Some thermodynamic effects of varying nonpolar surfaces in protein-ligand interactions

Article abstract of DOI:10.1016/j.ejmech.2020.112771

Understanding how making structural changes in small molecules affects their binding affinities for targeted proteins is central to improving strategies for rational drug design. To assess the effects of varying the nature of nonpolar groups upon binding entropies and enthalpies, we designed and prepared a set of Grb2-SH2 domain ligands, Ac–pTyr–Ac₆c–Asn–(CH₂)_n–R, in which the size and electrostatic nature of R groups at the pTyr+3 site were varied. The complexes of these ligands with the Grb2-SH2 domain were evaluated in a series of studies in which the binding thermodynamics were determined using isothermal titration calorimetry, and binding interactions were examined in crystallographic studies of two different complexes. Notably, adding nonpolar groups to the pTyr+3 site leads to higher binding affinities, but the magnitude and energetic origins of these effects vary with the nature of the R substituent. For example, enhancements to binding affinities using aliphatic R groups are driven by more favorable changes in binding entropies, whereas aryl R groups improve binding free energies through a combination of more favorable changes in binding enthalpies and entropies. However, enthalpy/entropy compensation plays a significant role in these associations and mitigates against any significant variation in binding free energies, which vary by only 0.8 kcal?mol^?1, with changes in the electrostatic nature and size of the R group. Crystallographic studies show that differences in ΔG° or ΔH° correlate with buried nonpolar surface area, but they do not correlate with the total number of polar or van der Waals contacts. The relative number of ordered water molecules and relative order in the side chains at pTyr+3 correlate with differences in –TΔS°. Overall, these studies show that burial of nonpolar surface can lead to enhanced binding affinities arising from dominating entropy- or enthalpy-driven hydrophobic effects, depending upon the electrostatic nature of the apolar R group.

Full text of DOI:10.1016/j.ejmech.2020.112771

European Journal of Medicinal Chemistry 208 (2020) 112771
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Some thermodynamic effects of varying nonpolar surfaces in protein-
ligand interactions
David L. Cramer, Bo Cheng, Jianhua Tian, John H. Clements, Rachel M. Wypych,
Stephen F. Martin*
Department of Chemistry and the Institute of Cellular and Molecular Biology, The University of Texas, Austin, TX 78712, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Understanding how making structural changes in small molecules affects their binding affinities for
targeted proteins is central to improving strategies for rational drug design. To assess the effects of
varying the nature of nonpolar groups upon binding entropies and enthalpies, we designed and prepared
a set of Grb2-SH2 domain ligands, AcepTyreAc6ceAsne(CH2)neR, in which the size and electrostatic
nature of R groups at the pTyrþ3 site were varied. The complexes of these ligands with the Grb2-SH2
domain were evaluated in a series of studies in which the binding thermodynamics were determined
using isothermal titration calorimetry, and binding interactions were examined in crystallographic
studies of two different complexes. Notably, adding nonpolar groups to the pTyrþ3 site leads to higher
binding affinities, but the magnitude and energetic origins of these effects vary with the nature of the R
substituent. For example, enhancements to binding affinities using aliphatic R groups are driven by more
favorable changes in binding entropies, whereas aryl R groups improve binding free energies through a
combination of more favorable changes in binding enthalpies and entropies. However, enthalpy/entropy
compensation plays a significant role in these associations and mitigates against any significant variation
in binding free energies, which vary by only 0.8 kcalꢀmolꢁ1, with changes in the electrostatic nature and
Received 2 July 2020
Received in revised form
11 August 2020
Accepted 15 August 2020
Available online 23 August 2020
Keywords:
Protein-ligand interactions
Grb2 SH2 domain
isothermal titration calorimetry
Thermodynamics
Hydrophobic effects
size of the R group. Crystallographic studies show that differences in
D D
Gꢂ or Hꢂ correlate with buried
nonpolar surface area, but they do not correlate with the total number of polar or van der Waals contacts.
The relative number of ordered water molecules and relative order in the side chains at pTyrþ3 correlate
with differences in eTD
Sꢂ. Overall, these studies show that burial of nonpolar surface can lead to
enhanced binding affinities arising from dominating entropy- or enthalpy-driven hydrophobic effects,
depending upon the electrostatic nature of the apolar R group.
© 2020 Elsevier Masson SAS. All rights reserved.
1. Introduction and background
energetically unfavorable [4]. Nondirectional van der Waals (vdW)
contacts are easier to enhance, but dispersion forces are less spe-
Optimizing binding affinities of small molecules for protein
targets is an early goal in drug discovery. Generally, binding affin-
cific and weaker relative to polar interactions.
Toward improving binding entropies, conformational con-
straints are often introduced into flexible molecules [5e7]. How-
ever, the increases in binding free energies are usually less than
energetic estimates associated with restricting independent rotors.
We have also shown that ligand preorganization can lead to higher
affinity ligands because of enhanced binding enthalpies rather than
the expected improvements to binding entropies [8]. Several
computational investigations have been reported seeking insights
into this paradoxical finding [9,10], which seems a more general
phenomenon. Indeed, Klebe and coworkers recently reported that
protein binding of more flexible ligands is characterized by more
favorable binding entropies [11]. Another common strategy for
ities,
optimizing nonbonded, bimolecular interactions to enhance bind-
ing enthalpy,
Hꢂ, and by making binding entropy, Sꢂ, more
D
Gꢂ, of small molecules for proteins can be increased by
D
D
positive by reducing unfavorable solvation and conformational
parameters [1e3]. Enhancing binding enthalpies by changing or
introducing functional groups in a ligand is difficult because
hydrogen bonding interactions are strongly distance- and angle-
dependent and because desolvating polar functionality is
* Corresponding author.
E-mail address: sfmartin@mail.utexas.edu (S.F. Martin).
https://doi.org/10.1016/j.ejmech.2020.112771
0223-5234/© 2020 Elsevier Masson SAS. All rights reserved.

2
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
enhancing binding entropy is based upon the hydrophobic effect
and involves adding nonpolar groups to a lead because the des-
olvation of well-ordered water molecules from a nonpolar surface
is typically regarded as being entropically favorable [12e14].
However, this view is not uniformly true as we [15] and others
[16,17] have found that increasing nonpolar surface area by adding
methylene groups to a molecule may lead to increases in binding
affinity because increasingly more favorable binding enthalpies
dominate increasingly less favorable binding entropies. Indeed,
there is not a good correlation between burial of nonpolar surface
area and protein binding entropies [18], a fact that may not be
widely recognized.
there was no correlation of buried nonpolar surface area with
as might have been expected [29,30]. This trend toward favorable
changes in
Gꢂ and Hꢂ was not observed in a related study where
we examined the effects of increasing the length of an alkyl group
at the pTyrþ1 site, perhaps because there is not an incremental
increase in the number of vdW contacts with longer chain length
[27]. Indeed, varying the length of an alkyl chain in related ligands
has been shown to lead to different thermodynamic protein bind-
ing signatures in several other systems [31e34].
To explore further the consequences of increasing the nonpolar
surface area of ligands that bind to the Grb2 SH2 domain, we
focused upon congeneric analogs of AcepTyreValeAsneNH2 in
which hydrophobic groups were added at the pTyrþ3 subsite of 1
(Fig. 1). Previous studies revealed that the IC50s of such compounds
improved when nonpolar groups were added at this position
[35,36]. Molecular modeling suggested this increased potency
might be a consequence of vdW interactions of the nonpolar groups
at the C-terminus of the bound ligands with an extended hydro-
phobic region of the domain formed by the side chains of Lys109,
Leu111, and Phe119. Because energetic parameters and structures of
complexes were not examined, we embarked upon a systematic
investigation to determine binding enthalpies and entropies and
structures for the associations of ligands derived from 1 with the
Grb2 SH2 domain.
DCp
D
D
Making accurate predictions of protein binding affinities for a
congeneric series of ligands is a significant challenge, because the
multifactorial nature of protein-ligand associations makes it diffi-
cult to deconvolute the overall energetics into individual contri-
butions. One productive strategy toward addressing this complex
problem involves performing systematic studies of structure, dy-
namics, and energetics for a series of closely related protein-ligand
interactions. However, even when such data are available, corre-
lating variations in thermodynamic parameters with differences in
structural features is often problematic because of enthalpy/en-
tropy compensation [19,20] and because changes in
D D
Hꢂ and eT Sꢂ
often do not correspond to changes in
D
Gꢂ [21]. Exacerbating the
difficulty is that contributions of solvent reorganization [22] and
protein dynamics [23] to binding free energies are poorly under-
stood. For example, recent studies of changes in protein confor-
mational entropy that occur upon ligand association show protein
dynamics contribute to binding entropies. Although the effects of
differential protein dynamics on binding entropies for a series of
closely related protein-ligand interactions are largely unknown, it
seems likely that differences in protein dynamics make significant
energetic contributions to such associations. In this context and
contrary to what is commonly believed, binding of structurally
distinct molecules to the same protein may have no effect on
overall motion of the protein in the complex, or it may induce
decreased or increased protein motions [24].
Toward unraveling some of the unknowns inherent in corre-
lating structure and energetics in protein-ligand interactions, we
became interested in studying complexes of the Src homology 2
(SH2) domain of the growth receptor binding protein 2 (Grb2).
Grb2 is a cytosolic adapter protein that participates in the Ras signal
transduction pathway [25], and binding of pTyr residues on re-
ceptor tyrosine kinases (RTKs) to the SH2 domain of Grb2 leads to
Ras activation and consequent cell growth and differentiation.
Compounds that inhibit binding of the Grb2 SH2 domain to RTKs
and modulate Ras signaling might exhibit anticancer properties, so
understanding how modifying such molecules affects their affinity
for the Grb2 SH2 domain could facilitate the rational design of new
inhibitors.
2. Results and discussion
Ligand design and thermodynamic parameters. There are several
notable design elements associated with selecting compounds 1 to
assess the effects of changing the nonpolar surface area at the
pTyrþ3 site upon binding to the Grb2 SH2 domain. Because we
wanted a high affinity ligand as a starting point, we adopted a
known tactic and replaced the Val residue at pTyrþ1 of the pro-
totypical tripeptide Ac-pTyreValeAsneNH2 with an
a,a-cyclo-
hexane amino acid to give the known parent tripeptide 2 [15,37].
Having made this substitution, two variables remained e the
number of methylene groups, n, linking the C-terminal nitrogen
atom with the nonpolar group R and the nature of the R group. We
settled on an experimental design in which we would first establish
a standard linker length, n, and then characterize the effects of
varying the R group.
To evaluate the effects of changing the number of methylene
groups, n, in the linking chain of 1, we first synthesized 3e7 via
routine peptide couplings of N-substituted asparagine amides,
which were prepared via amide forming reactions, with a protected
Ac-pTyreAc6ceOH dipeptide in accord with the bond disconnec-
tions shown in Fig. 2. The binding energetics for each of these li-
gands for the Grb2 SH2 domain was determined by ITC using
HEPES, which has a relatively low ionization enthalpy, as the buffer
(Table 1). Although the binding free energies, D
Gꢂ, for 3e7 vary over
We conducted a number of studies using different sets of
pseudopeptide analogs of AcepTyreValeAsneNH2, the tripeptide
consensus sequence for potent Grb2 SH2 binders [8,15,26e28].
Relevant to the investigation of hydrophobic effects presented
herein, we used isothermal titration calorimetry (ITC) and X-ray
crystallography to probe the consequences of replacing the pTyrþ1
Val residue with cyclic,
a,a-disubstituted surrogates having
different ring sizes [15]. The incrementally more favorable binding
affinities that accompanied increasing ring size arose from favor-
able changes in binding enthalpies; changes in binding entropies
actually became less favorable as methylene groups were added. X-
ray crystallographic analysis of the respective complexes showed
them to be highly similar and revealed that an incremental increase
in van der Waals (vdW) contacts in the pTyrþ1 binding pocket
Fig. 1. Analogs of Ac-pTyreValeAsneNH2 with Ac6c at pTyrþ1 and nonpolar groups at
pTyrþ3 with variable regions shown in red. (For interpretation of the references to
color in this figure legend, the reader is referred to the Web version of this article.)
correlated with the observed favorable changes in
D
Hꢂ. However,

