Design, Synthesis, and Characterization of Cyclic Peptidomimetics of the Inducible Nitric Oxide Synthase Binding Epitope That Disrupt the Protein-Protein Interaction Involving SPRY Domain-Containing Suppressor of Cytokine Signaling Box Protein (SPSB) 2 and Inducible Nitric Oxide Synthase

Article abstract of DOI:10.1021/acs.jmedchem.6b00386

SPRY domain-containing suppressor of cytokine signaling box protein (SPSB) 2-deficient macrophages have been found to exhibit prolonged expression of inducible nitric oxide synthase (iNOS) and enhanced killing of persistent pathogens, suggesting that inhibitors of the SPSB2-iNOS interaction have potential as novel anti-infectives. In this study, we describe the design, synthesis, and characterization of cyclic peptidomimetic inhibitors of the SPSB2-iNOS interaction constrained by organic linkers to improve stability and druggability. SPR, ITC, and ¹⁹F NMR analyses revealed that the most potent cyclic peptidomimetic bound to the iNOS binding site of SPSB2 with low nanomolar affinity (K_D 29 nM), a 10-fold improvement over that of the linear peptide DINNN (K_D 318 nM), and showed strong inhibition of SPSB2-iNOS interaction in macrophage cell lysates. This study exemplifies a novel approach to cyclize a Type II β-turn linear peptide and provides a foundation for future development of this group of inhibitors as new anti-infectives.

Full text of DOI:10.1021/acs.jmedchem.6b00386

Subscriber access provided by SIMON FRASER UNIV BURNABY
Article
Design, synthesis and characterization of cyclic peptidomimetics of
the inducible nitric oxide synthase (iNOS) binding epitope that disrupt
the protein-protein interaction involving SPRY domain-containing
suppressor of cytokine signaling box protein (SPSB) and iNOS
Jitendra R. Harjani, Beow Keat Yap, Eleanor WW Leung, Andrew J. Lucke, Sandra E. Nicholson,
Martin J Scanlon, David K Chalmers, Philip E. Thompson, Raymond S Norton, and Jonathan B. Baell
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00386 • Publication Date (Web): 23 May 2016
Downloaded from http://pubs.acs.org on June 5, 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 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
1
Design, synthesis and characterization of cyclic
peptidomimetics of the inducible nitric oxide
synthase (iNOS) binding epitope that disrupt the
proteinꢀprotein interaction involving SPRY domainꢀ
containing suppressor of cytokine signaling box
protein (SPSB) and iNOS
2
12
13
14
15
16
3
17
18
19
20
4
21
22
23
24
5
25
26
27
28
29
6
30
31
32
33
34
35
36
37
38
39
40
41
7
Jitendra R. Harjani,†§ Beow Keat Yap,†§ Eleanor W. W. Leung,† Andrew Lucke,† Sandra E.
Nicholson,‡ ¥ Martin J. Scanlon,† David K. Chalmers,† Philip E. Thompson,† Raymond S. Norton,
†* Jonathan B. Baell†*
8
9
42
43
44
10
† Medicinal Chemistry, Monash Institute of Pharmaceutical Sciences, Monash University,
Parkville, Victoria 3052, Australia
11
12
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
‡The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria 3052, Australia
13
14
¥The Department of Medical Biology, University of Melbourne, Parkville, Victoria 3052,
Australia
15
§These authors contributed equally
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 41
1
2
3
4
5
6
7
8
9
16
17
18
19
20
21
22
23
24
25
26
ABSTRACT: SPSB2 knockꢀout mice have been found to exhibit prolonged expression of iNOS
in macrophage cells and enhanced killing of persistent pathogens, suggesting that inhibitors of
SPSB2ꢀiNOS interaction have potential as novel antiꢀinfectives. In this study, we describe the
design, synthesis and characterization of cyclic peptidomimetic inhibitors of the SPSB2ꢀiNOS
interaction constrained by organic linkers to improve stability and druggability. SPR, ITC and
19F NMR analyses revealed that the most potent cyclic peptidomimetic bound to the iNOS
binding site of SPSB2 with low nanomolar affinity (KD 29 nM), a 10ꢀfold improvement over the
linear peptide DINNN (KD 318 nM), and showed strong inhibition of SPSB2ꢀiNOS interaction
in macrophage cell lysates. This study exemplifies a novel approach to cyclize a Type II βꢀturn
linear peptide and provides a foundation for future development of this group of inhibitors as
new antiꢀinfectives.
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
27
ACS Paragon Plus Environment
2

Page 3 of 41
Journal of Medicinal Chemistry
1
2
3
4
28
INTRODUCTION
5
6
7
8
29
30
31
32
33
34
35
36
The SPRYꢀdomain containing SOCS box protein 2 (SPSB2) modulates the lifespan of inducible
nitric oxide synthase (iNOS) in macrophages during infection as it mediates the proteasomal
degradation of this enzyme.1 It has been proven that the reduction of intracellular SPSB2 results
in prolonged iNOS expression and enhanced nitric oxide (NO) production, which ultimately
improves the killing of persistent pathogens. This suggests the potential of SPSB2ꢀiNOS
inhibitors as a new class of antiꢀinfectives. Xꢀray crystallographic studies showed that the SPRY
domain of SPSB2 binds to Asp184, Asn186 and Asn188 within the DINNN sequence of VASA,
which correspond to Asp23, Asn25 and Asn27 in the Nꢀterminal region of iNOS.1,2
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
37
38
39
40
41
42
43
44
45
46
47
The binding affinity, KD, of the 5ꢀresidue linear peptide sequence (DINNN), as determined by
SPR, was only 318 nM,3 about 25 times weaker than that of the 13ꢀresidue linear peptide
sequence (KEEKDINNNVKKT) of the Nꢀterminal region of iNOS (KD by isothermal titration
calorimetry (ITC) ≈ 13 nM)1 suggesting that the flanking residues of DINNN linear peptide
contributed to the low nanomolar binding between the Nꢀterminal region of iNOS and SPSB2.
Recently, we showed that the Nꢀacetylated cyclic peptide c[CVDINNNC]ꢀNH2 had a significant
improvement in binding to SPSB2 (KD by SPR: 4.4 nM ) when compared to its linear
counterpart.3 The disulphideꢀbridged cyclic peptide was also found to be stable in human plasma
and was stable against trypsin, αꢀchymotrypsin and pepsin, but the presence of a disulphide
bridge in this peptide rendered it susceptible to reduction in the cytoplasm, with a consequent
loss of the gain in affinity afforded by cyclization.
48
49
50
In this study, we describe a set of organic linkers that were designed in silico to optimally link
the Nꢀ and Cꢀ termini of the DINNN linear peptide. The organic linkers offer several advantages
over amino acid linkers, including tunability of size, shape and length, in addition to resistance to
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 41
1
2
3
4
5
6
7
8
9
51
52
53
54
55
56
57
58
59
60
proteolysis. Careful design of the organic linker imparts structural rigidity to the conformation
without interfering with the ability of the cyclic peptidomimetic to bind to the protein. Organic
linkers offer good redox stability when compared to the disulphideꢀbased cyclic peptide.3 We
describe the details of our design and synthesis of the building blocks that enabled us to
synthesize the cyclic peptidomimetics using Fmoc chemistry. The interaction of these cyclic
peptidomimetics with SPSB2 was characterized using SPR, ITC and 19F NMR experiments, and
their ability to displace fullꢀlength iNOS from its complex with SPSB2 in a macrophage cell
lysate was confirmed. A comparison of the binding properties presents an opportunity to
understand how the structurally different portions in these cyclic peptidomimetics contribute to
their binding affinities.
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
61
62
RESULTS AND DISCUSSION
Design of cyclic peptidomimetics M1, M2, M3 and M4
63
64
65
66
67
68
69
70
71
72
73
The design of the cyclic peptidomimetics was based on a technique called Interactive de novo
design,4 which we developed and which furnishes typeꢀIII mimetics,5 where the peptide
backbone is replaced by suitable organic scaffolds. In brief, the method is based on identifying
the vectors in a binding epitope that we wishes to match and replace with an organic scaffold
mimetic. While there are computational approaches to such problems, we have great success
adopting an interactive approach, where the medicinal chemist interactively builds scaffolds that
are simultaneously both synthetically accessible as well as conformationally appropriate. We
have applied this approach successfully to disparate, discontinuous binding epitopes where 3
vectors are matched in an organic scaffold. In the current problem, which relates more to design
of a molecular paperclip to furnish a cyclopeptidomimetics, we proposed to target just two
vectors. While relatively distant residues are readily spanned, it is convenient to focus on
ACS Paragon Plus Environment
4

