure
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
0b, 11, 35
3).
This is likely a result of the conformational dynamics of ERAP1, which samples more
open states than ERAP2 and IRAP with partially disordered S and S ′ subsites, rendering
1
2
6
c, 10b
inhibitor targeting quite challenging.
Still, some selectivity was evident for some of the
isoxazole derivatives, indicating that the P ′ position can be valuable in enhancing ERAP1
1
selectivity over ERAP2. For example, compound 13g is about 5ꢀfold more selective for ERAP1
versus ERAP2. In an attempt to rationalize this observation, we examined the predicted
conformations of 13g from docking calculations, which showed that the P ′ sideꢀchain extends in
1
a linear fashion down the elongated shallow S ′ pocket of ERAP1 (Figure 4A). This pocket is
1
much more shallower in ERAP2 (Figure 4B) making this conformation inaccessible to extended
P1′ sideꢀchains and possibly explaining the observed selectivity. Another informative observation
that supports the aforementioned shape description of S ′ subsite for ERAP1 and ERAP2 is the
1
fact that among the isoxazole series, orthoꢀsubstituted aryl substituents (13b, 13e and 13h) result
to the least ERAP1ꢀselective compounds (in fact, 13b bearing the bulkier –OMe group displays
higher potency for ERAP2 than ERAP1). Evidently, in contrast to 13g, inhibitors 13b, 13e and
1
3h require a wider, more spacious S ′ channel for favorable accommodation of their P ′ sideꢀ
1
1
chains, which results in reduction of their affinity for ERAP1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3: Ratio of IC values showing the selectivity of phosphinic tripeptide inhibitors (A) for
5
0
ERAP2 over ERAP1, (B) IRAP over ERAP1 and (C) IRAP over ERAP2.
ACS Paragon Plus Environment

Page 23 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: (A) Docked conformation of 13g in the active site of ERAP1 (PDB ID: 2YD0), where
the P1' sideꢀchain extends inside the deep, elongated S ′ pocket. (B) The active site of ERAP2
1
(
PDB ID: 4JBS) shown from an identical view, where the shallower S ′ subsite renders such an
1
extended configuration of 13g inaccessible due to steric clashes with R366 and W363. Colors are
as in Figure 1.
Both P ′ and P ′ side chains of phosphinic tripeptides offer opportunities for the optimization
1
2
of ERAP2 selective inhibitors. Screening of P ′ position resulted in several compounds that were
2
selective for ERAP2 over ERAP1 (Figure 3A). In particular, polar and charged sideꢀchains were
well tolerated by ERAP2 but not ERAP1, thus providing selective inhibitors, such as 6g that
exhibits 50ꢀfold selectivity for ERAP2. In regards to the P ′ position, some of the bulky sideꢀ
1
chains that were tolerated by ERAP2 resulted in highly selective inhibitors. For example, 22d
was found to be over 350ꢀfold selective for ERAP2 over ERAP1. Docking calculations suggest
that 22d binds to ERAP2 with the P ′ sideꢀchain abutting to the extension of the GAMEN loop
1
(Figure 5). At this site, the presence of two nonꢀconserved residues in ERAP1 (Q315 and S316)
that could interfere with this configuration of 22d possibly accounts for the observed selectivity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5: Lowest energy conformation of 22d (green C) docked inside the active site of ERAP2
(orange C), superimposed with ERAP1 (cyan C) to illustrate the conserved GAMEN motif and
the two nonꢀconserved preceding residues that probably restrict the accommodation of similar
conformations of 22d in ERAP1 (PDB IDs: 2YD0 for ERAP1, 4JBS for ERAP2).