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
3
Fig. 3. Varying nonpolar groups
pTyreAc6ceAsneNH(CH2)3eR.
R
of pTyrþ3 substituted analogs of Ac-
Fig. 2. Determining standard linker length, n, for pTyrþ3 substituted analogs of Ac-
pTyreAc6ceAsneNH(CH2)neR.
enthalpies, D
Hꢂ, which are identical within experimental error.
a range of only 0.5 kcalꢀmolꢁ1, the binding enthalpy,
D
Hꢂ, is
Examination of the binding parameters for compounds 5 and
14e16, which have phenyl and substituted phenyl groups for R,
reveals that each of these ligands binds with higher affinity than
the parent tripeptide 2. The enhanced affinities for 5 and 14e16
with respect to 2 results from a combination of more favorable
binding entropies and enthalpies although the enthalpic contri-
butions relative to 2 are smaller. It is difficult to assess the impor-
significantly more favorable for 5. Accordingly, we surmised that a
linker having three methylene groups might be a suitable standard
for studying binding energetics of a series of compounds 1, when R
was either a monocyclic or acyclic group. However, we were also
interested in examining bicyclic aromatic R substituents, so we
synthesized 8e10 to determine whether a linking chain having
fewer methylene groups might be needed for larger aryl R groups.
The ITC data for these compounds (Table 1) reveal that 10, which
also has a three-methylene linker, had a more favorable binding
enthalpy and free energy than the corresponding analogs having
one or two methylene groups in the chain. These preliminary ex-
periments were not designed to optimize the linker length for all
possible R groups, but they provided guidance to select a three-
carbon linker as a standard spacer.
tance of the variations in
D
Hꢂ values because of the associated
experimental error coupled with possible enthalpic contributions
from any differences in proton exchange upon ligand binding,
which we have shown can contribute to slight dissimilarities in D
Hꢂ
values of congeneric ligands related to Ac-pTyreValeAsneNH2 [8].
Comparing 5 with 14e16 reveals that modifying the nature of a
substituent on the phenyl ring from strongly electron donating to
electron withdrawing has little effect on binding energetics. When
the R group is a larger aromatic ring such as an indole or naphthyl
ring (e.g., 10, 17e19), the more favorable affinity relative to 2 arises
from significantly more favorable binding enthalpies, although
there is a trend for slightly more favorable binding entropies.
Adding nonpolar substituents at the C-terminal pTyrþ3 site of
the parent tripeptide 2 clearly increases binding affinity. Overall,
the addition of an aryl R group at the pTyrþ3 site enhances binding
affinity more than introducing an alkyl R group, but the specific
energetic origins of these increases varies with the nature of R. In
particular, incorporating purely alkyl substituents at pTyrþ3
Having established a suitable linker length for our studies, we
directed our attention to varying the nonpolar R substituents in a
congeneric set of ligands related to 1. Accordingly, peptides 11e19
(Fig. 3) were prepared in analogy with the syntheses of 3e10. The
binding energetics for each of these compounds with the Grb2 SH2
domain was determined by ITC (Table 2) as before. The Ac-
pTyreAc6ceAsneNH(CH2)3eR analogs 11e13 wherein R is an alkyl
or a cycloalkyl group are equipotent, and each binds with a higher
affinity than the parent tripeptide 2. In each case, the increase in
affinity relative to 2 arises from more favorable binding entropy
terms, eTD
Sꢂ, that overcome slightly less favorable binding
Table 1
,[b]
ITC data for Ac-pTyreAc6ceAsneNH(CH2)neR analogs wherein linker length, n, is varied[a]
.
Ligand
Ka (x 107 Mꢁ1
)
D
Gꢂ (kcalꢀmolꢁ1
)
D
Hꢂ (kcalꢀmolꢁ1
)
eT
D
Sꢂ (kcalꢀmolꢁ1
)
28
3
4
5
6
7
8
9
10
0.7 0.1
1.2 0.1
1.5 0.2
2.6 0.7
1.5 0.1
2.8 0.1
1.1 0.1
2.4 0.1
4.9 1.3
ꢁ9.3 0.1
ꢁ9.6 0.1
ꢁ9.8 0.1
ꢁ10.1 0.2
ꢁ9.8 0.1
ꢁ10.2 0.1
ꢁ9.5 0.1
ꢁ10.1 0.1
ꢁ10.4 0.2
ꢁ8.5 0.4
ꢁ7.5 0.3
ꢁ7.7 0.1
ꢁ8.8 0.1
ꢁ7.7 0.2
ꢁ7.9 0.1
ꢁ6.7 0.1
ꢁ8.1 0.2
ꢁ10.2 0.3
ꢁ0.8 0.4
ꢁ2.1 0.8
ꢁ2.1 0.3
ꢁ1.3 0.8
ꢁ2.1 0.2
ꢁ2.3 0.1
ꢁ2.8 0.1
ꢁ2.0 0.1
ꢁ0.2 0.7
[a]
ITC experiments were conducted at 25 ꢂC in HEPES (50 mM) with NaCl (150 mM) at pH 7.45 0.05 as previously reported [8]. Three or more experiments were performed
for each ligand, and the averages are reported following normalization of the n values for each experiment by adjusting ligand concentration (See Supporting Information).
[b]
Errors in the thermodynamic values were determined by the method of Krishnamurthy [38].

4
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
Table 2
ITC data for Ac-pTyreAc6ceAsneNH(CH2)3eR analogs[a]e[c]
.
Ligand
Ka (x 107 Mꢁ1
)
D
Gꢂ (kcalꢀmolꢁ1
)
D
Hꢂ (kcalꢀmolꢁ1
)
eT
D
Sꢂ (kcalꢀmolꢁ1
)
D
Cp (calꢀmolꢁ1ꢀKꢁ1
)
2
0.7 0.1
1.6 0.1
1.7 0.3
1.9 0.3
2.6 0.7
4.0 0.8
5.0 0.8
4.7 0.4
6.0 0.3
4.9 1.3
3.5 0.3
2.7 0.5
ꢁ9.3 0.1
ꢁ9.8 0.1
ꢁ9.8 0.1
ꢁ9.9 0.1
ꢁ10.1 0.2
ꢁ10.3 0.1
ꢁ10.5 0.1
ꢁ10.5 0.1
ꢁ10.6 0.1
ꢁ10.4 0.2
ꢁ10.3 0.1
ꢁ10.1 0.1
ꢁ8.5 0.4
ꢁ7.9 0.2
ꢁ7.8 0.1
ꢁ8.0 0.4
ꢁ8.8 0.1
ꢁ8.6 0.5
ꢁ9.1 0.8
ꢁ9.1 0.5
ꢁ9.6 0.1
ꢁ10.2 0.3
ꢁ9.1 0.1
ꢁ9.1 0.2
ꢁ0.8 0.4
ꢁ1.9 0.2
ꢁ2.0 0.4
ꢁ1.9 0.4
ꢁ1.3 0.8
ꢁ1.7 0.6
ꢁ1.4 0.4
ꢁ1.4 0.2
ꢁ1.0 0.1
ꢁ0.2 0.7
ꢁ1.2 0.2
ꢁ1.0 0.6
ꢁ179 10
11
12
13
5
14
15
16
17
10
18
19
e
e
ꢁ215
8
ꢁ208 12
e
e
e
ꢁ224
e
8
e
e 217 10
[a]
ITC experiments were conducted at 25 ꢂC in HEPES (50 mM) with NaCl (150 mM) at pH 7.45 0.05 as previously reported [8]. Three or more experiments were performed
for each ligand, and the averages are reported following normalization of the n values for each experiment by adjusting ligand concentration (See Supporting Information).
[b]
Errors in the thermodynamic values were determined by the method of Krishnamurthy [38].
Values for DCp were determined from slopes of plots of values of D
Hꢂ at 15, 25, and 35 ꢂC, and the error was determined from the standard error in the slope.
[c]
improves binding free energies compared with 2 because more
favorable binding entropies dominate binding enthalpies that
become less favorable. When the substituent at the pTyrþ3 site
contains an aromatic ring, the increased affinities relative to 2
generally arise from a combination of more favorable binding en-
tropies and enthalpies. The trend of more favorable binding en-
tropies for the alkylated analogs 11e13 relative those bearing an
aromatic ring (e.g., 5, 10 and 14e19) is consistent with the known
tendency of alkyl groups to have more favorable dehydration en-
tropies than aryl groups [39].
To assess whether differences in desolvation might play a role in
the binding energetics,
DCp values, which have been purported to
be a barometer of hydrophobic effects in protein-ligand in-
teractions [40], for several representative ligands were determined
(Table 2). The more negative DCp values observed for 13, 5, 17, and
19 relative to the parent tripeptide 2 are consistent with the burial
of nonpolar surface upon associating with the Grb2 SH2 domain.
Relative to 2 the burial of nonpolar surface area for the cyclohexyl
analog 13 is accompanied by a more favorable entropy of binding,
whereas more favorable binding enthalpies attend the burial of
nonpolar surface area for the aryl analogs 5, 17, and 19. Notably,
these experiments show that burial of nonpolar surface can lead to
enhanced binding affinities arising from entropy- or enthalpy-
driven hydrophobic effects, depending upon the electrostatic na-
ture of the hydrophobic surface.
Structural analysis of selected complexes of the Grb2 SH2 domain.
Toward correlating differences in binding energetics with structure,
we embarked on crystallographic studies, but we only succeeded in
obtaining suitable crystals by cocrystallization of the Grb2 SH2
domain with 5 and 11. The corresponding structures were deter-
mined by X-ray crystallography (1.8 Å resolution) by molecular
replacement using the structure of the complex of 2 with the
domain (PDB ID: 3S8O) [15]. Each of these two structures contains
two complexes within the asymmetric unit.
The two complexes in the asymmetric unit of 5 with the domain
align are closely similar with a root mean square deviation (RSMD)
of 0.5 Å. The major differences between these complexes are in the
relative orientations of the BC loops, which align with a RMSD of
1.5 Å (Fig. 4a). Indeed, the BC loop is positioned ~1.5 Å further away
from the phosphate group with which it interacts in one of the
Fig. 4. Complexes in the asymmetric unit of the Grb2 SH2 domain and 5 (PDB ID:
6WM1). Nitrogen, oxygen, and phosphorous atoms are colored blue, red, and orange,
respectively. (a) Overlay of the two complexes in the asymmetric unit of the Grb2 SH2
domain and 5 following alignment of backbone atoms belonging to the domain. Car-
bon atoms belonging to the a complex are colored green while those belonging to the b
complex are colored cyan. (b) Overlay of the two poses of 5 in the a and b complexes. A
water molecule contacting the pTyr-1 carbonyl oxygen atom of the ligand (dashed
black line) but no amino acid residues in the a complex is shown as a red sphere.
Dashed black lines also represent the internal hydrogen bonds between the pTyr
carbonyl oxygen atoms and the amide nitrogen atoms of the pTyrþ3 sites. (c) and (d)
Interactions of 5 (carbon atoms cyan) with domain (gray), which is depicted as a
Connolly water surface, showing the interactions of the pTyrþ3 group with Lys109 and
Leu111 (green) in complex a (c) and complex b (d). (For interpretation of the references
to color in this figure legend, the reader is referred to the Web version of this article.)
complexes. As in complexes of other Grb2 SH2 ligands, 5 adopts a b-
turn conformation in both complexes that features an internal
hydrogen bond (dashed line) between the pTyr carbonyl oxygen
atom and the amide nitrogen atom of the pTyrþ3 site (Fig. 4b). As
suggested by earlier modeling studies [35], the phenyl group at the
pTyrþ3 site packs against the side chains of Lys109 and Leu111,
which are oriented similarly in the two complexes (Fig. 4c and d).
Following alignment of the domain, the pTyrepTyrþ1epTyrþ2
portion of the bound ligands compare with an RMSD of 0.9 Å,
reflecting some differences in the relative positions of the AcepTyr

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
5
and pTyrþ1 residues, but the greatest variations in atomic positions
of ligand atoms are at the pTyrþ3 site (RMSD ¼ 1.6 Å) (Fig. 4b).
The backbone atoms belonging to the Grb2 SH2 domain in the
two complexes in the asymmetric unit of 11 align with a RSMD of
0.7 Å, reflecting only relatively small differences that are primarily
found in the loop regions, but especially the BG loop
(RMSD ¼ 2.2 Å), which does not interact with 11 (Fig. 5a). As
observed in the complexes of 5 with the domain, 11 in each com-
There is a significant difference in the relative positions of the BC
loops in the two complexes in the asymmetric unit of 5 bound to
the Grb2 SH2 domain (Fig. 4a). Displacement of the BC loop in the a
complex (green) results in the loss of two direct polar contacts
(Table 3) between the phosphate group and the side chains of Ser88
and Ser90 (see Supporting Information). Perhaps to compensate for
the loss of these interactions, the side-chain of Arg67 in the a
complex is rotated so the guanidine group makes an additional
polar contact with the phosphate moiety and not with the pTyre1
carbonyl oxygen atom as in the b complex (see Supporting Infor-
mation). Lacking this interaction with Arg67, the pTyre1 acetyl
group of the ligand is oriented away from the domain, forming a
hydrogen bond with an ordered water molecule that has no other
observable close contacts (Fig. 4b). Variations in the number of
direct polar contacts in the a and b complexes are approximately
offset by differences in the number of water-mediated polar con-
tacts because there are two additional ordered water molecules at
the protein-ligand interface of the a complex (Table 3). One of these
water molecules mediates a contact between the side chain of
Ser88 and the non-bridging oxygen atoms of the phosphate group
of 5, whereas the other mediates a hydrogen bond contact between
the backbone carbonyl oxygen atom of the pTyrþ2 residue and the
backbone carbonyl atom of Leu120 (see Supporting Information).
We inventoried the van der Waals (vdW) contacts in these com-
plexes, and the only difference involves two vdW contacts between
the cyclohexane ring at the pTyrþ1 site of 5 and the domain
(Table 3).
plex binds in a characteristic b-turn conformation that is stabilized
by an internal hydrogen bond between the pTyr carbonyl oxygen
atom and the amide nitrogen atom of the pTyrþ3 site (Fig. 5b).
Following alignment of the domain, the AcepTyrepTyrþ2 seg-
ments of the ligands overlay with an RMSD of 1.2 Å, reflecting
differences in relative positions of backbone atoms, the pTyre1
acetyl group, the phenyl ring and the bridging oxygen atoms of the
phosphate group of the pTyr side chain, and the pTyrþ1 side-chain.
As observed for the complexes of 5, the greatest variations in the
atomic positions in the two complexes are in the pTyrþ3 side chain
with the isopropyl groups occupying distinctly different spatial
orientations (Fig. 5b). The respective positions of the alkyl group at
pTyrþ3 of 11 and the side chains of Lys109 and Leu111 in the two
complexes are also markedly different (Fig. 5c and d). Moreover, the
orientation of the Lys109 side chain in complex a of 11 (Fig. 5c) is
notably distinct from those observed in the two complexes of 5 (cf
Fig. 4c and d).
A similar analysis of the polar contacts in the two complexes of
11 with the Grb2 SH2 domain reveals that the bound form of 11 in
the b complex makes three more direct polar contacts to the
domain than in the a complex (Table 3). This dissimilarity is partly
due to the atypical spatial orientation of the Glu87 side-chain,
which is turned toward the phosphate group of 11 in the b com-
plex, thus making two contacts with 11 that are not observed in the
a complex (see Supporting Information). The other added contact
in the b complex is between the side chain of Lys109 and the
bridging oxygen atom of the phosphate group (see Supporting In-
formation). An inventory of the vdW contacts in the two complexes
reveals that there are two more vdW contacts in the a complex, one
involving the cyclohexyl moiety of the pTyrþ1 residue and the
other involving the pTyrþ3 side chain of 11.
We compared structural features in the complexes of 5 and 11
with the Grb2 SH2 domain to ascertain whether any differences
might be correlated with relative binding energy parameters. In
keeping with an approach we have used in the past [8], we consider
dissimilarities in protein complexes of congeneric ligands to be
meaningful only if they are greater than the corresponding differ-
ences between multiple copies of the same protein-ligand complex
in a crystallographic asymmetric unit. An inventory of the polar
contacts in the complexes of 5 and 11 with the Grb2 SH2 domain
reveals there are more direct polar contacts between 11 and the
domain, whereas there are more water-mediated contacts between
5 and the domain (Table 3). These dissimilarities are small, how-
ever, and comparable to the corresponding differences between
each of the two distinct complexes of 5 and 11 in their respective
asymmetric units. Moreover, the relative enthalpic contributions of
direct versus water-mediated contacts between a protein and a
bound ligand cannot be reliably predicted. The total number of
vdW contacts in the complexes of 5 and 11 is about the same
(Table 3), even though there are significant differences in the
average number of vdW contacts at the pTyrþ1 and pTyrþ3 sites
for 5 and 11. Although the differences in the number of vdW con-
tacts at pTyrþ1 are subtle, those at pTyrþ3 are readily apparent by
comparing the orientations and interactions of the nonpolar R
groups at the pTryþ3 sites with Lys109 and Leu111 in the respective
Fig. 5. Complexes in the asymmetric unit of the Grb2 SH2 domain and 11 (PDB ID:
6WO2); heteroatom labels same as in Fig. 3. (a) Overlay of the two complexes in the
asymmetric unit of the Grb2 SH2 domain and 11 following alignment of backbone
atoms belonging to the domain; carbon atom labels same as in Fig. 3. (b) Overlay of the
two poses of 11 in the two complexes in the asymmetric unit showing intramolecular
hydrogen bond between the pTyr carbonyl oxygen atom and the amide nitrogen atom
of the pTyrþ3 site (dashed black lines) showing the markedly different orientations of
the isopropyl groups. (c) and (d) Interactions of 11 (carbon atoms cyan) with domain
(gray), which is depicted as a Connolly water surface, showing the interactions of the
pTyrþ3 group with Lys109 and Leu111 (green) in complex a (c) and complex b (d). (For
interpretation of the references to color in this figure legend, the reader is referred to
the Web version of this article.)