Page 5 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
74
constrictions and in this context observation of the DINNNNN heptapeptide revealed that the
Asp184 amide NH was relatively close to the Asn188 C=O at only 2.685 Å (Figure 1A).
Further, the CαꢀN bond vector relationship with that of the Asn188 amide group was deemed
suitable for mimetic design with the intention of keeping the DINNN sequence for binding to
SPSB2.
75
76
77
78
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
79
80
81
82
83
84
85
86
87
88
The crystal structure of SPSB2ꢀVASA peptide at 1.8 Å resolution (PDB ID: 3EMW)2 was
loaded into SYBYL®ꢀX (version 2.0, Tripos) in order to undertake the mimetic design. All
unnecessary residues were stripped and the Asn188 carboxamide group was flipped in silico to
give an amide that moved closer to the Asp184 nitrogen atom. A variety of scaffolds were then
built interactively and joined with Asp184. A preferred type III peptidomimetic is identified as
M1 (Figure 1B). In building this scaffold and prior to minimization, it was expected that the
torsion angles 1, 2, 3 and 4 of around 180°, 45°, ꢀ45° and ꢀ80°, respectively, had acceptably low
strain energies. Further, the ethereal oxygen atom was deliberately chosen to facilitate synthesis
and interact favourably with the anilide NH. This minimization led to a barely perceptible
change in the designed conformation, with a final NꢀO distance of 2.01 Å.
89
90
91
92
93
94
95
96
In this approach, initial identification of a suitably synthetically versatile and
conformationally appropriate scaffold is always deceptively difficult, but becomes progressively
easier once an initial scaffold has been identified. Hence it is relatively easy to envisage that
replacement of the M1 phenyl ring with an Nꢀmethylated cisoid amide bond would also furnish
an appropriate scaffold. The resulting acylhydrazide peptidomimetic M2 (Figure 1B) uses the
same ethereal oxygen linkage, as part of the Asp184 incorporation. However, it contrasts with
M1 in that it contains polar groups in the mimetic region rather than the hydrophobic phenyl
ring. While it was not known if one might be preferred over the other, it is nevertheless a
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 41
1
2
3
4
5
6
7
8
9
97
98
convenient matching pair to have. In both cases, there is the opportunity of derivatizing the
mimetic scaffold without any conformational interference. For M1, this would be through the
addition of other groups on the phenyl ring, while for M2 it would be the extension of the Nꢀ
methyl group. In neither scaffold does an appropriate side chain conformation have to be
designed; rather, this is inherited from Asp184.
99
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
100
101
102
103
104
105
106
107
108
109
110
111
To further explore the effect of linker rigidity on the designed conformation and binding to
SPSB2, the methyleneꢀoxy group in M1 was substituted with an amide bond as a more rigid
alternative scaffold, as shown in M3. We hypothesized that the benzyl group in M1 may also be
easily replaced with a more flexible aliphatic ethylene linker as in M4. Hence M3 relates
indirectly to the principles of Interactive de novo Design, while M4 diverges completely from
these principles. Nonetheless, energy minimization of in silico models of both M3 and M4 did
not result in significant changes to the overall conformation of the mimetics compared to the
crystal structure of SPSB2ꢀDINNN, although M4 was connected by a linker with one covalentꢀ
bond shorter than M1, M2 and M3. On the basis of these considerations, peptidomimetics M1,
M2, M3 and M4 were selected for synthesis (Figure S1).
112
Synthesis of building blocks required for the synthesis of M1, M2, M3 and M4
113
114
115
116
117
118
119
The principle used in designing the synthesis of the cyclic peptidomimetics is illustrated with
an example of an ethyleneꢀbased peptidomimetic M4. In the cyclic peptidomimetic M4 (Figure
2A), only amino acid residues designated here as I2, N3, N4 and N5 (or the INNN motif) were
connected to each other via peptide backbone. Thus, disconnection at an amide CꢀN bond that
connects the INNN motif to the rest of the cyclic molecule, i.e. at the Cꢀterminus of N5, will lead
to a linear structure 1, which may be cyclized to M4 by a peptide coupling strategy involving the
free terminal –NH2 and –COOH groups. This strategy requires that all other groups (such as –
ACS Paragon Plus Environment
6

Page 7 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
120
CONH2 and –COOH) in 1 that might be potentially participating in the coupling reaction or
suffer from side reactions during the coupling conditions be suitably protected. A further
disconnection of the INNN motif from the rest of the linear chain at an amide CꢀN bond at the Nꢀ
terminus of I2 results in 2, which is an amino dicarboxylic acid that required protection of the
malate C4 carboxylic acid group for the desired regioselective coupling. Further protection at the
–NH2 group in 2 as –NHFmoc and at the C4 carboxylate of malate as a tertꢀbutyl ester results in
a target molecule 3, which is compatible with standard Fmoc peptide chemistry and therefore
should be a good precursor for the synthesis of appropriately protected 1. The advantage of this
approach would be that, except for 3, all other precursors and reagents required for this synthesis
of appropriately protected 1 are commercially available. In addition to this, well optimized
peptide coupling, deprotection, peptide cleavage and cyclization strategies commonly used in
solid phase Fmoc peptide chemistry can be used in the synthesis of M4.
121
122
123
124
125
126
127
128
129
130
131
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
132
133
134
135
136
137
138
139
140
141
142
Applying similar principles of retrosynthesis to cyclic peptidomimetics M1, M2 and M3, we
arrive at building blocks such as 4, 5 and 6 (Figure 2B). Structurally fragmenting 4, 5 and 6
further leads to simple chemical structures such as Nꢀmethylhydrazine, malate, anthranilamide,
aspartate and asparagine. The simplicity of these fragments also suggests that inexpensive,
commercially available precursors may be useful for the synthesis of the peptidomimetic
building blocks 4, 5 and 6, which should be able to provide access to cyclic peptidomimetics
M1, M2 and M3, respectively (Figure 2B). The synthesis of building blocks 4, 5 and 6 relied on
the synthesis of the malate diester precursor with C1 as –COOBn and C4 as ꢀCOOtBu
(numbering sequence identified in 2). The synthesis of this diester resulted in racemization at the
chiral Cꢀatom of the malate. We decided to carry through the racemic malate diester using it in
the synthesis of 4, 5 and 6 that were racemic at the malate knowing that the final step in the
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 41
1
2
3
4
5
6
7
8
9
143
144
145
146
purification of peptides generally involves purification of the peptide by separation using
HPLC. We were expecting that isoforms of M1, M2 and M4 might be separated from their
attempted synthesis. Detailed descriptions of synthetic procedures used for the synthesis of
building blocks 3, 4, 5 and 6 are provided in the supporting information file.
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
147
148
Synthesis and characterization of peptidomimetics M1, M2, M3 and M4 by LC-MS and
NMR
149
150
151
152
153
154
Utilizing the synthesized building blocks, linear peptidomimetics M1, M2, M3 and M4 were
assembled onꢀresin (solidꢀphase peptide synthesis) using an automated Peptide Synthesizer 3
(PS3TM) and Fmoc chemistry (Figure S1).6 Following cyclization in solution, all sideꢀchain
protected groups were removed with TFA in presence of scavengers. The cyclized
peptidomimetics were purified by reverseꢀphase high performance liquid chromatography (RPꢀ
HPLC) and their identity and purity were determined by LCꢀMS and NMR experiments.
155
156
157
158
159
160
161
162
163
164
165
The analysis of crude M1 by LCꢀMS indicated two closelyꢀeluted peaks (i.e. P1 and P2) that
had the same observed m/z and therefore the same deconvulated mass (Figure S2A). After
separation by RPꢀHPLC, P1 and P2 were characterized by 1D and 2D NMR. The chemical shift
and integration of signals in 1D NMR of both P1 and P2 indicated that they were likely to be
isomers of M1. Both P1 and P2 exhibited similar groups of spin systems in 2D [1H, 1H]ꢀTOCSY
experiments7 and the backbone amide resonances of both P1 and P2 were able to be sequentially
1
1
assigned using 2D [1H, H]ꢀROESY8 and [1H, H] DQFꢀCOSY9 experiments (Table S1). To
investigate the identity of these isomers, NOE peaks observed in the 2D ROESY NMR for P1
and P2 were compared with the distances between atoms in the designed in silico SPSB2ꢀbound
model derived from the crystal structure of SPSB2ꢀDINNN linear peptide (PDB ID: 3EMW).2 In
principle, atoms separated through space with a distance of less than 5 Å can be observed as
ACS Paragon Plus Environment
8