The highest selectivity during our screening analysis was observed for IRAP, with several
derivatives reaching selectivity of over three orders of magnitude (Figure 2). This was true for
screening of both the P ′ and P ′ positions in respect to ERAP1 but only for the P ′ position in
1
2
1
respect to ERAP2. Taken together, the P ′ position proved to be an exceptional modification site
1
for generating IRAPꢀselective inhibitors. This ability of IRAP to accommodate in its S ′ subsite
1
very bulky groups justifies the nanomolar affinity for IRAP observed by all listed compounds of
Table 2. In fact, this feature is more striking in IRAP than in ERAP2, considering that in the
most extreme case of bulky P ′ sideꢀchain (22f) examined in this work, nanomolar affinity for
1
IRAP was still attainable (IC50 577 nM) whereas IC50 value for ERAP2 was dramatically
increased to 13 ꢁM. This may be due to the alternative configuration of the GAMEN loop in
IRAP that generates additional space around that site, making it more permissive, especially for
ACS Paragon Plus Environment

Page 25 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
bulky sideꢀchains. In an effort to put these observations under a structural perspective we docked
IRAPꢀselective inhibitor 22b in the active site of IRAP. Two major lowꢀenergy configurations
were identified and shown in Figure 6. Interestingly, in both configurations the P ′ sideꢀchain
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
does not interact with the GAMENꢀmotif residues but rather forces either the P or P ′ sideꢀchain
1
2
to occupy the free space adjacent to the GAMEN loop. This space is not readily available in
ERAP1, making these configurations inaccessible (Figure 6).
Figure 6: Two docked conformations of 22b (green C) inside the active site of IRAP (orange C),
superimposed with the structure of closed ERAP1 (cyan C). The more "closed" configuration of
the GAMEN motif and of the two preceding residues in ERAP1 with respect to IRAP probably
renders such bound poses of 22b inaccessible (PDB IDs: 2YD0 for ERAP1, 5C97 for IRAP).
Biological evaluation: To access the efficacy and selectivity of this group of inhibitors in
regulating antigenic peptide processing in live cells, we evaluated the ability of one of the IRAPꢀ
selective inhibitors, 22b (DG026A), to regulate crossꢀpresentation by Bone Marrowꢀderived
Dendritic Cells (BMDCs) as described in the experimental methods section. DCs can perform
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
crossꢀpresentation by at least two distinct pathways one of which is dependent on ERAP1
4
a
activity and the other is dependent on IRAP activity. To confirm the selectivity of any effect we
ꢀ/ꢀ
+
+
compared BMDCs from both wildꢀtype and IRAP knockout mice. Mouse CD8α and CD11b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
DCs were exposed to soluble ovalbumin (OVA) in the presence of increasing concentrations of
+
the inhibitor. After 6 hours the cells were fixed to stop antigen processing and mixed with CD8
ꢀ/ꢀ
Tꢀcells isolated from the lymph nodes of Ragꢀ1 C57Bl/6 OTꢀI mice. Tꢀcell activation was
+
evaluated by measuring the levels of secreted ILꢀ2 by the activated CD8 Tꢀcells. The OTꢀI T
cell response to increasing levels of soluble OVA (Figure S1) was first evaluated to determine
the optimal concentration of OVA for all subsequent experiments (set to 500ꢂg/mL) (Figure S1).
+
+
For both CD8α and CD11b DCs, titration of the inhibitor 22b resulted in a doseꢀdependent
reduction of cross presentation down to a plateau that corresponds to IRAPꢀindependent crossꢀ
ꢀ/ꢀ
presentation (Figure 7A). No effect was evident when using DCs isolated from IRAP knockout
mice, indicating that the effect was IRAPꢀspecific (Figure 7B). A limited additional effect was
seen at the highest inhibitor concentration (3.3 ꢂM) which may be attributable to toxicity. In
conclusion, we demonstrate a doseꢀdependent and targetꢀspecific effect of the IRAP selective
inhibitor 22b on the crossꢀpresentation by both subsets of conventional DCs, which indicates that
this compound may be useful in regulating DCꢀtriggered adaptive immune responses.
ACS Paragon Plus Environment

Page 27 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
ꢀ/ꢀ
Figure 7: Response of CD8 Tꢀcells isolated from the lymph nodes of Ragꢀ1 C57Bl/6 OTꢀI
+
+
mice to FACSꢀsorted splenic CD8α and CD11b conventional DCs exposed to soluble OVA in
ꢀ/ꢀ
the presence of the inhibitor 22b. DCs were isolated either from wildꢀtype (Panel A) or IRAP
knockout mice (Panel B). A doseꢀdependent inhibition of the response by DCs from wildꢀtype
mice is evident (Panel A) down to the level of activation seen for DCs isolated from knockout
mice (Panel B).