6
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
Table 3
Thermodynamic and protein-ligand contact data for the two complexes in the asymmetric units of 5 and 11.
Ligand
D
Go
D
Ho
eT
D
So
Polar Direct Contacts[a]
Water-Mediated Contacts
vdW contacts[b] to pTyrþ1
vdW contacts[b] to pTyrþ3
D
CSAnp (Å2)
(kcalꢀmolꢁ1
)
5
ꢁ10.2
ꢁ8.8
ꢁ7.8
ꢁ1.4
ꢁ2.1
11
14
13
16
4
2
1
1
12
14
18
17
8
8
3
2
250
197
218
215
11
ꢁ9.9
[a
] Total direct and single water-mediated hydrogen bonding contacts between protein and ligand for which non-hydrogen donor-acceptor distances are within the range
2.5e3.5 Å (see Supporting Information).
[b
]
A hydrophobic vdW contact exists when the interatomic distance between a methylene group in the pTyrþ1 residue and a carbon, nitrogen, or oxygen atom in the
domain is in the range of 3.4e4.2 Å.
complexes (cf Fig. 4c and d and 5c,d). Because the number of polar
and vdW interactions in the complexes of 5 and 11 with the Grb2
SH2 domain are similar, it is difficult to attribute differences in
binding enthalpies with variations in the number of either polar or
vdW contacts.
number of protein-ligand complexes [18]. However, because crys-
tallographicdatawereonlyobtainedfor5and11, theseaveragevalues
may not be representative of all the complexes studied herein.
Synthetic chemistry. All Grb2 SH2 binding ligands in this study
were prepared using standard procedures for peptide synthesis
(Scheme 1). Briefly, commercially available amino acid ester 20 was
coupled with Cbz-Tyr-OH using 1e(3edimethylamino)propyl]-3-
ethylcarbodiimideꢀHCl (EDCI) and 1ehydroxybenzotriazole (HOBt)
in the presence of N-methylmorpholine (NMM) in DMF to give 21 in
73% yield. Global removal of the protecting groups by catalytic
On the other hand, there are some structural dissimilarities in
the complexes of 5 and 11 with the Grb2 SH2 domain that are
consistent with the observed differences in binding entropies. For
example, there are more interfacial water molecules in the struc-
ture of 5 with the Grb2 SH2 domain. Fixing a water molecule at a
protein-ligand interface will have an unfavorable entropic conse-
quence [41]. Although this is but one contributing factor to overall
binding entropy, consideration of the number of interfacial water
molecules suggests that association of 5 with the domain should be
entropically less favorable than for 11, a prediction that is consis-
tent with the experimental data (Table 3). Notably, however, we
have also shown in another study of complexes of similar pseu-
dopeptide ligands with the Grb2 SH2 domain that complexes with
fewer interfacial water molecules do not necessarily enjoy an entropic
benefit [15]. One cannot thus assume that differences in the number
of ordered water molecules in a complex correlates with binding
entropies. The relative order of amino acid side chains has also been
linked with binding entropies in complexes of a protein with a set
of congeneric ligands [42,43]. To the extent that multiple orienta-
tions of the isopropyl group in the a and b complexes of 11 (see
Fig. 5b) reflect it is less ordered than the phenyl group in the
complex of 5, one might expect that complex formation of 11 with
the domain would be entropically more favorable as is observed by
ITC. The less ordered nature of the pTyrþ3 side chain in the com-
plexes of 11 relative to those in complexes of 5 is also reflected in
the respective 2fo-fc electron density maps (see Supporting
Information).
Another significant energetic component of protein-ligand asso-
ciations involves solvent reorganization. Toward assessing the
possible importance of differences in ligand desolvation in complex
formation of 5 and 11 with the Grb2 SH2 domain, we estimated the
changes in nonpolar Connolly surface areas DCSAnp that occur upon
binding for 5 and 11 (Table 3). Because there is no crystal structure of
the monomeric Grb2 SH2 domainwithout a bound ligand [15,44], the
calculated values for
complexes, thereby leading some uncertainties. Further complicating
the analysis, there are notable dissimilarities in the values of CSAnp
DCSAnp are based solely upon structures of the
D
that are calculated for the individual complexes of 5 in the crystal-
lographic asymmetric unit. These caveats notwithstanding, there are
good correlations between
with either Cp or eT
lated average values of
DCSAnp and D D
Hꢂ and Gꢂ, although none
D
D D
Sꢂ. Namely, the dependence of Hꢂ on calcu-
D
CSAnp for 5 and 11 is ꢁ38 2 calꢀmolꢁ1Åꢁ2
,
whereas the dependence of D DCSAnp corre-
Gꢂ on average values of
sponds to a contribution of ꢁ46 0.3 calꢀmolꢁ1Åꢁ2. It is noteworthy
that this later value is significantly larger than the value
of ꢁ12 calꢀmolꢁ1Åꢁ2 that is obtained from the analysis of a large
Scheme 1. Synthesis of Ac-pTyreAc6ceAsneNH(CH2)neR analogs 3e19.

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
7
hydrogenolysis, followed by N-acetylation of the intermediate amino
acid 22 provided 23 in 91% overall yield for the two steps. Phos-
phorylation of the tyrosine hydroxyl group then gave the dipeptide
fragment 24 in 70% yield. Subsequent coupling of 24 with the N-
substituted asparagines 25e41, which were prepared in good overall
yields by sequential coupling of N-protected asparagines with the
appropriate amine using EDCI and HOBt and deprotection, delivered
3e19 in generally good overall yields. These compounds were puri-
fied by reverse phase HPLC prior to the ITC studies.
the electrostatic nature and size of the R group. Detailed experi-
mental investigations such as these in combination with compu-
tational and NMR studies that probe the contributions of solvent
reorganization and protein dynamics will eventually lead to a
better understanding of the origins of differences in protein binding
energetics for a series of closely related small molecules.
4. Experimental section
General chemistry. Solvents and reagents were reagent grade
and were used without purification, unless otherwise noted. N,N-
Dimethylformamide (DMF) was dried by passage through two
columns of activated molecular sieves. Dichloromethane (CH2Cl2),
triethylamine (Et3N), 2,6-lutidine, and N-methylmorpholine
(NMM) were distilled from calcium hydride. Removal of solvent
under reduced pressure was performed using a rotary evaporator at
25e30 ꢂC (bath temperature). Flash chromatography was per-
formed with the indicated solvents and Merck 250e400 mesh silica
gel. Purification by reverse phase high performance liquid chro-
matography (RP HPLC) was conducted using a binary solvent sys-
tem, where solvent system A was 0.1% aqueous TFA and solvent
system B was 0.1% TFA in acetonitrile, with a C18 column (10 mm
particle size, 300 Å pore size), 22 mm diameter ꢃ 250 mm (flow
rate of 8 mL/min), being used for preparative work and a C18 col-
umn (10 mm particle size, 300 Å pore size), 4.6 mm diameter ꢃ 250
(flow rate of 1 mL/min), being used for analytical work. Melting
points were determined on a melting point apparatus and are
uncorrected. Proton (1H) nuclear magnetic resonance (NMR)
spectra were obtained at the indicated field strength as solutions in
the indicated solvent. Chemical shifts are reported in parts per
3. Summary and conclusions
Toward developing a better understanding of the energetics of
hydrophobic effects in protein-ligand interactions, we examined
the associations of the Grb2 SH2 domain with a series of congeneric
ligands 1 bearing different alkyl and alkyl aryl groups at the pTyrþ3
position. The first set of compounds 2e10 was designed to establish
a suitable linker length, n, in 1. These ligands were synthesized, and
their free energies, enthalpies, and entropies of binding to the Grb2
SH2 domain were determined using ITC. Based upon these exper-
iments, we selected a three-methylene linker as the standard
spacer and prepared another set of compounds 11e19 (i.e., 1, n ¼ 3)
in which the electrostatic nature and size of the hydrophobic group
R was varied The binding energetics of these compounds for the
Grb2 SH2 domain were determined by ITC. Although the binding
free energies of 5, and 11e19 differed by 0.8 kcalꢀmolꢁ1, the binding
enthalpies varied over a range of 2.4 kcalꢀmolꢁ1, revealing that
enthalpy/entropy compensation plays a significant role in the
binding energetics for these associations.
We determined the structures of the complexes of 5 and 11
bound to the Grb2 SH2 domain at 1.8 Å resolution as a step toward
correlating structure and energetics. Although there are slight
variations in the number of direct and water-mediated polar con-
tacts between 5 and 11 and the domain, there is no difference in the
total number of polar interactions that might be reliably correlated
with the substantive differences in binding enthalpies, DDHꢂ.
Likewise, even though there are notable differences in the number
of vdW contacts at the pTyrþ1 and pTyrþ3 sites in the complexes of
5 and 11, the total number of vdW contacts in the complexes of the
two ligands is similar, so it is not possible to correlate differences in
vdW contacts with DDHꢂ. Conversely, several structural features are
consistent with observed differences in binding entropies. For
example, the presence of additional ordered water molecules in the
complex of 5 and the apparent lower order of the pTyrþ3 side chain
of 11 in its complex with the domain are in accord with the more
favorable binding entropy observed for the association of 11 with
million (ppm, d
) referenced relative to the center of the residual 1H
resonance of the solvent (CD3OD 3.30 ppm; DMSO‑d6 2.49 ppm;
D2O 4.67 ppm). Coupling constants are reported in hertz (Hz).
Splitting patterns are designated as s ¼ singlet; d ¼ doublet;
dd ¼ doublet of doublet; dd ¼ doublet of doublet of doublets;
t ¼ triplet; q ¼ quartet; p ¼ pentuplet; hep ¼ heptet; m ¼ multiplet;
comp ¼ overlapping multiplets of magnetically nonequivalent
protons; br ¼ broad; app ¼ apparent. Carbon 13 (13C) NMR spectra
were obtained at the field indicated strength as solutions in the
indicated solvent. Resonances are reported in ppm referenced from
the center of the 13C multiplet of the solvent (CD3OD 49.0 ppm;
DMSO‑d6 39.5 ppm). Spectra taken in D2O were referenced utilizing
an external standard. All new compounds were judged to be >95%
pure based upon the 1H NMR spectra and HPLC.
Cbz-Tyr-Ac6c-OBn (21). To a solution of CbzeTyr (630 mg,
2.0 mmol) and benzyl 1-aminocyclohexanecarboxylate (20)
(513 mg, 2.2 mmol) in DMF (10 mL) at ꢁ20 ꢂC was consecutively
added NMM (0.66 mL, 6.0 mmol), HOBt$H2O (540 mg, 4.0 mmol)
and EDCI (460 mg, 2.4 mmol). The mixture was slowly warmed
to ꢁ10 ꢂC over 1 h then stirred at room temperature for 20 h. The
reaction mixture was concentrated under reduced pressure. The
residue was dissolved in 100 mL CH2Cl2 (100 mL), and the organic
phase was washed with 1 M HCl (75 mL), saturated NaHCO3
(75 mL) and brine (75 mL). The organic layers were collected and
dried (MgSO4). The crude product was purified by silica gel chro-
matography (hexanes-EtOAc 1:1) to yield 770 mg (73%) of 21 as an
the domain. Differences in
D
CSAnp for 5 and 11 correlate with the
D
measured values of
D
Gꢂ and
Hꢂ but not with Sꢂ.
D
Cp or eT
D
Collectively, these studies show that adding nonpolar groups to
the pTyrþ3 position of 2 leads to analogs with higher binding af-
finities, but the specific energetic origins of these effects vary with
the electrostatic nature of the R group at the terminus of the three-
methylene spacer. For example, introduction of aliphatic R groups
leads to binding enhancements that are driven by more favorable
binding entropies that dominate slight decreases in binding en-
thalpies. On the other hand, appending a more polar, monocyclic
aryl R group leads to improved binding free energies that arise from
a combination of more favorable binding enthalpies and entropies,
whereas more favorable binding enthalpies dominate when the
aryl moiety is bicyclic. Hence, the burial of nonpolar surface in these
protein-ligand interactions can lead to enhanced binding affinities
arising from dominating entropy- or enthalpy-driven hydrophobic
effects, depending upon the electrostatic nature of the apolar R
group. However, enthalpy/entropy compensation mitigates against
any significant variation in binding free energies with changes in
amorphous white solid. 1H NMR (400 MHz, CDCl3)
d 7.38e7.27
(comp, 10 H), 7.00 (d, J ¼ 7.6 Hz, 2 H), 6.70 (d, J ¼ 7.6 Hz, 2 H), 6.20
(br s,1 H), 6.05 (br s,1H), 5.41 (br s,1 H), 5.16e5.03 (comp, 4 H), 4.33
(d, J ¼ 6.4 Hz, 1 H), 2.95 (dd, J ¼ 13.6, 5.6 Hz, 1 H), 2.85 (dd, J ¼ 13.6,
5.6 Hz, 1 H), 1.90 (d, J ¼ 13.2 Hz, 2 H), 1.76 (t, J ¼ 11.6 Hz, 2 H),
1.58e1.42 (comp, 3 H), 1.28e1.05 (comp, 3 H); 13C NMR (100 MHz,
CDCl3)
d 173.6, 170.2, 155.0, 135.8, 130.5, 128.5, 128.5, 128.2, 128.0,
115.6, 67.0, 58.9, 56.1, 37.2, 32.2, 31.9, 24.9, 21.2, 21.0; HRMS (CI) m/z
531.2490 [C31H35N2O6 (M þ H) requires 531.2495].