Page 9 of 41
Journal of Medicinal Chemistry
1
2
1
3
4
166
NOE peaks in 2D [1H, H]ꢀNOESY or ROESY experiments. The interꢀresidue NOEs between
assigned peaks in the 2D [1H, 1H]ꢀROESY spectra of both P1 and P2 in water at pH 4.8 acquired
at a 300 ms mixing time at 10 °C were compared to each other. A strong NOE crossꢀpeak
between resonances at 4.22 and 4.50 ppm (Hα and Hlb) and a weaker NOE crossꢀpeak between
resonances at 4.22 and 4.61 ppm (Hα and Hla) was apparent in the 2D [1H, 1H]ꢀROESY spectra of
P1 (Figure S3A). However, similar NOE crossꢀpeaks between Hα and Hl* were not observed in
the 2D [1H, 1H]ꢀROESY spectra of P2 (Figure S3B). When the distances between the Hα and Hl*
in SPSB2ꢀM1, i.e. the bound model of mimetic M1, were measured (Figure S3C), we observed
that Hα and Hl* were spatially separated by 2.4 ꢀ 3.5 Å. This model suggested that one strong and
one weak interꢀresidue NOE cross peaks are expected in the 2D [1H, 1H]ꢀROESY spectra of M1,
indicating that P1 is the originally designed mimetic represented as M1 (Figure 1). As the
racemic malate was used for the synthesis of M1, we assumed that perhaps P2 may be the epimer
of the designed M1, although future work is warranted to confirm its identity.
5
6
7
8
9
10
11
12
13
14
15
16
17
167
168
169
170
171
172
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
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
A similar observation was made for the crude product obtained when the synthesis of M2 was
attempted. Two peaks were observed in this case and designated P3 and P4. When the purified
products corresponding to peaks P3 and P4 of mimetic M2 were analyzed by NMR, we observed
1
that the number of proton resonances in the backbone amide region of H NMR spectra of both
P3 and P4 was more than the expected number of amide resonances of the designed mimetic,
which could be due to conformational isomerism of the peptidomimetic or impurities from either
a deletion peptide or organic impurities that coꢀelute with P3 and/or P4 during the purification. In
order to investigate the possible origin of the additional amide resonances, we conducted a
variable temperature NMR study on both P3 and P4. An increase in the temperature from 10 to
1
30 °C resulted in broadening of 10 out of 12 resonances in the backbone amide region of the H
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 41
1
2
3
4
5
6
7
8
9
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
NMR spectra of P3, with sharpening of 2 of the existing resonances observed after 25 °C
(Figure S4A, left), suggesting that some amide protons may be involved in conformational
exchange while others were rapidly exchanged with the solvent at high temperature. 2D [1H,1H]ꢀ
TOCSY experiments on P3, however, revealed three spin systems with an A3MPT(B3)Xꢀlike
pattern, a common pattern of the spin system for an Ile. In addition, 9 backbone amide proton
peaks were observed, which is more than twice the expected number of amide resonances for M2
(Figure S4B, left), suggesting that P3 may not be the designed peptidomimetic M2. In contrast,
we were able to sequentially assign the resonances of P4 via 2D [1H,1H]ꢀROESY and [1H,1H]ꢀ
DQFꢀCOSY NMR experiments (Table S2), confirming it to be the designed mimetic M2. In
addition, the variableꢀtemperature NMR experiments on peak P4 indicated that the number of
broadened peaks in the amide region with the increase in temperature matches the expected
number of the amide resonances for M2 (Figure S4A, right). Notably, three of the resonances
(as shown in the 2D [1H, 1H]ꢀTOCSY experiment) belong to spin systems with an AMX pattern
(i.e. Asn3, Asn4 and Asn5), while another resonance belongs to the spin system with the
A3MPT(B3)X pattern (i.e. Ile) (Figure S4B, right), consistent with that of the designed mimetic
M2. The presence of other temperatureꢀresistant proton resonances in the NMR spectra of P4
(including possible aromatic peaks between 7ꢀ8 ppm), suggests that the purified P4 peak also
contains some coꢀeluting organic impurities from purification.
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
207
208
209
210
211
Syntheses of mimetics M3 and M4 yielded single products. This was indicated by a single
peak in the UV trace with a matching m/z in their LCꢀMS profiles (Figures S2C and S2D). In
addition, all backbone amide proton resonances observed for the purified mimetic M3 and M4
(Figures S5C and S5D) could be sequentially assigned via [1H, 1H]ꢀROESY and [1H, 1H] DQFꢀ
1
COSY with the number and characteristics of the spin systems observed by [1H, H]ꢀTOCSY
ACS Paragon Plus Environment
10

Page 11 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
212
matching the predicted spin systems for each mimetic (Tables S3 and S4). These observations
thus confirmed the identity of the products as the designed mimetic M3 and M4.
213
214
SPR analysis of M1, M2, M3 and M4 binding to SPSB2
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
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
The binding of M1, M2, M3 and M4 to SPSB2 was analyzed by SPR (Figure 3A). The KD
values for M1, M2, M3 and M4 ranged from 29 to 465 nM (Table 1), with M1 bound more
tightly to SPSB2 than M2 (p<0.05) and M4 (p<0.01), and comparable to M3 (p>0.05).
Consistent with the observations made for Nꢀacetylated cyclic peptide c[CVDINNNC]ꢀNH2, M1,
M2 and M3 demonstrated a faster association rate (1–2 ×106 Mꢀ1sꢀ1) than the linear peptide,
presumably by offering more rigid binding conformations.3 These observations suggest that
cyclization of DINNN via the organic linkers improved their binding affinity for SPSB2. The
offꢀrates were found to be modestly different from each other, with the slowest offꢀrate being
that of M1, followed by M3 and M2. M4, however, did not show improvement in affinity for
SPSB2 when compared with the linear peptide DINNN with a KD of approximately 465 nM, and
a koff of 0.3 sꢀ1. This indicates that the structural characteristics of the linker are crucial in
determining how these cyclic peptidomimetics interact with SPSB2. The peptidomimetics
corresponding to peaks P2 and P3 of M1 and M2, respectively, were also found to bind to
SPSB2, although their binding affinity as determined by SPR, was approximately two to three
times lower than that of P1 and P4 from the respective syntheses (Figure S6).
230
ITC analysis of M1, M2, M3 and M4 binding to SPSB2
231
232
233
To further investigate the thermodynamics of these interactions, ITC titration curves were
obtained for the binding of M1, M2, M3 and M4 to SPSB2 (Figure 3C). In all four cases, the
binding of peptidomimetic to SPSB2 was an exothermic event, with negative peaks (ꢁH <0)
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 41
1
2
3
4
5
6
7
8
9
234
235
236
237
238
239
following each injection. Table 1 summarizes the thermodynamic parameters obtained during
this study. The binding stoichiometry was close to 1:1 in all cases. Both M1 and M3 bound to
SPSB2 with stronger affinities, with KD values of 45 and 49 nM, respectively. M2 had a KD of 83
nM and the weakest binder M4, with a KD of ~442 nM. All KD values measured by ITC match
with SPR except for mimetic M1 where an approximately 1.5ꢀfold lower KD value was observed
by SPR.
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
240
241
242
243
244
245
246
247
Although both M1 and M3 showed similar KD values by ITC, the enthalpy change associated
with the binding of M1 to SPSB2 (ꢀ26 kcal/mol at 25 ºC) was about 4 kcal molꢀ1 higher in
magnitude than for M3. The observed enthalpy change in M1 upon binding to SPSB2 was also
the highest in magnitude of the four cyclic peptidomimetics investigated. In general, the enthalpy
changes associated with the binding of these peptidomimetics were significantly larger in
magnitude than those observed for small molecules (typically about ꢀ15 kcal molꢀ1).10 One
plausible explanation for this might be an extensive hydrogen bonding network predicted to form
between these peptidomimetics and SPSB2.
248
19F NMR analysis of the interaction of SPSB2 with M1, M2, M3 and M4
19
249
250
251
252
253
254
255
256
Owing to the high sensitivity of F chemical shifts to changes in their local environment, we
utilized 19F NMR as a versatile tool to probe ligand binding to SPSB2. Our previous study
successfully established a fluorineꢀlabelling scheme for SPSB2 whereby all six Trp residues in
SPSB2 were replaced with 5ꢀfluorotryptophan (5ꢀFꢀTrp) using a Trp auxotroph Escherichia coli
strain. These six Trp residues were wellꢀresolved in the 19F NMR spectrum and were assigned by
utilizing siteꢀdirected mutagenesis. Four of the Trp residues were unambiguously assigned to Trp
48, Trp131, Trp133 and Trp207, while the remaining two Trp residues were unassigned, because
mutating these residues negatively impacted on the structure of the protein. We have shown
ACS Paragon Plus Environment
12