CONCLUSIONS
In this study, we expanded on our previous observation that phosphinic pseudotripeptides are
potent inhibitors of aminopeptidases that are involved in the generation of antigenic peptides, by
thoroughly investigating the P ′ and P ′ positions for their effect on inhibitor potency and
1
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
selectivity. Special attention was given to the control of chirality during the preparation of target
compounds, considering the difference in affinity of 2ꢀ3 orders of magnitude previously
observed for the P ′ꢀepimers 6e and 6e′ and verified in this study for several more cases of
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
phosphinic tripeptides. This requirement is of paramount importance for the accurate evaluation
of reported structure–activity relationships and the discovery of improved inhibitors, especially
in terms of selectivity. In this regard, appropriately adjusted synthetic plans able to confer
stereochemical control were devised and employed. Our strategy succeeded in providing several
low nanomolar inhibitors for all APAs, such as 14 which is among the most potent inhibitors
ever reported for all three enzymes. Important inhibitors in terms of selectivity have also been
identified, such as the ERAP2 selective inhibitor 6g and the IRAP selective inhibitor 22b. Our
analysis suggests that the P ′ position is more important for controlling inhibitor potency for
2
ERAP1, than for ERAP2 and IRAP. Moreover, selectivity for ERAP2 can be controlled by polar
residues at P ′ position while selectivity for IRAP is mainly determined by the presence of bulky
2
groups at P ′ position. A notable observation concerning the effect of shape of P ′ sideꢀchains in
1
1
ERAP1/ERAP2 selectivity is that ERAP1 accommodates better than ERAP2 long, linear ligands
whereas this preference is inversed in the case of more “expanded” bulky groups (e.g. 22d).
Biological evaluation demonstrated that one of the selective IRAP inhibitors (22b) was able to
regulate crossꢀpresentation by dendritic cells exꢀvivo in a doseꢀdependent and targetꢀspecific
fashion. Our results suggest that this class of compounds can be useful for the targeted regulation
of adaptive immune responses and encourage their preclinical evaluation for applications in
cancer immunotherapy or the control of autoimmunity. Possible bioavailability concerns
associated with the high polarity of the described structures can be overcome either by prodrugꢀ
36
based strategies, as it has been successfully demonstrated in the past with fosinopril, a known
ACS Paragon Plus Environment

Page 29 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
marketed phosphinic inhibitor of angiotensinꢀconverting enzyme, and with a series of
structurally relevant phosphinic inhibitors of NEP and APN (aminopeptidase N) that exhibit
2
2a
analgesic properties , or with intravenous (i.v.) or intraperitoneal (i.p.) administration to
animals.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Materials and General Procedures: Solvents for reactions were purchased as anhydrous
grade and stored over 4Å activated molecular sieves before use. Reagents were purchased from
commercial sources and were used without further purification. Column chromatography was
performed on silica gel (E. Merck, 70ꢀ230 mesh). TLC analyses were performed on silica gel
plates (E. Merck, silica gel 60 F254), and components were visualized by the following methods:
UV light absorbance, ninhydrin spraying and/or charring after spraying with a solution
ofNH HSO4. Coupling reactions were monitored by LC/MS analysis. Melting points (measured
4
on an Electrothermal apparatus) are uncorrected. Optical rotation data were acquired on a Perkinꢀ
Elmer 343 polarimeter at 25 ºC. NMR spectra were recorded on a Varian 200 MHz Mercury or a
1
13
Bruker Avance DRXꢀ500 instrument at 25 ºC. H and C spectra are referenced according to the
37 31
residual peak of the solvent based on literature data.
P NMR chemical shifts are reported in
ppm downfield from 85% H PO (external standard) and the corresponding spectra are fully
3
4
proton decoupled. The following abbreviations are used to designate the multiplicities: s, singlet;
d, doublet; t, triplet; q, quartet; m, multiplet; and br, broad. ESI mass spectral analysis was
performed on a MSQ Surveyor, Finnigan mass spectrometer, using direct sample injection.
Negative or positive ion ESI spectra were acquired by adjusting the needle and cone voltages
accordingly. HRMS spectra were recorded on a Bruker Maxis Impact QTOF spectrometer or a
ACS Paragon Plus Environment