8
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
H2N-Tyr-Ac6c-OH (22). The benzyl carbamate 21 (625 mg;
16 h. The reaction mixture was concentrated under reduced pres-
sure, the residue was triturated with Et2O, and the precipitate thus
formed was collected by filtration. The solid was washed with Et2O
(10 mL), 1 M HCl (10 mL), saturated NaHCO3 (10 mL), and H2O
(10 mL) to afford CbzeNHeAsn-NHR, which was purified when
necessary by silica gel chromatography using a suitable mixture of
hexanes/EtOAc. A solution of the crude CbzeNHeAsn-NHR in
MeOH (40 mL) containing 10% Pd/C (40 mg) was stirred for 6e12 h
under H2 (1 atm) at room temperature. The mixture was filtered
through a pad of Celite, and the pad was washed with MeOH
(10 mL). The combined filtrates were concentrated under reduced
pressure to afford the asparagine amides H2N-Asn-NHR, which
were typically >95% pure by 1H NMR and used without further
purification.
1.18 mmol) was dissolved in MeOH (15 mL) containing 10% Pd/C
(100 mg, 10 wt %), and the resulting mixture was purged four times
with H2 and then stirred under H2 (1 atm) for 14 h. The mixture was
filtered through a pad of Celite, and the pad was washed with
MeOH. The combined filtrate and washings were concentrated to
dryness under reduced pressure to yield 343 mg (95%) of 23 as a
clear glass. 1H NMR (400 MHz, CD3OD)
d
7.13 (d, J ¼ 8.4 Hz, 2 H),
6.76 (d, J ¼ 8.4 Hz, 2 H), 4.04 (t, J ¼ 7.6 Hz, 1 H), 3.05 (dd, J ¼ 13.8,
8.2 Hz, 1 H), 2.96 (dd, J ¼ 13.8, 8.2 Hz, 1 H), 2.23 (d, J ¼ 13.6 Hz, 1 H),
1.85e1.76 (comp, 2 H), 1.67e1.63 (m, 1 H), 1.52e1.12 (comp, 5 H),
1.28e1.05 (q, J ¼ 12.8 Hz, 1 H); 13C NMR (100 MHz, CD3OD)
d 181.3,
169.6, 158.0, 131,6, 126.9116.7, 62.9, 56.2, 38.1, 35.4, 31.7, 26.6, 22.8,
22.7; HRMS (ESI) m/z 307.1652 [C16H23N2O4 (M þ H)þ requires
307.1657].
H-Asn-NH-CH2Ph (25). Prepared in 64% yield as a white solid;
Ac-Tyr-Ac6c-OH. (23) Acetic anhydride (7.5 mL) and 22 (634 mg,
2.07 mmol) in dioxane/H2O (1:1, 60 mL). This mixture was stirred at
room temperature for 2 h whereupon it was concentrated to dry-
ness under reduced pressure. Azeotropic removal of Ac2O was
accomplished via the addition of toluene (1.5 mL) to the reaction
vessel followed by concentration under reduced pressure. This
procedure was repeated two times followed by drying in vacuo to
afford 693 mg (96%) of 23 as a glass. 1H NMR (400 MHz, CD3OD)
mp 143e144 ꢂC. 1H NMR (400 MHz, CD3OD)
d 7.31e7.29 (comp,
3 H), 7.25e7.21 (comp, 2 H), 3.70 (dd, J ¼ 8.0, 5.1 Hz, 1 H), 3.31e3.30
(comp, 2 H), 2.61 (dd, J ¼ 15.5, 8.0 Hz, 1 H), 2.47 (dd, J ¼ 15.5, 8.0 Hz,
1 H), 1.82 (p, J ¼ 7.5 Hz, 2 H); 13C NMR (150 MHz)
d 175.1, 174.7,
138.6, 128.4, 127.4, 127.0, 52.2, 42.9, 39.9; HRMS (ESI) m/z 222.1237;
C
11H16N3O2 (M þ H) requires 222.1240.
H-Asn-NH-(CH2)2Ph (26). Prepared in 94% yield as a white
solid; mp 143e144 ꢂC. 1H NMR (400 MHz, CD3OD)
d 7.30e7.26
d
7.08 (d, J ¼ 8.4 Hz, 2 H), 6.79 (d, J ¼ 8.4 Hz, 2 H), 4.64 (dd, J ¼ 8.6,
(comp, 3 H), 7.24e7.18 (comp, 2 H), 3.68 (dd, J ¼ 8.0, 5.1 Hz, 1 H),
6.2 Hz, 1 H), 3.00 (dd, J ¼ 13.8, 6.2 Hz, 1 H), 2.76 (dd, J ¼ 13.8, 6.2 Hz,
1 H), 2.08e1.97 (comp, 2 H), 1.90 (s, 3 H), 1.76 (td, J ¼ 12.8, 3.6 Hz,
3.46e3.40 (comp, 2 H), 2.61 (dd, J ¼ 15.4, 8.0 Hz, 1 H), 2.47 (dd,
J ¼ 15.2, 8.0 Hz, 1 H); 13C NMR (150 MHz)
d 175.1, 174.5, 139.3, 128.7,
2 H), 1.61e1.21 (comp, 6 H); 13C NMR (100 MHz, CD3OD)
d
177.8,
128.4, 126.2, 52.0, 40.8, 39.5, 35.3; HRMS (ESI) m/z 236.1394;
173.2, 173.0, 157.2, 131.4, 129.2, 116.2, 60.3, 56.0, 38.2, 33.4, 33.0,
26.4, 22.5, 22.4; HRMS (ESI) m/z 349.1758 [C18H25N2O5 (M þ H)
requires 349.1763].
C
12H18N3O2 (M þ H) requires 236.1394.
H-Asn-NH-(CH2)3Ph (27). Prepared in 24% yield as a white solid;
mp 170e172 ꢂC. 1H NMR (400 MHz, CD3OD)
d 7.27e7.14 (comp,
(S)-1-{2-Acetamido-3-[4-{bis(benzyloxy)phosphoryloxy}
phenyl]propanamido}cyclohexane carboxylic acid (24). NMM
5 H), 3.64 (dd, J ¼ 8.0, 5.5 Hz, 1 H), 3.23 (dd, J ¼ 10.0, 5.5 Hz, 2 H),
2.66e2.58 (comp, 3 H), 2.46 (dd, J ¼ 15.0, 8.0 Hz, 1 H), 1.82 (p,
(42
m
L, 0.38 mmol) and TBSCl (57.1 mg, 0.38 mmol) were added
J ¼ 7.52 Hz, 2 H); 13C NMR (150 MHz)
d 176.2, 175.8, 143.0, 129.5,
129.4, 126.9, 53.3, 41.2, 40.0, 34.1, 32.2; HRMS (ESI) m/z 250.1304;
sequentially to a solution of the 23 (132 mg, 0.38 mmol) in DMF
(5 mL) at room temperature. After stirring for 1 h at room tem-
perature, the reaction was cooled to 0 ꢂC, 1-H-tetrazole (133 mg,
1.89 mmol) and dibenzyl diisopropylphosphoramidite (583 mg,
0.57 mL, 1.52 mmol) were added, and the solution was stirred at
0 ꢂC for 1 h and then 14.5 h at room temperature. The reaction was
cooled to 0 ꢂC, and 6 M tert-butyl hydroperoxide in decane (1 mL)
was added. After stirring for 30 min at 0 ꢂC and 4 h at room tem-
perature, the reaction was cooled to 0 ꢂC, and 10% aqueous NaHSO3
(5 mL) was added. The solution was stirred at 0 ꢂC for 30 min and at
room temperature for 2 h H2O (30 mL) was added, and the reaction
mixture was extracted with EtOAc (50 mL x 3). The combined
organic layers were washed with saturated NaHCO3 (25 mL), dried
(MgSO4), and concentrated under reduced pressure. The residue
was purified by silica gel chromatography (CH2Cl2eMeOH 10:1) to
afford 160 mg (70%) 24 as an amorphous white solid. 1H NMR
C
13H20N3O2 (M þ H) requires 250.1304.
H-Asn-NH-(CH2)4Ph (28). Prepared in 50% yield as a white
solid; mp 119e121 ꢂC. 1H NMR (400 MHz, CD3OD)
d 7.26e7.13
(comp, 5 H), 3.62 (dd, J ¼ 8.0, 5.5 Hz, 1 H), 3.23e3.19 (comp, 2 H),
2.66e2.58 (comp, 3 H), 2.46 (dd, J ¼ 15.0, 8.0 Hz, 1 H), 1.82 (p,
J ¼ 7.5 Hz, 2 H), 1.56 (p, J ¼ 7.5 Hz, 2 H); 13C NMR (150 MHz)
d 176.2,
175.8, 143.0, 129.5, 129.4, 126.9, 53.3, 41.2, 40.0, 34.1, 32.2, 30.9;
HRMS (ESI) m/z 286.1526; C14H21N3O2Na (M þ Na) requires
286.1531.
H-Asn-NH-(CH2)5Ph (29) Prepared in 70% yield as a glass. 1H
NMR (400 MHz, CD3OD)
d
7.26e7.11 (comp, 5 H), 3.63 (dd, J ¼ 8.0,
5.2 Hz, 1 H), 3.22 (dd, J ¼ 14.2, 6.8 Hz, 1 H), 3.15 (dd, J ¼ 14.2, 7.2 Hz,
1 H), 2.63e2.57 (comp, 3 H), 2.43 (dd, J ¼ 15.2, 8.0 Hz, 1 H), 1.63 (p,
J ¼ 7.6, 2 H), 1.54 (p, J ¼ 7.2 Hz, 2 H), 1.40e1.32 (comp, 2 H); 13C NMR
(100 MHz, CD3OD) d 176.3, 175.9, 143.8, 129.4 (2C), 129.3 (2C), 126.7,
(400 MHz, CDCl3)
d
7.69 (br s, 1 H), 7.34e7.27 (comp, 10 H), 7.18 (d,
53.4, 41.3, 40.4, 36.8, 32.4, 30.3, 27.5; HRMS (ESI) m/z 278.1863;
C
J ¼ 8.4 Hz, 2 H), 7.06 (d, J ¼ 8.4 Hz, 2 H), 5.08 (app d, J ¼ 6.4 Hz, 4 H),
4.83 (dd, J ¼ 14.4, 6.8 Hz, 1 H), 3.11e3.97 (comp, 2 H), 1.98e1.90
(comp, 2 H), 1.93 (s, 3 H), 1.77e1.65 (comp, 2 H), 1.56e1.02 (comp,
15H24N3O2 (M þ H)þ requires 278.1860.
H-Asn-NH-CH2(1-Me-1H-indol-3-yl) (30). Prepared in 83%
yield as a glass. 1H NMR (400 MHz, CD3OD)
d
7.58 (d, J ¼ 8.0 Hz,1 H),
6 H); 13C NMR (100 MHz, CDCl3)
d
176.7, 171.5, 171.4, 149.3, 149.2,
7.30 (d, J ¼ 8.4 Hz, 1 H), 7.16 (app td, J ¼ 7.6, 1.0 Hz, 1 H), 7.10 (s, 1 H),
7.04 (app td, J ¼ 7.6, 1.0 Hz, 1 H), 4.52 (dd, J ¼ 19.2, 14.8 Hz, 2 H), 3.70
(s, 3 H), 3.64 (dd, J ¼ 8.2, 5 Hz, 1 H), 2.61 (dd, J ¼ 15.2, 4.8 Hz, 1 H),
135.3,135.2,133.8,130.8,128.7, 128.6,128.0,120.0,119.98, 70.1, 70.0,
59.1, 54.4, 36.8, 32.4, 31.4, 25.0, 22.7, 21.2, 21.1; HRMS (ESI) m/z
607.2211 [C32H37N2O8P (M þ H) requires 607.2215].
2.42 (dd, J ¼ 15.2, 8.4 Hz, 1 H); 13C NMR (100 MHz, CD3OD)
d 175.9
General procedure for the synthesis of aspargine amides. N-
(2 C) 138.7, 129.1, 128.4, 122.8, 120.12, 119.8, 112.2, 110.3, 53.4, 41.3,
35.7, 32.7; HRMS (ESI) m/z 275.1501; C14H19N4O2 (M þ H)þ requires
275.1508.
Methylmorpholine
1ehydroxybenzotriazole hydrate (HOBt$H2O) (200 mg, 1.3 mmol)
and
(NMM)
(0.22
mL,
2.0
mmol),
1e(3edimethylamino)propyl]-3-ethylcarbodiimideꢀHCl
H-Asn-NH-(CH2)2(1-Me-1H-indol-3-yl) (31). Prepared in 85%
(EDCI) (150 mg, 0.78 mmol) were added sequentially to a solution
of Cbz-AsneH (190 mg, 0.71 mmol) and the appropriate amine,
RNH2, (0.65 mmol) in DMF (5 mL) at ꢁ30 ꢂC. The mixture was
allowed to warm slowly to room temperature and then stirred for
yield as a glass. 1H NMR (400 MHz, CD3OD)
d
7.57 (d, J ¼ 6.8 Hz,1 H),
7.29 (d, J ¼ 8.4 Hz, 1 H), 7.15 (dd, J ¼ 8.2, 7.2, 1.2 Hz, 1 H), 7.03 (dd,
J ¼ 7.8, 6.8, 0.8 Hz, 1 H), 6.99 (s, 1 H), 3.72 (s, 3 H), 3.62 (dd, J ¼ 8.0,
4.8 Hz,1 H), 3.48 (td, J ¼ 7.4, 4.0 Hz, 2 H), 2.93 (t, J ¼ 7.2 Hz, 2 H), 2.58