Page 13 of 41
Journal of Medicinal Chemistry
1
2
3
4
257
previously that ligands targeting the iNOSꢀbinding site of SPSB2 induce significant chemical
5
6
7
8
19
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
shift perturbations of the 5ꢀFꢀTrp indole resonances of SPSB2, particularly the F resonance of
Trp207.11 Peptides containing the DINNN sequence typically cause a downfield shift of about 3
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
19
ppm of the F resonance of Trp207. As shown in Figure 4, the binding of M1 and M2 to 5ꢀFꢀ
TrpꢀSPSB2 caused a similar downfield shift of the Trp207 resonance to that caused by
c[CVDINNNC]ꢀNH2, although the chemical shift changes were slightly larger and the peak
widths were broader for M1 and M2 (Figure 4). Although the spectral perturbations for M1 and
M2 were relatively similar, a sharper Trp207 resonance was observed with M1 than with M2.
Interestingly, resonance at ꢀ44.5 ppm was also perturbed and was only influenced by M1, M2
and the control peptide, Nꢀacetylated cyclic c[CVDINNNC]ꢀNH2. Although this resonance (peak
3) was unassigned, it is likely that this resonance is originated from Trp95 and that its adjacent
shoulder peaks are minor conformers. According to the crystal structure, Trp95 is approximately
19
6Å distance away from Trp207 and hence, is likely to have a greater influence on the F NMR
spectrum in presence of peptidomimetic/peptide as compared to Trp91 which is about 18Å from
Trp207. In the presence of M3, the Trp207 resonance broadened beyond detection, suggesting
that the complex with M3 samples multiple conformations. In contrast, mimetic M4 (with
expected bound occupancy of 99.8% at KD ≈ 442 nM) did not perturb the chemical shift of
Trp207 resonance, but rather causes an enhancement of the resonance, suggesting that M4 may
bind to SPSB2 in such a way that Trp207 is not perturbed in the same way as with the control
cyclic peptide and other mimetics. In most cases, Trp207 was significantly perturbed by the
presence of peptide/mimetic. In some instances, the presence of mimetic (e.g. for M1, M2 and
M3) caused the appearance of some minor peaks. The resonance of Trp207 in the presence of
M4 was about 25% higher in peak height when compared to its unbound counterpart.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 41
1
2
3
4
280
Simulation of M1 bound to SPSB2
5
6
7
8
9
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
To further understand the observed binding dynamics of the most potent peptidomimetic M1
by 19F NMR on SPSB2, a 1 ꢀs MD simulation was performed of the modelled SPSB2ꢀbound M1
on the crystal structure of SPSB2 protein (PDB ID: 3EMW).2 Throughout the simulation, the
protein backbone rmsd was 2.7 Å relative to the initial structure. In addition, mimetic M1
remains closely bound to the iNOS binding site on SPSB2, with the side chain of M1ꢀN4 located
above the side chain of Trp207 of SPSB2, with an average distance between side chain nitrogens
of 5.1 Å. Further analysis of hydrogen bonding interactions revealed seven persistent hydrogen
bonds with occupancy of > 25% between SPSB2 and M1 (Table S5). The hydrogen bonding
present during simulation was consistent with that inferred from the binding of VASA peptide to
SPSB2,2 with key hydrogen bonds observed between the equivalent amide side chain (NH) of
residue N5 of mimetic M1 (M1ꢀN5) and the backbone carbonyl group of Val206 of SPSB2, as
well as between the carbonyl side chain of M1ꢀN5 and NH side chain of Arg68 of SPSB2
(Figure 5). Furthermore, residue M1ꢀN3 made a side chain donor interaction with the side chain
of Thr102, side chain acceptor interaction with the backbone of Thr102 and backbone acceptor
interaction with the main chain of Gly208 of SPSB2.
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
296
297
298
299
300
301
302
The side chain of M1ꢀBzD1 was also found to accept hydrogen bonds from the Tyr120 side
chain and the backbone NH of Gln73. Nevertheless, the hydrogen bond between the side chain
of M1ꢀBzD1 and Tyr120 of SPSB2 was not maintained throughout the simulation as the
hydrogen bond breaks and forms during the simulation coincident with the side chain of Tyr120
adopting a transient second orientation away from the side chain of M1ꢀBzD1. Furthermore, the
distance between the side chain nitrogens of Trp207 and M1ꢀN4 was found to fluctuate between
3.1 and 10.3 Å throughout the simulations. These observations suggest that although mimetic
ACS Paragon Plus Environment
14

Page 15 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
303
M1 is bound to the SPSB2 protein, it did not lock the iNOS binding site of SPSB2 into a single
bound conformation, which is consistent with the observation of a broader 19F resonance for
Trp207 in the presence of M1 compared with the disulphideꢀbridged cyclic peptide.
304
305
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
306
307
Displacement of full-length iNOS from SPSB2 in macrophage cell lysates by M1, M2, M3
and M4
308
309
310
311
312
313
314
315
316
317
The ability of these cyclic peptidomimetics to inhibit the interaction between fullꢀlength iNOS
and SPSB2 in physiological conditions was investigated. M1, M2, M3 and M4 were incubated
in lysates from BMDM stimulated by LPS and IFNꢀγ. Recombinantly expressed GSTꢀtagged
SPSB2 protein was used to pull down iNOS from the macrophage cell lysate. As shown in the
Western blots with antiꢀiNOS antibodies in Figure 6, affinity purification with GSTꢀSPSB2 in
the presence of peptidomimetics resulted in a significant reduction in the amount of iNOS pulled
down, suggesting that all peptidomimetics were competing with fullꢀlength iNOS for binding to
the SPSB2 protein in the macrophage cell lysate. Mimetic M1 was found to give the strongest
reduction, followed by mimetic M3, M2 and M4, consistent with the observations by both SPR
and ITC.
318
319
320
321
322
323
324
325
To elucidate the effect of structural differences of the organic scaffold used for cyclization of
peptidomimetics on their ability to bind to SPSB2, the results of SPR, ITC and 19F NMR
experiments were compared. The SPR and 1TC assays showed low nanomolar binding affinities
for all cyclic peptidomimetics on SPSB2 except M4, with M1 showing the strongest affinity to
SPSB2 (KD ≈ 29–45 nM), followed by M3 (KD ≈ 49–54 nM), M2 (KD ≈ 83–99 nM) and M4 (KD
≈ 442–465 nM). Mimetic M4, which has a covalent bond shorter than other mimetics, did not
show any improvement over the linear peptide DINNN, presumably due to its subꢀoptimal
conformation for binding to SPSB2, as supported by a 4ꢀfold faster koff than the linear peptide
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 41
1
2
DINNN (koff ≈ 0.07 sꢀ1).3 In contrast, M2 and M3 showed similar offꢀrates to the linear peptide,
while M1 gave a slowest offꢀrate, comparable to the observed koff for the disulphideꢀbridged
cyclic peptide (koff = 0.02 sꢀ1).3
3
4
5
6
7
8
9
326
327
328
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
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
M1, M2 and M3 perturbed the 5ꢀFꢀTrp207 resonance of SPSB2, indicating that these
mimetics bound to the iNOS binding site of SPSB2 in a similar pose to those of the cyclic
peptide c[CVDINNNC]ꢀNH2, although significant differences in peak line width were evident. A
sharper 5ꢀFꢀTrp207 resonance was observed in the presence of M1; molecular dynamic
simulations of this complex revealed that M1 maintained its bound conformation on SPSB2
throughout the 1 µs simulations, with seven persistent hydrogen bonds observed between M1
and key residues in the iNOS binding site of SPSB2, consistent with its slow offꢀrate and strong
19
binding enthalpies according to SPR and ITC. Intriguingly, the resonance of 5ꢀFꢀTrp207 in F
NMR was not perturbed in the presence of mimetic M4. A slight sharpening of the peak was
observed, suggesting that mimetic M4 may have adopted a different binding conformation on
SPSB2, locking the iNOS binding site of SPSB2 in the unbound state. When the mimetic M4
was incubated with macrophage cell lysates, a significant reduction in the amount of iNOS
pulled down by GSTꢀtagged SPSB2 from the cell lysates was observed confirming that M4 was
still bound to the iNOS binding site of SPSB2 although the reduction was smaller than those
caused by M1 and M3, consistent with their higher binding affinities by SPR and ITC.
344
345
346
347
348
While cyclic peptides of this size tend to have much better bioavailability than larger
peptides,12 cellular permeability can still be problematic, especially for molecules with low
overall lipophilicity.13 However, since macrophages are an important target for our mimetics,
and since these cells can ingest particles of diameter 0.5 ꢀ10 µm,14 the next step in progression of
these molecules will be to couple them to polymers that favour uptake by macrophages.
ACS Paragon Plus Environment
16