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
9
(dd, J ¼ 15.2, 4.8 Hz, 1 H), 2.38 (dd, J ¼ 15.2, 8.4 Hz, 1 H); 13C NMR
3.17e3.08 (comp, 2 H), 2.92e2.72 (comp, 2 H), 2.00e1.93 (comp,
2 H); HRMS (ESI) m/z 300.1712; C17H22N3O2 (M þ H) requires
300.1707.
(100 MHz, CD3OD)
d 176.1, 175.8, 138.7, 129.3, 128.1, 122.5, 119.7,
119.6, 112.6, 110.2, 53.3, 41.4, 41.1, 32.7, 26.1; HRMS (ESI) m/z
311.1479; C15H20N4O2Na (M þ Na)þ requires 311.1484.
H-Asn-NH-(CH2)3-(1-H-1H-indol-3-yl) (40). Prepared in 47%
H-Asn-NH-(CH2)3(1-Me-1H-indol-3-yl) (32). Prepared in 61%
yield as a yellow oil. 1H NMR (400 MHz, CD3OD)
d
7.52 (d, J ¼ 8.0 Hz,
yield as a yellow oil. 1H NMR (400 MHz, CD3OD)
d
7.52 (d, J ¼ 8.0 Hz,
1 H), 7.32 (d, J ¼ 8.0 Hz,1 H), 7.07 (t, J ¼ 7.6 Hz,1 H), 7.02 (s,1 H), 6.98
(t, J ¼ 7.6 Hz, 1 H), 3.61 (dd, J ¼ 8.0, 5.2 Hz, 1 H), 3.28e3.22 (m, 2 H),
2.77 (t, J ¼ 7.4 Hz, 2 H), 2.55 (dd, J ¼ 29.8, 15.2, 5.4 Hz, 1 H), 2.41 (dd,
J ¼ 15.2, 8.2 Hz, 1 H), 1.90 (p, J ¼ 7.2 Hz, 2 H). 13C NMR (100 MHz,
1H), 7.26 (d, J ¼ 8.4 Hz, 1 H), 7.13 (t, J ¼ 7.6 Hz,1 H), 7.01 (t, J ¼ 7.6 Hz,
1 H), 6.91 (s, 1H), 3.67 (s, 3 H), 3.61 (dd, J ¼ 8.0, 5.2 Hz, 1 H),
3.28e3.18 (comp, 2 H), 2.75 (t, J ¼ 7.4 Hz, 2 H), 2.59 (dd, J ¼ 15.2,
4.8 Hz, 1 H), 2.41 (dd, J ¼ 15.4, 8.2 Hz, 1 H), 1.88 (p, J ¼ 7.4 Hz, 2 H);
CD3OD)
d 176.0, 175.9, 138.2, 128.7, 123.0, 122.2, 119.5, 119.4, 115.6,
13C NMR (100 MHz, CD3OD)
d
176.3, 175.9, 138.7, 129.2, 127.6, 122.3,
112.2, 53.3, 41.1, 40.3, 30.9, 23.6; HRMS (ESI) m/z 289.1658;
119.7, 119.5, 115.0, 110.1, 53.4, 41.3, 40.3, 32.6, 31.0, 23.4; HRMS (ESI)
m/z 303.1816; C16H23N4O2 (M þ H) requires 303.1822.
C
15H21N4O2 (M þ H)þ requires 289.1663].
H-Asn-NH-(CH2)3-(1H-indol-1-yl) (41). Prepared in 58% yield as
H-Asn-NH-(4-methylpentyl) (33). Prepared in 67% yield as a
a yellow oil after purification by silica gel chromatography eluting
glass. 1H NMR (400 MHz, CD3OD)
d
3.41 (dd, J ¼ 7.1, 5.5 Hz, 1 H),
with CH2Cl2/MeOH (9:1) then CH2Cl2/MeOH (2:1). 1H NMR
3.24e3.12 (comp, 2 H), 2.62 (dd, J ¼ 15.2, 7.1 Hz, 1 H), 2.53e2.19
(comp, 2 H), 1.59e1.47 (comp, 3 H), 1.24e1.20 (m, 1 H), 0.89 (d,
J ¼ 6.5 Hz, 6 H); HRMS (ESI) m/z 216.1706; C10H22N3O2 (M þ H)
requires 216.1705.
(400 MHz, CD3OD)
d
7.53 (d, J ¼ 8.0 Hz, 1 H), 7.32 (dd, J ¼ 8.4, 0.8 Hz,
1 H), 7.21 (d, J ¼ 3.2 Hz, 1 H), 7.14 (td, J ¼ 7.6, 1.0 Hz, 1 H), 7.01 (td,
J ¼ 7.4, 0.8 Hz, 1 H), 6.42 (dd, J ¼ 3.2, 0.8 Hz, 1 H), 4.19 (t, J ¼ 7.0 Hz,
2 H), 3.59 (dd, J ¼ 7.6, 5.2 Hz, 1 H), 3.17 (td, J ¼ 6.8, 1.6 Hz, 2 H), 2.59
(dd, J ¼ 15.4, 5.4 Hz, 1 H), 2.44 (dd, J ¼ 15.4, 7.8 Hz, 1 H), 2.00 (p,
H-Asn-NH-(5-methylhexyl) (34). Prepared in 98% yield as a
white solid; mp 210 ꢂC (dec). 1H NMR (400 MHz, CD3OD)
d
3.61 (dd,
J ¼ 6.8 Hz, 2 H); 13C NMR (100 MHz, CD3OD)
d 175.7, 175.5, 137.4,
J ¼ 8.2, 5.1 Hz, 1 H), 3.24e3.12 (comp, 2 H), 2.62 (dd, J ¼ 15.2, 8.2 Hz,
130.3, 129.2, 122.4, 121.7, 120.2, 110.5, 101.9, 53.1, 44.7, 40.5, 38.1, 30.9;
HRMS (ESI) m/z 289.1659C; 15H21N4O2 (M þ H) requires 289.1668.
General procedure for the synthesis of tripeptides 3e19.
1 H), 2.42 (dd, J ¼ 15.2, 8.2 Hz, 2 H),1.59e1.47 (comp, 3 H),1.24e1.20
(m, 1 H), 0.89 (d, J ¼ 6.5 Hz, 6 H); 13C NMR (150 MHz)
d 221.5, 39.3,
39.1, 29.5, 27.9, 24.6, 21.8; HRMS (ESI) m/z 252.1682; C11H23N3O2Na
(M þ Na) requires 252.1688.
NMM (31 mL, 0.28 mmol), HOBt$H2O (29 mg, 0.19 mmol) and EDCI
(22 mg, 0.11 mmol) were added sequentially to a solution of the
appropriate asparagine amide (0.11 mmol) and 24 (57 mg,
0.09 mmol) in DMF (5 mL) at ꢁ30 ꢂC. The mixture was allowed to
warm slowly to room temperature and then stirred overnight. The
reaction mixture was concentrated under reduced pressure, and
the residue was triturated with Et2O to form a precipitate that was
collected by filtration. The solid was washed with Et2O (5 mL), 1 M
HCl (5 mL), saturated NaHCO3 (5 mL) and H2O (5 mL) to afford the
crude protected tripeptide. This was dissolved in MeOH (6 mL)
containing 10% Pd/C (10 mg), and the mixture was stirred for
6e12 h under H2 (1 atm) at room temperature. The mixture was
filtered through a pad of Celite, and the pad was washed with
MeOH (10 mL). The combined filtrates were concentrated under
reduced pressure to afford the crude tripeptide that was purified
via preparative RP HPLC as described.
H-Asn-NH-(CH2)3-cyclohexyl (35). Prepared in 72% yield as a
white solid; mp 167e168 ꢂC. 1H NMR (400 MHz, CD3OD)
d 4.13 (dd,
J ¼ 8.5, 4.8 Hz, 1 H), 3.25e3.12 (comp, 2 H), 2.84 (dd, J ¼ 17.2, 4.8 Hz,
1 H), 2.74 (dd, J ¼ 17.2, 4.8 Hz, 1 H), 1.79e1.62 (comp, 5 H), 1.55e1.51
(comp, 2 H), 1.27e1.18 (comp, 6 H), 0.94e0.84 (comp, 2 H); 13C NMR
(150 MHz) d 221.5, 211.3, 39.8, 37.5, 35.1, 34.5, 33.3, 26.6, 26.5, 26.2;
HRMS (ESI) m/z 256.2025; C13H26N3O2 (M þ H) requires 256.2020.
H-Asn-NH-(CH2)3-(p-tolyl) (36) Prepared in 40% overall yield as
a white solid; mp 123e125 ꢂC. 1H NMR (400 MHz, CD3OD)
d 7.06 (br
s, 4 H), 4.88 (comp, 5 H), 3.62 (dd, J ¼ 5.2, 8.0 Hz, 1 H), 3.19 (m, 2 H),
2.62e2.57 (comp, 3 H), 2.43, (dd, J ¼ 8.0, 15.2 Hz, 1 H), 2.27 (s, 3 H),
1.77e1.74 (comp, 2 H); 13C NMR (100 MHz, CD3OD)
d 174.9, 174.5,
138.4, 134.9, 128.6, 127.9, 52.0, 39.9, 38.6, 32.3, 30.9, 19.6; HRMS
(ESI) m/z 264.1705; C14H22N3O2 (M þ H) requires 264.1712.
H-Asn-NH-(CH2)3-(4-methoxyphenyl) (37). Prepared in 99%
yield as a white solid; mp 204e205 ꢂC. 1H NMR (400 MHz, CD3OD)
Tripeptide 3. Prepared from 25 in 72% overall yield according to
the general procedure for the synthesis of tripeptides. The crude
material was purified via preparative RP HPLC using a gradient of
d
7.10 (d, J ¼ 6.5 Hz, 2 H), 6.8 (d, J ¼ 6.5 Hz, 2 H), 3.66e3.62 (m, 1 H),
3.23e3.19 (comp, 2 H), 2.66e2.44 (comp, 4 H), 2.15, (s, 3 H), 1.78 (p,
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.31e7.24
J ¼ 7.1 Hz, 2 H); 13C NMR (150 MHz)
d
176.2, 175.8, 143.0, 129.5,
(comp, 5 H), 7.19e7.14 (comp, 4 H), 4.60 (dd, J ¼ 13.3, 4.8 Hz, 2 H),
4.47 (dd, J ¼ 15.2, 5.1 Hz, 1 H), 4.35 (dd, J ¼ 13.3 Hz, 4.8 Hz, 1 H), 2.97
(dd, J ¼ 14.4, 5.1 Hz, 1 H), 2.86 (dd, J ¼ 14.4, 5.1 Hz, 1 H), 2.78e2.71
(comp, 2 H), 2.00e1.92 (comp, 2 H), 1.82 (s, 3 H), 1.76e1.69 (comp,
2 H), 1.56e1.46 (comp, 3 H), 1.27e1.18 (comp, 3 H); 13C NMR
129.4, 126.9, 53.3, 41.2, 40.0, 34.1, 32.2, 30.9; HRMS (ESI) m/z
280.1658; C14H22N3O3 (M þ H) requires 280.1656.
H-Asn-NH-(CH2)3-(4-chlorophenyl) (38) Prepared in 70% yield
as a white solid, but cleavage of Cbz group was achieved by stirring
in neat CF3CO2H containing TfOH (8 equiv) and anisole (3 equiv) at
room temperature for 1 h CF3CO2H was removed under reduced
pressure, and the residue was triturated with Et2O. The precipitate
thus formed was collected by filtration and washed with Et2O
(2 ꢃ 5 mL) to afford the desired product; mp 166e168 ꢂC. 1H NMR
(150 MHz, CD3OD)
d 176.7, 175.6, 174.4, 174.3, 173.6, 139.8, 134.1,
131.3, 129.4, 128.4, 128.1, 121.3, 61.3, 61.2, 56.3, 52.3, 44.1, 37.1, 36.7,
33.0, 32.60, 26.2, 22.2; HRMS (ESI) m/z 654.2300; C29H38N5O9PNa
(M þ Na) requires 654.2305.
Tripeptide 4. Prepared from 26 in 34% overall yield according to
the general procedure for the synthesis of tripeptides. The crude
material was purified via preparative RP HPLC using a gradient of
(400 MHz, CD3OD)
d
7.26 (d, J ¼ 8.8 Hz, 2 H), 7.19 (d, J ¼ 8.0 Hz, 2 H),
4.12 (dd, J ¼ 5.2, 8.4 Hz, 1 H), 3.23 (comp, 2 H), 2.84 (dd, J ¼ 4.4,
16.8 Hz, 1 H), 2.75 (dd, J ¼ 8.4, 17.2 Hz, 1 H), 2.63 (t, J ¼ 7.6 Hz, 2 H),
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.27e7.24
1.82 (p, J ¼ 6.8 Hz, 2 H); 13C NMR (100 MHz, CD3OD)
d
171.8, 167.9,
(comp, 6 H), 7.17e7.14 (comp, 3 H), 4.65 (dd, J ¼ 9.5, 6.3 Hz, 1 H),
4.53 (dd, J ¼ 6.9, 5.1 Hz, 1 H), 3.42e3.38 (comp, 2 H), 3.13e3.09 (m,
1 H), 2.94e2.86 (m, 1 H), 2.84e2.68 (comp, 4 H), 1.98e1.95 (comp,
2 H), 1.88 (s, 3 H), 1.8e1.71 (comp, 3 H), 1.54 (comp, 3 H), 1.40e1.20
140.2, 131.3, 129.6, 128.0, 49.9, 38.7, 34.9, 31.8, 30.5; HRMS (ESI) m/z
284.1160 [C13H19N3O2Cl (M þ H) requires 284.1166].
H-Asn-NH-(CH2)3-(1-naphthyl) (39). Prepared in 62% crude
yield and used without purification in the next step. 1H NMR
(comp, 4 H); 13C NMR (150 MHz, CD3OD)
d 176.6, 175.7, 174.2, 173.6,
(400 MHz, CD3OD)
d
8.07 (d, J ¼ 8.3 Hz, 1 H), 7.89e7.82 (dd, J ¼ 7.6,
172.9, 140.5, 134.7, 131.4, 129.8, 129.4, 127.3, 121.4, 61.3, 56.3, 52.1,
42.4, 37.2, 36.7, 36.5, 32.8, 26.2, 22.3; HRMS (ESI) m/z 668.2456;
1.1 Hz,1 H), 7.72 (d, J ¼ 7.8 Hz,1 H), 7.56e7.43 (comp, 2 H), 7.43e7.32
(comp, 2 H), 4.16 (dd, J ¼ 4.9, 8.5 Hz, 1 H), 3.44e3.31 (comp, 2 H),
C30H39N5O9P (M ꢁ H) requires 668.2464.