Page 17 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
349
Preliminary results on the peptidomimetic analogue M1 show some uptake by murine BMDMs
even in the absence of polymeric carriers, but there is considerable scope for enhancement of this
process.
350
351
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
352
353
354
355
356
357
358
359
360
361
Given a positive outcome of design and synthesis in this study, it is worth mentioning that as
we have described in detail elsewhere,4dꢀe this approach of designing peptidomimetics is quite
different to those involved in design of scaffolds to replace secondary structures such as βꢀturns,
successful reports of which first emerged more than 35 years ago.15 Our method is applicable to
any binding epitope, weather continuous or discontinuous and as there is a wealth of structural
knowledge around biologically active peptides in current databases, we believe that this
approach has broad applicability. In our hands, it is routinely successful yet appears to be seldom
reported by others, if at all. With respect to its application reported herein, our results provide a
foundation for future development of cyclic peptidomimetics targeting SPSB2 protein as
potential novel class of antiꢀinfectives.
362
363
364
CONCLUSION
In order to design new inhibitors of the SPSB2ꢀiNOS interaction, we performed an interactive de
novo design using the crystal structure of the linear peptide DINNN bound to SPSB2. This led to
designed cyclic peptidomimetics M1 and M2 and partners M3 and M4, which incorporate
organic scaffold between the Nꢀ and Cꢀtermini of the DINNN sequence. Four building blocks
were synthesized for incorporation into solid phase peptide synthesis by Fmoc chemistry. These
building blocks were used for the synthesis of M1-4. Synthesis and further characterization of
the final purified products by LCꢀMS and 2D NMR confirmed the identity of the designed
mimetics M1-4. The most potent cyclic peptidomimetic inhibitor identified in this study, M1,
bound strongly to SPSB2 (KD 29 nM) and inhibited fullꢀlength iNOS from interacting with
365
366
367
368
369
370
371
372
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 41
1
2
3
4
5
6
7
8
9
373
374
375
376
377
378
SPSB2 in macrophage cell lysate. Although M1 is less potent than a previously identified
disulphideꢀbridged cyclic peptide, it is smaller in size and is redoxꢀstable, suggesting that it has a
greater potential to be developed into a new class of antiꢀinfective drugs. As the antibiotic
resistance continues to be a significant public health problem, newer antiꢀinfectives that have a
different mode of action, such as the ones discussed in this work, would be of interest to broad
research community.
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
379
380
381
382
EXPERIMENTAL PROCEDURES
Synthesis and purification of M1, M2, M3 and M4
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
All reactions were carried out at room temperature. The linear peptidomimetics required for the
synthesis of M1, M2, M3 and M4 were synthesized onꢀresin (solidꢀphase peptide synthesis)
using an automated Peptide Synthesizer 3 (PS3TM) and Fmoc chemistry.6 0.1 mmol (1 equiv.)
peptidomimetic building block 3, 4, 5 or 6 (synthesized according to methods outlined in
supporting information) was loaded on 0.2 mmol of 2ꢀchlorotrityl chloride resin (1.2 mequiv./g
substitution) followed by chain assembly via the coupling of 0.3 mmol (3 equiv.) of Fmocꢀ
protected amino acids to the resin bound 3, 4, 5 or 6 (Figure S1). The coupling reactions
between the Cꢀterminal residue and the resin, and those involving subsequent amino acid
residues were carried out using a mixture of 0.3 mmol (3 equiv.) HCTU and 0.6 mmol (6 equiv.)
DIPEA over 50 min. Double couplings were carried out to couple the second Asn to the first Asn
as well as the third Asn to the second Asn, as single coupling was found to result in substantial
deletion(s) of Asn from the peptide sequence. The Fmoc protecting group was removed with
20% (v/v) piperidine in DMF after each coupling/double coupling cycle was completed. In each
case, the sideꢀchainꢀprotected linear peptidomimetic was cleaved from the resin using 30%
hexafluoroisopropanol (HFIP) in DCM. The Nꢀ and Cꢀ termini were coupled (cyclization) in
ACS Paragon Plus Environment
18

Page 19 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
398
solution at a peptidomimetic concentration of 0.5 mM over 24 h using 3 equiv. HCTU and 6
equiv. DIPEA in DMF. The protecting groups (Tr for amide and tertꢀbutyl ester for carboxylic
acid) were removed with a solution containing 95% (v/v) TFA, 2.5% (v/v) triisopropylsilane
(TIPS) and 2.5% (v/v) 3,6ꢀdioxaꢀ1,8ꢀoctanedithiol (DODT), the latter two being the scavengers.
After cleavage, the peptidomimetics were precipitated with cold diethyl ether twice and
dissolved in a 1:1 mixture of water (H2O) and acetonitrile (MeCN) prior to lyophilization. The
crude peptidomimetics M2, M3 and M4 were purified by RPꢀHPLC) on a Phenomenex® Luna
C18 (2) column (100 Å, 5 ꢂm, 100 × 10 mm) using a gradient of 5–40% B (A: 99.9% H2O, 0.1%
TFA; B: 80% MeCN, 19.9% H2O, 0.1% TFA) over 35–70 min. For the crude product obtained
during the synthesis of peptidomimetic M1, peaks P1 and P2 in the UV trace of LCꢀMS shown
in Figure S2A were separated using the same column, but at a gradient of 20–60% B over 40
min (A: 10 mM ammonium acetate pH 3.8; B: 50% 10 mM ammonium acetate pH 3.8, 50%
MeCN). Under these conditions, peak P2 was found to elute faster than P1.
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
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
Purity of peptide building blocks and mimetics M1, M2, M3 and M4
To establish the purity of 3, 4, 5 and 6, low resolution mass spectrometry analyses were
performed with an Agilent 6100 Series Single Quad LC/MS coupled with an Agilent 1200 Series
HPLC, 1200 Series G1311A quaternary pump, 1200 series G1329A thermostated autosampler
and 1200 series G1314B variable wavelength detector. The conditions for liquid chromatography
were: reverse phase HPLC analysis fitted with a Phenomenex Luna C8(2) 5 µm (50 x 4.6 mm)
100 Å column; column temperature: 30 °C; injection volume: 5 µL; solvent: 99.9% acetonitrile,
0.1% formic acid; gradient: 5–100% of solvent over 10 min; detection: 254 nm. The conditions
for mass spectrometry were: quadrupole ion source; ion mode: multimodeꢀES; drying gas temp:
300 °C; vaporizer temperature: 200 °C; capillary voltage: 2000 V (positive), 4000 V (negative);
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 41
1
2
3
4
5
6
7
8
9
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
scan range: 100ꢀ1000 m/z; step size: 0.1 sec; acquisition time: 10 min. In case of 23, 5 min
acquisition time was used. The purity of 3, 4, 5 and 6 was ˃ 95%.
For M1, M2, M3 and M4, the purity and molecular mass of the synthesized cyclic
peptidomimetic were confirmed by LCꢀMS on a Shimadzu LCMS2020 instrument, incorporating
a Phenomenex® Luna C8 column (100 Å, 3 ꢂm, 100 × 2 mm), using a linear gradient of 100%
H2O (0.05% TFA) for 4 min, followed by 0–60% MeCN (0.05% TFA) in water over 10 min at a
flow rate of 0.2 ml/min. The purity of M1, M2, M3 and M4 was ˃ 95%.
SPSB2 protein expression and purification
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 GSTꢀtagged SPSB2 (GSTꢀSPSB2) expression construct has been described previously.16
The SPSB2 and 5ꢀFꢀTrp labeled SPSB2 (5ꢀFꢀTrp SPSB2) were both prepared as described
previously.11
NMR spectroscopy
All NMR experiments were recorded on a Bruker Avance 600 MHz spectrometer, equipped with
a 5 mm TCI cryoprobe. For chemical shift assignments of M1, M2, M3 and M4, the cyclic
2
peptidomimetic was dissolved in water at pH 4.7–4.8 with 10% H2O. Homonuclear 2D NMR
spectra were recorded at a temperature that allowed for the optimal resolution of the backbone
1
amide peaks, which was at 10 °C for M1, M2 and M3 and 25 °C for M4. H chemical shifts
13
15
were referenced to DSS, while C and N chemical shifts were referenced indirectly through
their gyromagnetic ratios.17 The 2D TOCSY spectrum (2048 × 300) was acquired using the
DIPSIꢀ2 pulse sequence7a with excitation sculpting for water suppression,18 a relaxation delay of
1.3 s and a spinꢀlock time of 70 ms. ROESY spectra (2048 × 300)8 with WATERGATE for
water suppression19 were acquired with a 300 ms spinꢀlock time and a relaxation delay of 3 s.
2
DQFꢀCOSY spectra (2048 × 300)9 were acquired in 99.9% H2O, with a relaxation delay of 1.3
ACS Paragon Plus Environment
20