10
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
Tripeptide 5. Prepared from 27 in 28% overall yield according to
the general procedure for the synthesis of tripeptides. The crude
2.98 (t, J ¼ 5.0 Hz, 2 H), 2.86 (dd, J ¼ 9.6, 6.0 Hz, 1 H), 2.81 (dd,
J ¼ 10.4, 4.8 Hz, 1 H), 2.72 (dd, J ¼ 10.4, 3.6 Hz, 1 H), 1.96e1.94 (m,
2 H), 1.97 (s, 3 H), 1.74e1.68 (comp, 2 H), 1.57e1.52 (comp, 3 H),
1.34e1.29 (m, 1 H), 1.22 (app t, J ¼ 6.0 Hz, 2 H); 13C NMR (150 MHz,
material was purified via preparative RP HPLC using a gradient of
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.24e7.19
(comp, 6 H), 7.14e7.12 (comp, 3 H), 4.65 (dd, J ¼ 9.5, 6.3 Hz, 1 H),
4.53 (dd, J ¼ 6.9, 5.1 Hz, 1 H), 3.26e3.22 (comp, 2 H), 3.13e3.11 (m,
1 H), 2.92e2.71 (comp, 3 H) 2.66e2.62 (t, J ¼ 7.5 Hz, 2 H), 2.05e1.93
(comp, 2 H), 1.88 (s, 3 H), 1.80e1.69 (comp, 3 H), 1.62e1.50 (comp,
CD3OD) d 176.6, 175.7, 174.2, 174.1, 173.6, 152.0, 138.7, 134.5, 131.4,
129.3, 128.2, 122.4, 121.4, 119.63, 119.60, 112.6, 110.2, 61.2, 56.3, 52.2,
41.6, 37.2, 36.8, 32.9, 32.73, 32.71, 26.3, 25.9, 22.4, 22.3; HRMS (ESI)
m/z 699.2899; C33H44N6O9P (M ꢁ H) requires 699.2902.
3 H), 1.38e1.19 (comp, 3 H); 13C NMR (150 MHz, CD3OD)
d
175.5,
Tripeptide 10. Prepared from 32 in 95% crude overall yield ac-
cording to the general procedure for the synthesis of tripeptides.
The crude tripeptide was purified via preparative RP HPLC with a
gradient of 15% B to 95% B over 30 min with a flow rate at 10 mL/
min to give 10 as a white solid; mp 139e143 ꢂC. 1H NMR (600 MHz,
174.5, 173.1, 172.4, 171.8, 151.6, 139.3, 133.1, 130.1, 128.6, 128.3, 126.1,
120.2, 60.0, 55.2, 51.0, 41.2, 36.0, 35.6, 35.3, 31.6 (2), 25.0, 21.1;
HRMS (ESI) m/z 660.2793; C31H43N5O9P (M þ H) requires 660.2793.
Tripeptide 6. Prepared from 28 in 15% overall yield according to
the general procedure for the synthesis of tripeptides. The crude
material was purified via preparative RP HPLC using a gradient of
CD3OD)
d
8.04 (s,1 H), 7.54 (dt, J ¼ 7.8,1.2 Hz,1 H), 7.27 (d, J ¼ 8.4 Hz,
1 H), 7.16 (d, J ¼ 8.4 Hz, 2 H), 7.13e7.10 (comp, 3 H), 6.98 (dd, J ¼ 7.8,
7.2, 1.2 Hz, 1 H), 6.94 (s, 1 H), 4.63 (dd, J ¼ 9.3, 6.3 Hz, 1 H), 4.55 (dd,
J ¼ 6.9, 5.1 Hz, 1 H), 3.68 (s, 3 H), 3.33e3.23 (m, 2 H), 3.09 (dd,
J ¼ 14.4, 6.0 Hz, 1 H), 2.87e2.73 (comp, 5 H), 2.04e1.93 (comp, 4 H),
1.86 (s, 3 H),1.82e1.70 (comp, 2 H), 1.59e1.52 (comp, 3 H), 1.37e1.19
10% B to 95% B over 30 min.1H NMR (400 MHz, CD3OD)
d 7.23e7.19
(comp, 6 H), 7.14e7.12 (comp, 3 H), 4.66 (dd, J ¼ 9.5, 6.1 Hz, 1 H),
4.53 (dd, J ¼ 6.9, 5.1 Hz, 1 H), 3.26e3.08 (comp, 3 H), 2.92e2.76
(comp, 3 H), 2.66e2.62 (t, J ¼ 7.5 H, 2 H), 2.05e1.93 (br s, 2 H), 1.88
(s, 3 H), 1.80e1.71 (comp, 3 H), 1.61e1.50 (comp, 3 H), 1.37e1.18
(comp, 3 H); 13C NMR (150 MHz, CD3OD)
d 176.7, 175.7, 174.3, 173.7,
(comp, 3 H); 13C NMR (150 MHz, CD3OD)
d
176.6, 175.6, 174.2, 173.6,
172.9, 138.7, 134.6, 131.4, 129.3, 127.6, 122.3, 121.4, 121.3, 119.8, 119.4,
115.3, 111.2, 110.1, 61.2, 56.3, 52.3, 40.6, 37.1, 36.7, 33.2, 32.6, 32.5,
31.0, 26.3, 23.4, 22.3, 22.3 (2C); HRMS (ESI) m/z 711.2913;
172.9, 143.7, 131.4, 129.5, 129.3, 126.7, 121.4, 61.2, 52.3, 40.4, 37.1,
36.7, 36.5, 33.0, 32.6, 29.9, 29.8, 26.2, 22.3, 22.2; HRMS (ESI) m/z
672.2804; C32H43N5O9P (M ꢁ H) requires 672.2802.
C
34H44N6O9P (M ꢁ H) requires 711.2903.
Tripeptide 7. Prepared from 29 in 56% overall yield according to
the general procedure for the synthesis of tripeptides. This material
was further purified via preparative RP HPLC with a gradient of 15%
B to 95% B over 30 min with a flow rate at 20 mL/min to give 7 as a
Tripeptide 11. Prepared from 33 in 87% overall yield according
to the general procedure for the synthesis of tripeptides. The crude
tripeptide was purified via preparative RP HPLC using a gradient of
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.26 (d,
white solid; mp 136e138 ꢂC. 1H NMR (500 MHz, CD3OD)
d
8.06 (br
J ¼ 8.2 Hz, 2 H), 7.14 (d, J ¼ 8.2 Hz, 2 H), 4.62e4.68 (m, 1 H),
4.47e4.54 (m, 1 H), 3.08e3.20 (comp, 3 H), 2.91 (dd, J ¼ 9.2 Hz,
7.2 Hz, 1 H), 2.80 (dd, J ¼ 12.3 Hz, 7.0 Hz, 1 H), 2.73 (dd, J ¼ 12.3 Hz,
7.0 Hz, 1 H), 2.00e1.92 (comp, 2 H), 1.91 (s, 3 H), 1.74e1.61 (comp,
2 H),1.52e1.49 (comp, 5 H),1.38 (m,1 H) 1.22e1.17 (comp, 4 H), 0.87
(d, J ¼ 6.5 Hz, 6 H); 13C NMR (125 MHz, CD3OD) 176.6, 175.7, 174.3,
173.6, 152.0, 134.6, 131.4, 121.4, 61.2, 56.4, 52.3, 41.0, 37.14, 37.05,
33.0, 32.5, 29.0, 28.2, 26.2, 23.0, 22.3, 22.2; HRMS (ESI) m/z
624.2806; C28H43N5O9P (M ꢁ H) requires 624.2804.
s, 0.9 H), 7.27 (d, J ¼ 10.2 Hz, 2 H), 7.24e7.21 (comp, 2 H), 7.15e7.10
(comp, 5 H), 4.66 (dd, J ¼ 11.1, 7.5 Hz, 1 H), 4.51 (dd, J ¼ 8.4, 6.0 Hz,
1 H), 3.19 (td, J ¼ 8.4, 3.6 Hz, 2 H), 3.13 (dd, J ¼ 17.0, 7.5 Hz, 1 H), 2.91
(dd, J ¼ 17.0, 10.8 Hz, 1 H), 2.81 (dd, J ¼ 18.6, 8.4 Hz, 1 H), 2.72 (dd,
J ¼ 18.6, 6.0 Hz, 1 H), 2.59 (t, J ¼ 9.3 Hz, 2 H), 2.02e1.94 (comp, 2 H),
1.90 (s, 3 H),1.80e1.69 (comp, 2 H),1.65e1.52 (comp, 7 H), 1.41e1.33
(comp, 3 H), 1.23 (app t, J ¼ 11.4 Hz, 2 H); 13C NMR (125 MHz,
CD3OD)
d 176.6, 175.7, 174.3, 173.7, 172.8, 151.9, 151.9, 143.9, 134.6,
131.4, 129.4 (2C), 129.3 (2C), 126.7, 121.4, 121.4, 61.3, 56.4, 52.3, 40.6,
37.1, 36.9, 36.7, 33.1, 32.5, 32.5, 30.1, 27.6, 26.3, 22.4, 22.3 (2C);
HRMS (ESI) m/z 686.2960; C33H45N5O9P (M ꢁ H) requires 686.2958.
Tripeptide 8. Prepared from 30 in 84% crude overall yield ac-
cording to the general procedure for the synthesis of tripeptides.
The crude tripeptide was purified via preparative RP HPLC with a
gradient of 15% B to 95% B over 30 min with a flow rate at 20 mL/
min to give 8 as a white solid; mp 162e164 ꢂC. 1H NMR (600 MHz,
Tripeptide 12. Prepared from 34 in 22% overall yield according
to the general procedure for the synthesis of tripeptides. The crude
tripeptide was purified via preparative RP HPLC using a gradient of
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.19 (d,
J ¼ 8.2 Hz, 2 H), 7.05 (d, J ¼ 8.2 Hz, 2 H), 4.57e4.55 (dd, 1 H, J ¼ 8.9,
6.2 Hz, 1 H), 4.54e4.48 (dd, J ¼ 7.2, 5.5 Hz, 1 H), 3.10e3.01 (comp,
3 H), 2.85e2.79 (m, 1 H), 2.74e2.61 (comp, 2 H), 2.00e1.88 (comp,
5 H), 1.74e1.61 (comp, 2 H), 1.52e1.49 (comp, 6 H), 1.29e1.12 (comp,
7 H) 0.78e0.76 (comp, 6 H); 13C NMR (125 MHz, CD3OD) 176.6,
175.7, 174.3, 173.6, 152.0, 134.6, 131.4, 121.4, 61.2, 56.4, 52.3,41.1, 37.1,
37.0, 33.0, 32.5, 29.0, 28.2, 26.2, 23.0, 22.2, 22.2; HRMS (ESI) m/z
638.2963; C29H45N5O9P (M ꢁ H) requires 638.2960.
CD3OD)
d
8.98 (br s, 0.8 H), 7.51 (d, J ¼ 5.6 Hz,1 H), 7.25 (d, J ¼ 5.6 Hz,
1 H), 7.15e7.13 (m, 2 H), 7.08 (d, J ¼ 5.6 Hz, 2 H),7.05e7.02 (m, 3 H),
4.66 (d, J ¼ 9.6 Hz, 1 H), 4.62 (dd, J ¼ 4.6, 3.4 Hz, 1 H), 4.51 (dd,
J ¼ 6.2, 4.2 Hz, 1 H), 4.47 (d, J ¼ 9.6 Hz, 1 H), 3.65 (s, 3 H), 2.84e2.75
(m, 3 H), 2.62 (dd, J ¼ 9.6, 6.0 Hz, 1 H), 1.94e1.89 (comp, 2 H), 1.76 (s,
3 H), 1.74e1.64 (comp, 2 H), 1.64e1.50 (comp, 3 H), 1.32e1.17 (comp,
Tripeptide 13. Prepared from 35 in 75% overall yield according
to the general procedure for the synthesis of tripeptides. The crude
tripeptide was purified via preparative RP HPLC using a gradient of
3 H); 13C NMR (150 MHz, CD3OD)
d 176.7, 175.7, 174.1, 173.6, 172.6,
151.9, 138.7, 134.6, 131.4, 129.1, 128.5, 122.7, 121.3, 120.1, 120.0, 112.5,
110.3, 61.1, 56.1, 52.2, 36.9, 36.8, 35.9, 33.2, 32.8, 32.4, 26.2, 22.3,
22.2; HRMS (ESI) m/z 707.2566; C32H41N6O9NaP (M þ Na)þ requires
707.2565.
Tripeptide 9. Prepared from 31 in 88% crude overall yield ac-
cording to the general procedure for the synthesis of tripeptides.
The crude tripeptide was purified via preparative RP HPLC with a
gradient of 15% B to 95% B over 30 min with a flow rate at 20 mL/
min to give 9 as a white solid; mp 157e159 ꢂC. 1H NMR (600 MHz,
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.28e7.26
(comp, 2 H), 7.15e7.13 (comp, 2 H), 4.68e4.64 (m, 1 H), 4.54e4.48
(m, 1 H), 3.19e3.07 (comp, 3 H), 2.92e2.89 (dd, J ¼ 13.7 Hz, 9.6 Hz,
1 H), 2.83e2.79 (comp, 2 H), 2.00e1.92 (br d, 2 H), 1.88 (s, 3 H),
1.74e1.61 (comp, 7 H), 1.59e1.49 (comp, 5 H), 0.87e0.85 (comp,
2 H), 1.24e1.08 (comp, 9 H) 0.95e0.81 (comp, 2 H); 13C NMR
(125 MHz, CD3OD) 177.60, 176.58, 175.6, 174.3, 173.7, 173.0, 172.8,
152.2, 134.3, 131.4, 131.3, 121.4, 61.3, 61.2, 60.2, 56.4, 55.7, 52.2, 41.0,
38.8, 38.0, 37.0, 36.7, 35.7, 34.5, 33.3, 31.8, 32.4, 27.77, 27.68, 26.4,
26.2, 22.45, 22.42, 22.37, 22.31, 22.2; HRMS (ESI) m/z 664.3120;
CD3OD)
d
8.02 (br s, 0.65 H), 7.57 (d, J ¼ 5.2 Hz, 1 H), 7.28 (d,
J ¼ 5.6 Hz,1 H), 7.22 (d, J ¼ 5.6 Hz, 2 H), 7.15e7.12 (m, 3 H), 7.02e7.00
(m, 2 H), 4.63 (dd, J ¼ 6.0, 4.0 Hz, 1 H), 4.56 (dd, J ¼ 4.6, 3.4 Hz, 1 H),
3.73 (s, 3 H), 3.49 (t, J ¼ 4.8 Hz, 2 H), 3.07 (dd, J ¼ 9.4, 4.2 Hz, 1 H),
C
31H47N5O9P (M ꢁ H) requires 664.3117.
Tripeptide 14. Prepared from 36 in 36% yield according to the
general procedure for the synthesis of tripeptides. The crude

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
11
tripeptide was purified by RP HPLC using a gradient of 10% B to 95%
B (20 mL/min) to give 14 as a white solid; mp 156e158 ꢂC. 1H NMR
CD3OD)
d
8.04 (s, 1 H), 7.54 (d, J ¼ 7.8, 1 H), 7.30 (d, J ¼ 7.8 Hz, 1 H),
7.18 (d, J ¼ 8.4 Hz, 2 H), 7.11 (d, J ¼ 8.4 Hz, 2 H), 7.05 (app td, J ¼ 7.5,
1.2 Hz, 1 H), 7.03 (s, 1 H), 6.95 (dd, J ¼ 7.8, 7.2, 1.2 Hz, 1 H), 4.63 (dd,
J ¼ 9.3, 6.3 Hz, 1 H), 4.55 (dd, J ¼ 6.9, 5.1 Hz, 1 H), 3.32e3.24 (comp,
2 H), 3.09 (dd, J ¼ 14.1, 6.3 Hz, 1 H), 2.88e2.73 (comp, 5 H),
2.03e1.94 (comp, 4 H), 1.86 (s, 3 H), 1.81e1.70 (comp, 2 H),
1.59e1.51 (comp, 3 H), 1.38e1.28 (m, 1 H), 1.26e1.19 (comp, 2 H);
(600 MHz, CD3OD)
d
7.22 (d, J ¼ 8.4 Hz, 2 H), 7.12 (dd, J ¼ 1.2, 9.0 Hz,
2 H), 7.06 (d, J ¼ 7.8 Hz, 2 H), 7.01 (d, J ¼ 7.8 Hz, 2 H), 4.64 (dd, J ¼ 6.6,
9.0 Hz, 1 H), 4.51 (dd, J ¼ 4.8, 6.6 Hz, 1 H), 3.21 (comp, 2 H), 3.10 (dd,
J ¼ 6.0, 13.8 Hz, 1 H), 2.88 (dd, J ¼ 9.0, 14.4, 1 H), 2.82 (dd, J ¼ 7.2,
15.6 Hz,1 H), 2.73 (dd, J ¼ 4.8,15.6 Hz,1 H), 2.58 (comp, 2 H), 2.25 (s,
3 H), 1.97 (comp, 2 H), 1.88 (s, 3 H), 1.76 (comp, 4 H), 1.56e1.51
(comp, 3 H), 1.34e1.21 (comp, 3 H); 13C NMR (150 Hz, CD3OD)
13C NMR (150 MHz, CD3OD)
d 176.7, 175.7, 174.3, 173.7, 172.9, 138.3,
134.4, 131.3, 128.8, 123.0, 122.2, 121.4, 119.5, 119.4, 115.8, 112.2, 111.2,
61.3, 56.4, 52.3, 40.7, 37.2, 36.8, 33.2, 32.6, 31.0, 26.3, 23.6, 22.3, 22.3
(2C); HRMS (ESI) m/z 697.2756; C33H42N6O9P (M ꢁ H) requires
697.2754.
Tripeptide 19. Prepared from 41 in 54% crude overall yield ac-
cording to the general procedure for the synthesis of tripeptides.
The crude tripeptide was purified via preparative RP HPLC with a
gradient of 15% B to 95% B over 30 min with a flow rate at 10 mL/
min to give 20 as a white solid; mp 145e148 ꢂC. 1H NMR (600 MHz,
d
176.6, 175.7, 174.2, 173.6, 172.9, 151.8, 151.7, 140.0, 136.3, 134.7,
131.4, 130.0, 129.41, 129.38, 121.32, 121.29, 61.2, 56.3, 52.3, 40.3, 37.1,
36.7, 33.6, 33.1, 32.5, 32.3, 26.2, 22.3, 22.2, 21.1; HRMS (ESI) m/z
696.27756; C32H44N5O9NaP (M þ Na)þ requires 696.27689.
Tripeptide 15. Prepared from 37 in 72% overall yield according
to the general procedure for the synthesis of tripeptides. The crude
tripeptide was purified via preparative RP HPLC using a gradient of
10% B to 95% B over 30 min 1H NMR (400 MHz, CD3OD)
d 7.22 (d,
J ¼ 8.2 Hz, 2 H), 7.14e7.10 (comp, 4 H), 6.77 (d, J ¼ 8.2 Hz, 2 H), 4.65
(dd, 1.0 H, J ¼ 9.2, 6.2 Hz, 1 H), 4.54e4.50 (m, 1 H), 3.73 (s, 3 H),
3.25e3.16 (comp, 2 H), 3.11 (dd, J ¼ 14.0, 6.2 Hz, 1 H), 2.91e2.75
(comp, 3 H), 2.57 (br t, J ¼ 7.52 Hz, 2 H), 2.05e1.88 (comp, 2 H), 1.88
(s, 3 H), 1.85e1.68 (comp, 5 H), 1.62e1.48 (comp, 4 H), 1.40e1.18
CD3OD)
d
8.11 (s,1 H), 7.51 (dt, J ¼ 7.8, 0.9 Hz,1 H), 7.44 (d, J ¼ 8.4 Hz,
1 H), 7.28 (s, 1 H), 7.22 (d, J ¼ 8.4 Hz, 2 H), 7.13e7.09 (comp, 3 H),
6.99 (dd, J ¼ 7.5, 7.2, 0.6 Hz, 1 H), 6.39 (dd, J ¼ 3.0, 1.2 Hz, 0.1 H), 4.66
(dd, J ¼ 9.0, 6.6 Hz, 1 H), 4.51 (dd, J ¼ 7.2, 4.8 Hz, 1 H), 4.23 (t,
J ¼ 7.2 Hz, 2 H), 3.26e3.18 (comp, 2 H), 3.11 (dd, J ¼ 13.8, 6.6 Hz,1 H),
2.91e2.84 (comp, 2 H), 2.78 (dd, J ¼ 15.6, 4.8 Hz, 1 H), 2.10e1.92
(comp, 4 H), 1.88 (s, 3 H), 1.82e1.72 (m, 2 H), 1.60e1.52 (comp, 3 H),
1.39e1.32 (m, 1 H), 1.26e1.19 (comp, 2 H); 13C NMR (150 MHz,
(comp, 4 H); 13C NMR (125 MHz, CD3OD)
d 176.6, 175.7, 174.3, 173.6,
172.9, 159.3, 151.74, 151.68, 135.2, 134.7, 131.4, 130.4, 121.33, 121.29,
114.8, 61.2, 56.3, 55.6, 52.3, 40.3, 37.1, 36.6, 33.14, 33.11, 32.4, 26.2,
22.3, 22.2; HRMS (ESI) m/z 689.2799; C32H43N5O10P (M ꢁ H) re-
quires 689.2784.
Tripeptide 16. The protected tripeptide was prepared from 38 in
16% overall yield according to the general procedure for the syn-
thesis of tripeptides, but debenzylation was achieved by stirring in
neat CF3CO2H with added anisole (3 equiv) at room temperature for
50 min. The solvent was removed under reduced pressure, and the
tripeptide was purified by RP HPLC using a gradient of 10% B to 95%
B (20 mL/min) to give 16 as a white solid; mp 165e167 ꢂC. 1H NMR
CD3OD)
d 176.7, 175.7, 174.5, 173.7, 173.2, 137.4, 134.2, 131.3, 130.2,
129.3, 122.3, 121.6, 121.4, 121.4, 120.1, 110.6, 61.2, 56.4, 52.5, 44.6,
38.1, 37.2, 36.6, 33.1, 32.7, 30.8, 26.3, 22.3, 22.3, 22.3; HRMS (ESI) m/z
697.2756; C33H42N6O9P (M ꢁ H) requires 697.2749.
General biochemical and biophysical procedures. All solu-
tions were prepared from distilled water that was passed through a
Nanopure water purification system (Barnstead) to give a resistivity
within the range 17.0e18.2 MUcm. Chromatography columns were
stored in a 20:80 v/v ethanol/water mixture at 4 ꢂC.
(600 MHz, CD3OD)
d
7.22e7.18 (comp, 6 H), 7.13 (d, J ¼ 8.4 Hz, 2 H),
Expression and purification of Grb2 SH2 domain. Expression
and purification of the Grb2 SH2 domain has been described in
detail elsewhere [8]. Briefly, the DNA construct containing the QE60
plasmid and residues 53e163 of the Grb2 SH2 domain was ob-
tained from the Schering-Plough Research Institute and expressed
in E. coli (SG13009, Qiagen). Cultures were grown at 30 ꢂC in LB
media containing 0.1 g/L ampicillin (Acros Organics) and 0.035 g/L
kanamycin (Sigma-Aldrich) to an OD600nm of 0.5e0.8 at which
4.64 (dd, J ¼ 6.6, 9.6 Hz, 1 H), 4.50 (dd, J ¼ 4.8, 6.6 Hz, 1 H), 3.20 (m,
2 H), 3.10 (dd, J ¼ 6.6, 14.4 Hz, 1 H), 2.87 (dd, J ¼ 9.0, 14.4 Hz, 1 H),
2.81 (dd, J ¼ 7.2, 15.6 Hz, 1 H), 2.74 (dd, J ¼ 6.8, 15.6 Hz, 1 H), 2.63
(comp, 2 H), 1.97 (comp, 2 H), 1.88 (s, 3 H), 1.83 (p, J ¼ 6.6 Hz, 2 H),
1.75 (comp, 2 H), 1.56e1.52 (comp, 3 H), 1.36e1.22 (comp, 3 H); 13
C
NMR (150 MHz, CD3OD)
d 176.6, 175.7, 174.3, 173.6, 173.0, 142.0,
132.6, 131.24, 131.19, 129.4, 121.38, 121.35, 61.2, 56.4, 52.3, 40.1, 37.1,
36.6, 33.3, 33.2, 32.4, 32.0, 26.2, 22.3, 22.2; HRMS (ESI) m/z
692.2258; C31H40ClN5O9P (M ꢁ H) requires 692.2252.
time
expression
was
induced
by
isopropyl-b-D-1-
thiogalactopyranoside (IPTG, Acros Organics, 1 mM) After 15 h of
incubation followed by sedimentation of the suspension by
centrifugation, the cells were resuspended in 25 mM TRIS, 1 mM
EDTA, pH 7.5 and lysed using a French Press under a pressure of 500
psi. The lysate was centrifuged and the supernatant purified on a
phosphotyrosine affinity column followed by dialysis of the eluent
containing the protein in 25 mM TRIS, 1 mM EDTA, pH 7.5. The
dialysed protein was further purified on a Q-Sepharose FF column
(GE Healthcare).
Tripeptide 17. Prepared from 39 in 64% overall yield according
to the general procedure for the synthesis of tripeptides. The crude
tripeptide was purified via preparative RP HPLC using a gradient of
10% B to 95% B over 30 min 1H NMR (600 MHz, CD3OD)
d 8.09 (d,
J ¼ 8.5 Hz, 1 H), 7.82 (d, J ¼ 8.0 Hz, 1 H), 7.68 (d, J ¼ 7.9 Hz, 1 H),
7.48e7.41 (comp, 2 H), 7.39e7.32 (comp, 2 H), 7.16 (d, J ¼ 8.5, 2 H),
7.09 (d, J ¼ 8.5 Hz, 2 H), 4.63 (dd, J ¼ 6.3, 9.2 Hz, 1 H), 4.55 (dd,
J ¼ 5.0, 6.9 Hz, 1 H), 3.37e3.31 (comp, 3 H), 3.12 (app t, J ¼ 7.8 Hz,
2 H), 3.07 (dd, J ¼ 6.3, 14.7 Hz, 1 H), 2.87e2.81 (comp, 2 H), 2.75 (dd,
J ¼ 5.0, 15.6 Hz, 1 H), 1.84 (s, 3 H), 2.03e1.93 (comp, 3 H), 1.81e1.68
(comp, 2 H), 1.58e1.50 (comp, 3 H), 11.36e1.20 (comp, 3 H); 13C
Isothermal Titration Calorimetry. The eluent fractions from
the Q-Sepharose column containing the Grb2 SH2 domain were
combined, placed in 15 mL Centriplus concentrators (Millipore,
MWCO ¼ 2000), and centrifuged (3000 g) at 4 ꢂC until the volume
NMR (150 MHz, CD3OD)
d 176.6, 175.7, 174.3, 173.6, 173.0, 139.3,
135.5,134.4,133.2,131.3,129.7,127.7,127.1,126.9,126.6,126.5,125.0,
121.31, 121.28, 66.9, 61.2, 57.5, 57.4, 56.3, 52.3, 40.7, 37.1, 36.6, 33.2,
32.5, 31.6, 31.2, 26.2, 22.26, 22.23; HRMS (ESI) m/z 710.2940;
had been reduced to 800e1000 mL (ca 20e30 h) giving a concen-
tration of ~30e100 mg/mL of a mixture of monomeric and dimeric
Grb2 SH2 domain [44]. This solution of Grb2 SH2 domain was
directly injected using a 1 mL syringe onto a XK column with AK
fittings packed with Superose 12 size-exclusion media prepared
according to the manufacturer instructions (GE Healthcare) [45].
The protein was eluted by washing the column with a 70:30 v/v
mixture of buffer A and buffer B (0.5 mL/min), and dimeric and
monomeric Grb2 SH2 domain were eluted sequentially using
C
35H45N5O9P (M þ H) requires 710.2949.
Tripeptide 18. Prepared from 40 in 70% crude overall yield ac-
cording to the general procedure for the synthesis of tripeptides.
The crude tripeptide was purified via preparative RP HPLC with a
gradient of 15% B to 95% B over 30 min with a flow rate at 10 mL/
min to give 19 as a white solid; mp 162e165 ꢂC. 1H NMR (600 MHz,