Page 21 of 41
Journal of Medicinal Chemistry
1
2
13
s. For carbon and nitrogen chemical shift assignments, [1H, C]ꢀHSQC (2048 × 256) and [1H,
3
4
444
5
6
7
8
9
10
11
12
445
15N]ꢀSOFAST HMQC (2048 × 64) spectra20 were acquired at pH 4.8 using 99.9% 2H2O and 90%
1H2O: 10% 2H2O, respectively. Spectra were processed with a sineꢀbell squared window function
using Bruker TopSpin (Version 3.2) and analyzed using CcpNmrꢀAnalysis (Version 2.1.5). The
chemical shift assignments of M1, M2, M3 and M4 are summarized as Table S1, S2, S3 and S4,
respectively.
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
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
Surface plasmon resonance
SPR experiments were performed at 25 °C on a Biacore™ T200. SPSB2 immobilization was
carried out via the standard amine coupling method on a CM5 sensor chip using a running buffer
of 10 mM HEPES, 150 mM NaCl, 3 mM EDTA and 0.05% Tween 20 at pH 7.4, as described
previously.3 The binding of M1, M2, M3 and M4 was investigated using the same running
buffer. Each peptidomimetic was injected onto the surface with a contact time of 60ꢀ100 s, flow
rate of 100 µL/ min, and dissociation time of 200ꢀ300 s. The sensorgrams were doubleꢀcorrected
for nonꢀspecific binding to the surface by subtracting (1) the signals on the reference surface
from that of the target protein surface, and (2) the signals of the blank sample (no mimetics) from
that of the mimetic containing samples. The corrected signal was fitted to a steady state 1:1
interaction model with Biacore T200 Evaluation Software (version 2.0) and the binding affinity,
KD, association rate, ka, and dissociation rate, koff, were determined. Differences between the KD
of each analogue were evaluated using oneꢀway ANOVA and multiple Tukey’s comparison tests
with GraphPad Prism (version 6.05).
Isothermal titration calorimetry
ITC measurements were acquired at 25 ºC using a MicroCal ITC200 instrument. A solution of 50
µM SPSB2 was prepared in 50 mM phosphate, pH 7.4, 150 mM NaCl using a PDꢀ10 column
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 41
1
2
3
4
5
6
7
8
9
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
(GE Healthcare). Solutions of M1, M2, M3 and M4 were prepared as 5 mM stocks in the same
buffer. In each ITC experiment, the sample cell was filled with 300 µL of SPSB2 solution and
the syringe was filled with 40 µL of the peptidomimetic solution. During the experiment, the
SPSB2 solution was stirred at 1,000 rpm. The titration was initiated with the first injection of 0.2
µL followed by a series of 29 periodic injections (1 µL each). The data were analyzed using the
evaluation software Origin 7 (MicroCal Inc). Thermodynamic parameters were derived from a
nonlinear least squares fit to the data, using the “one set of sites” model.
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
19F NMR spectroscopy
19F NMR studies of M1, M2, M3 and M4 binding to 5ꢀFꢀTrpꢀSPSB2 were performed as
described previously.11 Experiments were conducted at 30ºC in a 50 mM phosphate buffer, pH
7.0, 50 mM NaCl, 2 mM DTT, 2 mM EDTA, 0.02% sodium azide at 564 MHz on a Bruker
19
Avance 600 MHz spectrometer equipped with a cryogenic probe tuned to F. Chemical shifts
were referenced against 0.005% trifluoroethanol as an internal reference. NMR experiments
were conducted with 50 µM 5ꢀFꢀTrpꢀSPSB2 and a 5ꢀfold molar excess of peptidomimetic (M1,
M2, M3 and M4). All spectra were processed using an exponential window function with a lineꢀ
broadening of 40 Hz.
Molecular dynamics simulations
MD simulations were performed with GROMACS 5.0.621 using the GROMOS 54a7 force
field22. The starting model of M1 bound to SPSB2 was built using the complex of the 20ꢀresidue
VASA peptide (PDB ID: 3EMW) as a template. The protein was prepared using Maestro
(Schrödinger, version 10.3); missing atoms were added and protonation states of ionisable
residues adjusted to pH 7.0. All hydrogens and added atoms were subjected to restrained
minimization with default settings. M1 was constructed from VASA peptide residues Asp184 to
ACS Paragon Plus Environment
22

Page 23 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
490
Asn188, which were modified by removing the Nꢀterminus nitrogen and replacing it with an
ether linkage to a 2ꢀoxyphenylamine group. An amide bond was created from the phenylamine
nitrogen to the Cꢀterminus of Asn188. Finally the SPSB2 proteinꢀM1 complex was minimized.
ATB23 was used to generate topology information for residue BzD1, which was added to the
GROMOS force field files. A pdb2gmx generated topology file for the linear peptide mimetic Hꢀ
BzD1ꢀI2ꢀN3ꢀN4ꢀN5ꢀOH was modified to make a suitable topology for M1.
491
492
493
494
495
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
496
497
498
499
500
501
502
503
504
505
506
MD simulations were performed on a SPSB2ꢀM1 complex solvated with SPC water model
containing 150 mM NaCl. Periodic boundary conditions were used with a dodecahedral cell.
Simulations were performed using the particle mesh Ewald method.24 Van der Waals and shortꢀ
range Ewald cutꢀoffs were set to 0.9 nm, bonds to hydrogen atoms were constrained with the
LINCS algorithm, temperature coupling was controlled via the vꢀrescale method25 at 300 K and
pressure was controlled via ParrinelloꢀRahman isotropic coupling26 at 1 bar. Following an initial
minimization, 200 ps of NVT ensemble simulation for equilibration was performed with
positionꢀrestraints on the heavy atoms of SPSB2 and M1 using a 2 fs time step. Similarly, 200 ps
of NPT ensemble simulation was performed with constraints. Finally, to achieve a NPT
production simulation time of 1 ꢀs, a time step of 5 fs with HMR27 to slow high frequency
motions was utilized.
507
508
509
510
511
512
Cell lysate inhibition assay
The cell lysate inhibition assay was performed on murine BMDM extracted and cultured
according to the protocol described previously.28 Cell lysates were incubated with 2.5 ꢂg/mL
recombinantly expressed GST protein (as a negative control), or GSTꢀtagged SPSB2ꢀSPRY
domain, both in the absence and presence of 10 ꢂM cyclic peptidomimetics M1, M2, M3 and
M4 for 2 h, followed by another 1 h of incubation with glutathione Sepharose 4B beads (GE
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 41
1
2
3
4
5
6
7
8
9
513
514
515
516
517
518
519
520
521
522
523
524
Healthcare) to pull down the SPSB2ꢀiNOS protein complex from the cell lysate. Proteins were
separated by SDSꢀPAGE under reducing conditions and electrophoretically transferred to
nitrocellulose membranes (Millipore). The membranes were blocked overnight in 10% (w/v)
skim milk prior to incubation with mouse monoclonal antiꢀiNOS primary antibody (BD
Biosciences) for 2 h. This was followed by another 1 h of incubation with peroxidaseꢀconjugated
sheep anti–mouse Ig antibody (GE Healthcare). Antibody binding was visualized with the ECL
system (GE Healthcare) after extensive washing with a phosphate buffer, saline and 0.1% Tween
20 (2 h, with a change of buffer every 20 min). To confirm equivalent protein loading, the
membrane was reblotted with mouse monoclonal antiꢀGST primary antibody for 2 h after
stripping off the antiꢀiNOS and peroxidaseꢀconjugated sheep anti–mouse Ig antibodies in 0.1 M
Gly, pH 2.9. The antiꢀGST antibody binding was visualized with peroxidaseꢀconjugated sheep
anti–mouse Ig antibody (GE Healthcare) and the ECL system (GE Healthcare).
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
525
ASSOCIATED CONTENT
526
527
528
529
530
531
Supporting information. Schematic representation and detailed description of methods used for
the synthesis and characterization of the building blocks required for the synthesis of
peptidomimetics M(1ꢀ4), schematic summary of methods used for the synthesis of
1
peptidomimetics M(1ꢀ4), HPLC profiles of the crude peptidomimetics M(1ꢀ4), H NMR and
HPLC profiles of purified peptidomimetics M(1ꢀ4) and chemical shift assignments for peptides
M(1ꢀ4). This material is available free of charge via the Internet at http://pubs.acs.org.
532
533
534
AUTHOR INFORMATION
Corresponding Author
* Phone: +613 9903 9167; Fax: +613 9903 9582; Email: Ray.Norton@monash.edu
ACS Paragon Plus Environment
24

Page 25 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
535
*Phone +613 9903 9044; Fax: +613 9903 9581; Email: Jonathan.Baell@monash.edu
Notes
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
7
8
9
The authors declare no competing financial interest.
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
ACKNOWLEDGMENTS
We thank Kade Roberts for helpful discussions. B. K. Yap is the recipient of the Academic Staff
Training Scheme Fellowship of Universiti Sains Malaysia. R. S. Norton and J. B. Baell
acknowledge fellowship support from the National Health and Medical Research Council of
Australia. This study was supported by the Australian National Health Medical Research Council
(NHMRC) Grants 1022693, as well as an NHMRC IRIISS grant 361646 and a Victorian State
Government Operational Infrastructure Scheme grant.
ABBREVIATIONS USED
SPSB2, SPRY domainꢀcontaining suppressor of cytokine signaling box protein 2; iNOS,
inducible nitric oxide synthase; KD, binding affinity; ITC, isothermal titration calorimetry; SPR,
surface plasmon resonance; RPꢀHPLC, reverseꢀphase high performance liquid chromatography;
koff, dissociation rate; BMDM, bone marrowꢀderived macrophages; IFNꢀγ, interferonꢀgamma;
HCTU, Oꢀ(1Hꢀ6ꢀChlorobenzotriazoleꢀ1ꢀyl)ꢀ1,1,3,3ꢀtetramethyluronium hexafluorophosphate;
DIPEA, N,Nꢀdiisopropylethylamine; HFIP, hexafluoroisopropanol; DODT, 3,6ꢀdioxaꢀ1,8ꢀ
octanedithiol; MeCN, acetonitrile; DSS, 3ꢀtrimethylsilylꢀ1ꢀpropanesulfonic acid sodium sulfate;
TOCSY, total correlation spectroscopy; DQFꢀCOSY, doubleꢀquantum filtered correlation
spectroscopy; DIPSI, decoupling in the presence of scalar interactions; WATERGATE, water
suppression by gradient tailored excitation; SOFAST, bandꢀselective optimized flipꢀangle short
transient; HEPES, 4ꢀ(2ꢀhydroxyethyl)ꢀ1ꢀpiperazineethanesulfonic acid; ATB, automated
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 41
1
2
3
4
5
6
7
8
9
557
558
559
topology builder; LINCS, linear constraint solver; HMR, hydrogen mass repartitioning; kon,
association rate; ECL, electrochemiluminescence.
REFERENCES
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
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
(1)
Kuang, Z.; Lewis, R. S.; Curtis, J. M.; Zhan, Y.; Saunders, B. M.; Babon, J. J.; Kolesnik,
T. B.; Low, A.; Masters, S. L.; Willson, T. A.; Kedzierski, L.; Yao, S.; Handman, E.;
Norton, R. S.; Nicholson, S. E., The SPRY domainꢀcontaining SOCS box protein SPSB2
targets iNOS for proteasomal degradation. J. Cell Biol. 2010, 190, 129ꢀ141.
(2)
Filippakopoulos, P.; Low, A.; Sharpe, T. D.; Uppenberg, J.; Yao, S.; Kuang, Z.; Savitsky,
P.; Lewis, R. S.; Nicholson, S. E.; Norton, R. S.; Bullock, A. N., Structural basis for Parꢀ
4 recognition by the SPRY domainꢀ and SOCS boxꢀcontaining proteins SPSB1, SPSB2,
and SPSB4. J. Mol. Biol. 2010, 401, 389ꢀ402.
(3)
(4)
Yap, B. K.; Leung, E. W.; Yagi, H.; Galea, C. A.; Chhabra, S.; Chalmers, D. K.;
Nicholson, S. E.; Thompson, P. E.; Norton, R. S., A potent cyclic peptide targeting
SPSB2 protein as a potential antiꢀinfective agent. J. Med. Chem. 2014, 57, 7006ꢀ7015.
(a) Andersson, A.; Baell, J. B.; Duggan, P. J.; Graham, J. E.; Lewis, R. J.; Lumsden, N.
G.; Tranberg, C. E.; Tuck, K. L.; Yang, A., ωꢀconotoxin GVIA mimetics based on an
anthranilamide core: effect of variation in ammonium side chain lengths and
incorporation of fluorine. Bioorg. Med. Chem. 2009, 17, 6659ꢀ6670; (b) Baell, J. B.;
Duggan, P. J.; Forsyth, S. A.; Lewis, R. J.; Lok, Y. P.; Schroeder, C. I., Synthesis and
biological evaluation of nonpeptide mimetics of ωꢀconotoxin GVIA. Bioorg. Med. Chem.
2004, 12, 4025ꢀ4037; (c) Baell, J. B.; Duggan, P. J.; Forsyth, S. A.; Lewis, R. J.; Lok, Y.
P.; Schroeder, C. I.; Shepherd, N. E., Synthesis and biological evaluation of
anthranilamideꢀbased nonꢀpeptide mimetics of ωꢀconotoxin GVIA. Tetrahedron 2006,
62, 7284ꢀ7292; (d) Baell, J. B.; Forsyth, S. A.; Gable, R. W.; Norton, R. S.; Mulder, R. J.,
Design and synthesis of typeꢀIII mimetics of ωꢀconotoxin GVIA. J. Comput. Aided Mol.
Des. 2001, 15, 1119ꢀ1136; (e) Baell, J. B.; Harvey, A. J.; Norton, R. S., Design and
synthesis of typeꢀIII mimetics of ShK toxin. J. Comput. Aided Mol. Des. 2002, 16, 245ꢀ
262; (f) Brady, R. M.; Baell, J. B.; Norton, R. S., Strategies for the development of
conotoxins as new therapeutic leads. Mar. Drugs 2013, 11, 2293ꢀ2313; (g) Brady, R. M.;
Hatzis, E.; Connor, T.; Street, I. P.; Baell, J. B.; Lessene, G., Synthesis of
ACS Paragon Plus Environment
26

Page 27 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
587
conformationally constrained benzoylureas as BH3ꢀmimetics. Org. Biomol. Chem. 2012,
10, 5230ꢀ5237; (h) Brady, R. M.; Zhang, M.; Gable, R.; Norton, R. S.; Baell, J. B., De
novo design and synthesis of a µꢀconotoxin KIIIA peptidomimetic. Bioorg. Med. Chem.
Lett. 2013, 23, 4892ꢀ4895; (i) Harvey, A. J.; Gable, R. W.; Baell, J. B., A threeꢀresidue,
continuous binding epitope peptidomimetic of ShK toxin as a Kv1.3 inhibitor. Bioorg.
Med. Chem. Lett. 2005, 15, 3193ꢀ3196; (j) Tranberg, C. E.; Yang, A.; Vetter, I.;
McArthur, J. R.; Baell, J. B.; Lewis, R. J.; Tuck, K. L.; Duggan, P. J., ωꢀConotoxin
GVIA mimetics that bind and inhibit neuronal Ca(v)2.2 ion channels. Mar. Drugs 2012,
10, 2349ꢀ2368.
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
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
(5)
(6)
Ripka, A. S.; Rich, D. H., Peptidomimetic design. Curr. Opin. Chem. Biol. 1998, 2, 441ꢀ
452.
Atherton, E.; Logan, C. J.; Sheppard, R. C., Peptide synthesis. Part 2. Procedures for
solidꢀphase synthesis using Nαꢀfluorenylmethoxycarbonylaminoꢀacids on polyamide
supports. Synthesis of substance P and of acyl carrier protein 65ꢀ74 decapeptide. J.
Chem. Soc., Perkin. Trans. 1 1981, 538ꢀ546.
(7)
(8)
(9)
(a) Shaka, A. J.; Lee, C. J.; Pines, A., Iterative schemes for bilinear operators; application
to spin decoupling. J. Magn. Reson. (1969) 1988, 77, 274ꢀ293; (b) Braunschweiler, L.;
Ernst, R. R., Coherence transfer by isotropic mixing: Application to proton correlation
spectroscopy. J. Magn. Reson. (1969) 1983, 53, 521ꢀ528.
(a) Bax, A.; Davis, D. G., Practical aspects of twoꢀdimensional transverse NOE
spectroscopy. J. Magn. Reson. (1969) 1985, 63, 207ꢀ213; (b) Hwang, T. L.; Shaka, A. J.,
Cross relaxation without TOCSY: transverse rotatingꢀframe Overhauser effect
spectroscopy. J. Am. Chem. Soc. 1992, 114, 3157ꢀ3159.
Derome, A. E.; Williamson, M. P., Rapidꢀpulsing artifacts in doubleꢀquantumꢀfiltered
COSY. J. Magn. Reson. (1969) 1990, 88, 177ꢀ185.
(10) Klebe, G., Applying thermodynamic profiling in lead finding and optimization. Nat. Rev.
Drug Discov. 2015, 14, 95ꢀ110.
(11) Leung, E. W.; Yagi, H.; Harjani, J. R.; Mulcair, M. D.; Scanlon, M. J.; Baell, J. B.;
19
Norton, R. S., F NMR as a probe of ligand interactions with the iNOS binding site of
SPRY domainꢀcontaining SOCS box protein 2. Chem. Biol. Drug Des. 2014, 84, 616ꢀ
625.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 41
1
2
3
4
5
6
7
8
9
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
(12) Bock, J. E.; Gavenonis, J.; Kritzer, J. A., Getting in shape: controlling peptide bioactivity
and bioavailability using conformational constraints. ACS Chem. Biol. 2013, 8, 488ꢀ499.
(13) Hewitt, W. M.; Leung, S. S. F.; Pye, C. R.; Ponkey, A. R.; Bednarek, M.; Jacobson, M.
P.; Lokey, R. S., Cellꢀpermeable cyclic peptides from synthetic libraries inspired by
natural products. J. Am. Chem. Soc. 2015, 137, 715ꢀ721.
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
(14) Makino, K.; Yamamoto, N.; Higuchi, K.; Harada, N.; Ohshima, H.; Terada, H.,
Phagocytic uptake of polystyrene microspheres by alveolar macrophages: effects of the
size and surface properties of the microspheres. Colloids Surf. B Biointerfaces 2003, 27,
33ꢀ39.
(15) Ball, J. B.; Alewood, P. F., Conformational constraints: nonpeptide βꢀturn mimics. J.
Mol. Recognit. 1990, 3, 55ꢀ64.
(16) Yao, S.; Masters, S. L.; Zhang, J. G.; Palmer, K. R.; Babon, J. J.; Nicola, N. A.;
1
15
Nicholson, S. E.; Norton, R. S., Backbone H, 13C and N assignments of the 25 kDa
SPRY domainꢀcontaining SOCS box protein 2 (SSBꢀ2). J. Biomol. NMR 2005, 31, 69ꢀ70.
(17) Wishart, D. S.; Bigam, C. G.; Yao, J.; Abildgaard, F.; Dyson, H. J.; Oldfield, E.;
1
13
15
Markley, J. L.; Sykes, B. D., H, C and N chemical shift referencing in biomolecular
NMR. J. Biomol. NMR 1995, 6, 135ꢀ140.
(18) Hwang, T. L.; Shaka, A. J., Water suppression that works. Excitation sculpting using
arbitrary waveꢀforms and pulsedꢀfield gradients. J. Magn. Reson. A 1995, 112, 275ꢀ279.
(19) (a) Piotto, M.; Saudek, V.; Sklenar, V., Gradientꢀtailored excitation for singleꢀquantum
NMR spectroscopy of aqueous solutions. J. Biomol. NMR 1992, 2, 661ꢀ665; (b) Sklenar,
1
V.; Piotto, M.; Leppik, R.; Saudek, V., Gradientꢀtailored water suppression for Hꢀ15N
HSQC experiments optimized to retain full sensitivity. J. Magn. Reson., Series A 1993,
102, 241ꢀ245.
(20) Schanda, P.; Kupče, E.; Brutscher, B., SOFASTꢀHMQC experiments for recording twoꢀ
dimensional heteronuclear correlation spectra of proteins within a few seconds. J.
Biomol. NMR 2005, 33, 199ꢀ211.
(21) Pronk, S.; Pall, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.; Apostolov, R.; Shirts, M. R.;
Smith, J. C.; Kasson, P. M.; van der Spoel, D.; Hess, B.; Lindahl, E., GROMACS 4.5: a
highꢀthroughput and highly parallel open source molecular simulation toolkit.
Bioinformatics 2013, 29, 845ꢀ854.
ACS Paragon Plus Environment
28

Page 29 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
649
(22) Schmid, N.; Eichenberger, A. P.; Choutko, A.; Riniker, S.; Winger, M.; Mark, A. E.; van
Gunsteren, W. F., Definition and testing of the GROMOS forceꢀfield versions 54A7 and
54B7. Eur. Biophys. J. 2011, 40, 843ꢀ856.
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
(23) Malde, A. K.; Zuo, L.; Breeze, M.; Stroet, M.; Poger, D.; Nair, P. C.; Oostenbrink, C.;
Mark, A. E., An automated force field topology builder (ATB) and repository: version
1.0. J. Chem. Theor. Comput. 2011, 7, 4026ꢀ4037.
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
(24) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G., A
smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103, 8577ꢀ8593.
(25) Bussi, G.; Donadio, D.; Parrinello, M., Canonical sampling through velocity rescaling. J.
Chem. Phys. 2007, 126, 014101.
(26) Parrinello, M.; Rahman, A., Polymorphic transitions in single crystals: A new molecular
dynamics method. J. App. Phys. 1981, 52, 7182ꢀ7190.
(27) Feenstra, K. A.; Hess, B.; Berendsen, H. J. C., Improving efficiency of large timeꢀscale
molecular dynamics simulations of hydrogenꢀrich systems. J. Comput. Chem. 1999, 20,
786ꢀ798.
(28) (a) Wormald, S.; Zhang, J. G.; Krebs, D. L.; Mielke, L. A.; Silver, J.; Alexander, W. S.;
Speed, T. P.; Nicola, N. A.; Hilton, D. J., The comparative roles of suppressor of
cytokine signalingꢀ1 and ꢀ3 in the inhibition and desensitization of cytokine signaling. J.
Biol. Chem. 2006, 281, 11135ꢀ11143; (b) Nicholson, S. E.; Novak, U.; Ziegler, S. F.;
Layton, J. E., Distinct regions of the granulocyte colonyꢀstimulating factor receptor are
required for tyrosine phosphorylation of the signaling molecules JAK2, Stat3, and p42,
p44MAPK. Blood 1995, 86, 3698ꢀ3704.
672
673
674
675
ACS Paragon Plus Environment
29