12
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
approximately 0.75e1 bed volumes. The fractions containing pure
monomeric Grb2 SH2 domain were combined to a final concen-
tration of 250e500 mM and dialyzed (3 h minimum equilibration
time) three times in 3.5 L ITC buffer (50 mM HEPES, 150 mM NaCl,
pH 7.50) at 4 ꢂC. The final HEPES dialysis buffer was retained and its
pH re-adjusted to 7.50 using either HCl or NaOH. The dialyzed
centrifuged at 3000 g and 4 ꢂC until the protein concentration was
within the range 5e15 mg/mL. The ligands were each dissolved in
this solution to give a protein-ligand molar ratio of 1:2. These so-
lutions were each heated to 50 ꢂC for 10 min to convert any Grb2
SH2 domain-swapped dimer to the monomeric domain [44],
filtered through a 0.45 mm PVDF filter disk, cooled, and stored at
protein was diluted to a concentration of 65e85
mM in the final
4 ꢂC. Crystal Screens I, II, and Lite (Hampton Research) were used to
identify initial crystallization conditions; additional screening was
performed if necessary. Crystals were cultivated by the hanging-
HEPES dialysis buffer prior to ITC experiments.
Solutions of ligands 2e19 were prepared by dissolving the
lyophilized ligand, which was weighed to the nearest 0.01 mg, in
the final HEPES dialysis buffer to give an approximately 1 mM so-
lution. The buffer and the solutions of protein and ligands were
drop, vapor-diffusion method, employing 7 mL drops and 350 mL
well solutions in standard 24-well, flat-bottom, polystyrene plates.
Specific details pertaining to the growth and diffraction of crystals
of Grb2 SH2 complexed with each of the ligands are given.
Grb2 SH2 complexed with 5. An aqueous solution containing a
1.5 M ratio of 5 to protein (ca. 10 mg/mL) was prepared. This so-
filtered (0.20 mm PDVF) and degassed prior to the ITC experiment.
ITC experiments were performed in at least duplicate experi-
ments employing independently prepared aliquots of ligand, and in
some cases protein, using a Microcal VP-ITC equilibrated at 25 ꢂC or
other appropriate temperature. Between experiments involving the
same ligand, the sample cell was thoroughly rinsed with water
followed by the final HEPES dialysis buffer. The injection apparatus
was cleaned sequentially with water and methanol and dried with
a stream of dry argon. Between experiments involving different
ligands, the sample cell and injection apparatus were cleaned by
soaking in 5% Contrad solution (Decon Laboratories) for 24 h, fol-
lowed by the above-mentioned rinse procedures. Typically, ~35 x
lution (4.0
0.1 M HEPES, pH 7.5, and 25% w/v polyethylene glycol, MW 10,000
(3.0 L) and allowed to equilibrate with the aforementioned pre-
mL) was mixed with a precipitant solution containing
m
cipitant well solution (350
4 weeks.
mL) at 298 K. Useable crystals grew after
Grb2 SH2 complexed with 11. An aqueous solution containing a
1.5 M ratio of 11 to protein (ca. 10 mg/mL) was prepared. This so-
lution (4.0
0.2 M sodium citrate tribasic dihydrate, 0.1 M HEPES, pH 7.5, and
20% v/v 2-propanol (3.0 L Hampton crystal screen I, condition no.
27) and allowed to equilibrate with the aforementioned precipitant
mL) was mixed with a precipitant solution containing
7e9
mL injections (200 s intervals) of ligand were made into the
m
sample cell containing protein, preceded by one 2
mL injection to
account for initial heat of mixing at the tip, which was omitted prior
to data analysis. Raw data were integrated, background heats for
injecting ligand into buffer were subtracted, and Microcal Origin
graphing software (version 7.0) was used to determine the ther-
well solution (350
weeks.
mL) at 298 K. Useable crystals grew after four
Collection of X-ray diffraction data. Prior to data collection,
crystals were cryoprotected by transferring them to solutions
containing salt, buffer and precipitant concentrations equal to their
theoretical initial concentrations in the hanging drop experiment,
yet containing glycerol in the concentration range 25e35% (v/v).
Crystals were allowed to equilibrate in this solution for 0.5e1 min.
Once equilibrated, the crystals were removed with a standard wire
loop, flash frozen in liquid N2, and equilibrated for about 30 s prior
to affixing to the goniometer. X-ray diffraction data were collected
at 100 K using a Rigaku RAXIS IV area detector positioned on a
Rigaku RU200 rotating angle X-ray generator operated at 40 kV and
modynamic parameters n (number of binding sites), Ka and
D
Hꢂ.
Titrations for which the initial concentration yielded n values > 1.20
were rejected, and the ligand concentration was adjusted as needed
to give a normalized n value of 1.00 in accord with the known 1:1
binding stoichiometry.
applying the modified Arrhenius equation,
calculated by applying the Gibbs relationship
D
Gꢂ was calculated indirectly from Ka by
D
Gꢂ ¼ eRTlnKa.
D
Sꢂ was
D
Gꢂ
¼
D D
Hꢂ e T Sꢂ.
Prior to performing ITC titrations of flexible and constrained ligands
with new batches of purified Grb2 SH2 domain, an ITC experiment
using the tripeptide AcepTyreValeAsneNH2 was performed, and
the thermodynamic values obtained were compared to those pre-
viously obtained in order to verify protein purity and folding.8
Representative ITC data for each ligand are given in Supporting
Information.
70 mA producing CuK
a graphite-monochromatic radiation
(1.5418 nm). Data frames were collected in 0.5ꢂ intervals with
exposure times of 120
100e120 mm.
s at crystal-to-detector distances of
Data processing and structure refinement. Data frames were
processed and scaled using HKL2000 [46]. CCP4 (Phaser) [47] was
used to identify a molecular replacement solution from a published
Grb2 SH2/peptide crystal structure (PDB ID: 3C7I) Structures were
manipulated using the program COOT [48] and refined using CCP4
(Refmac5). Density maps were created using CCP4 (FFT). Ligands
containing amino acid replacements were docked into the protein
model and the necessary topology files were built using the online
program PRODRG [49]. The PDB ID codes for the structures of 5 and
11 with the Grb2 SH2 domain are 6WM1 and 6WO2, respectively.
Molecular surface area calculations. Changes in nonpolar and
The temperature dependence of
D
Hꢂ,
D
Cp, of binding for ligands
2, 5, 13, 17, and 19 was obtained by measuring
D
Hꢂ in duplicate at
three different temperatures between 15 and 35 ꢂC according to the
procedure outlined above and determining the slope of the plot of
D
Hꢂ vs. temperature. The error in
the standard error in the slope.
DCp was obtained by determining
Treatment of concentration error. Errors in the thermody-
namic values were determined as previously described [38]. Briefly,
errors of 5.0 and 1.5% were assumed for weight and volume,
respectively, in the preparation of ligand solutions, giving an error
of 5.2% in ligand concentration. This error was propagated through
polar Connolly surface area,
DCSAnp and DCSAp [50], were calcu-
the thermodynamic relations
D
Gꢂ ¼ eRTlnKa and
D
Gꢂ
¼
D
Hꢂ
Gꢂ,
e
lated using Macromodel v. 9.1 [51] based on the assumption that no
change in the conformation of either the ligand or the protein oc-
curs on complex formation. Specifically, water and other solvent
molecules were removed from each X-ray structure and the CSAs of
the complexes were calculated. The ligands were removed from
each complex and the CSAs of the remaining structures were
assumed to be that of the apo protein, which varied <0.3% from one
structure to another. Likewise, CSAs of the ligands removed from
the complexes were assumed to be that of the unbound ligands in
solution. Subtraction of the latter two CSAs from the former gave
TD
Sꢂ to arrive at errors in the thermodynamic parameters Ka,
D
D
Hꢂ, and
D
Sꢂ.
Preparation of protein/ligand solutions for crystallization.
Prior to the cultivation of crystals of the domain complexed with
each ligand, purified protein was dialyzed twice in 3.5 L of distilled
water that had been passed through a Nanopure water purification
system (Barnstead) to give
17.0e18.2 M cm. The resulting protein solution was placed in
15 mL Centriplus concentrators (Millipore, MWCO ¼ 2000) and
a resistivity within the range
U

D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
13
correct ligand preorganization for the S2 pocket of thrombin: a study by
isothermal titration calorimetry, molecular dynamics, and high-resolution
crystal structures, ChemMedChem 11 (2016) 309e319.
D
CSA values for the formation of each complex. A probe radius of
1.4 Å was employed.
[7] A. Sandner, T. Huefner-Wulsdorf, A. Heine, T. Steinmetzer, G. Klebe, Strategies
for late-stage optimization: profiling thermodynamics by preorganization and
salt bridge shielding, J. Med. Chem. 62 (2019) 9753e9771.
Associated content
[8] J.E. DeLorbe, J.H. Clements, M.G. Teresk, A.P. Benfield, H.R. Plake,
L.E. Millspaugh, S.F. Martin, Thermodynamic and structural effects of confor-
mational constraints in protein-ligand interactions. Entropic paradoxy asso-
ciated with ligand preorganization, J. Am. Chem. Soc. 131 (2009)
16758e16770.
[9] Y. Shi, C.Z. Zhu, S.F. Martin, P. Ren, Probing the effect of conformational
constraint on phosphorylated ligand binding to an SH2 domain using polar-
izable force field simulations, J. Phys. Chem. B 116 (2012) 1716e1727.
[10] A. Li, M.K. Gilson, Protein-ligand binding enthalpies from near-millisecond
simulations: analysis of a preorganization paradox, J. Chem. Phys. (2018)
149, 072311.
[11] B. Wienen-Schmidt, H.R.A. Jonker, T. Wulsdorf, H.-D. Gerber, K. Saxena,
D. Kudlinzki, S. Sreeramulu, G. Parigi, C. Luchinat, A. Heine, H. Schwalbe,
G. Klebe, Paradoxically, most flexible ligand binds most entropy-favored:
intriguing impact of LigandFlexibility and solvation on drug-kinase binding,
J. Med. Chem. 61 (2018) 5922e5933.
Supporting Information
Representative ITC data for compound 3e19 at 298 K and for
compounds 5, 13, 17, and 19 at 288 and 308 K; DCp plots for 5, 13, 17,
and 19; crystallographic data for complexes of the Grb2 SH2
domain with 5 and 11 include complete diffraction data and
refinement statistics, density difference maps and omit maps of the
ligands in the complexes, protein-ligand polar contact diagrams,
table of B-factors, and tables correlating thermodynamic and DCSA
data and listing the number of direct and water-mediated protein-
ligand contacts as well as vdW contacts.
[12] J. Kyte, The basis of the hydrophobic effect, Biophys. Chem. 100 (2003)
193e203.
Accession codes
[13] D. Chandler, Interfaces and the driving force of hydrophobic assembly, Nature
437 (2005) 640e647.
[14] K.A. Dill, T.M. Truskett, V. Vlachy, B. Hribar-Lee, Modeling water, the hydro-
phobic effect, and ion solvation, Annu. Rev. Biophys. Biomol. Struct. 34 (2005)
173e199.
[15] J.M. Myslinski, J.E. DeLorbe, J.H. Clements, S.F. Martin, Protein-ligand in-
teractions: thermodynamic effects associated with increasing nonpolar sur-
face area, J. Am. Chem. Soc. 133 (2011) 18518e18521.
Atomic coordinates have been deposited in the Protein Data
Bank (https://www.rcsb.org) for the structures of the Grb2 SH2
domain complexed with 5 (PDB ID: 6WM1)and 11 (PDB ID: 6WO2).
Authors will release the atomic coordinates and experimental data
upon article publication.
[16] R. Malham, S. Johnstone, R.J. Bingham, E. Barratt, S.E.V. Phillips, C.A. Laughton,
S.W. Homans, Strong solute-solute dispersive interactions in a protein-ligand
complex, J. Am. Chem. Soc. 127 (2005) 17061e17067.
Declaration of competing interest
[17] C. Carey, Y.-K. Cheng, P. Rossky, Hydration structure of the a-chymotrypsin
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
substrate binding pocket: the impact of constrained geometry, J. Chem. Phys.
258 (2000) 415e425.
[18] T.S.G. Olsson, M.A. Williams, W.R. Pitt, J.E. Ladbury, The thermodynamics of
protein-ligand interaction and solvation: insights for ligand design, J. Mol.
Biol. 384 (2008) 1002e1017.
[19] T.S.G. Olsson, J.E. Ladbury, W.R. Pitt, M.A. Williams, Extent of enthalpy-
entropy compensation in protein-ligand interactions, Protein Sci. 20 (2011)
1607e1618.
Acknowledgments
We thank the National Institutes of Health (GM 84965), National
Science Foundation (CHE 0750329), Robert A. Welch Foundation (F-
0652), Norman Hackerman Advanced Research Program, and Texas
Institute for Drug and Diagnostic Development through the Welch
Foundation Grant #HeF-0032 for support of this research. We
thank Ms. Andrea Beckham for experimental and technical assis-
tance. Instrumentation and technical assistance for this work were
provided by the Macromolecular Crystallography Facility, with
financial support from the College of Natural Sciences, the Office of
the Executive Vice President and Provost, and the Institute for
Cellular and Molecular Biology at the University of Texas at Austin.
[20] J.D. Chodera, D.L. Mobley, Entropy-Enthalpy Compensation: Role and Rami-
fications in Biomolecular Ligand Recognition and Design. Annu. Rev. Biophys
42 (2013) 121e142.
[21] C.H. Reynolds, M.K. Holloway, Thermodynamics of ligand binding and effi-
ciency, ACS Med. Chem. Lett. 2 (2011) 433e437.
[22] K. Haider, A. Cruz, S. Ramsey, M.K. Gilson, T. Kurtzman, Solvation Structure
and Thermodynamic Mapping (SSTMap): An open-source, flexible package for
the analysis of water in molecular dynamics trajectories, J. Chem. Theor.
Comput. 14 (2018) 418e425.
[23] For some leading references, see: (a) S.-R. Tzeng, C.G. Kalodimos, Protein ac-
tivity regulation by conformational entropy, Nature 488 (2012) 236e240;
(b) J.A. Caro, K.W. Harpole, V. Kasinath, J. Lim, J. Granja, K.G. Valentine,
K.A. Sharp, A.J. Wand, Entropy in molecular recognition by proteins, Proc. Nat.
Acad. Sci USA 114 (2017) 6563e6568;
€
(c) M.L. Verteramo, O. Stenstrom, M.M. Ignjatovic, O. Caldararu, M.A. Olsson,
F. Manzoni, H. Leffler, E. Oksanen, D.T. Logan, U.J. Nilsson, U. Ryde, M. Akke,
Interplay between conformational entropy and solvation entropy in
ProteinꢁLigand binding, J. Am. Chem. Soc. 141 (2019) 2012e2026.
[24] (a) L. Zidek, M.V. Novotny, M.J. Stone, Increased protein backbone confor-
mational entropy upon hydrophobic ligand binding, Nat. Struct. Biol. 6 (1999)
1118e1121;
Appendix A. Supplementary data
Supplementary data and Supporting Information to this article
can be found online at https://doi.org/10.1016/j.ejmech.2020.
112771.
(b) R.J. Bingham, J.H.B.C. Findlay, S.-Y. Hsieh, A.P. Kalverda, A. Kjellberg,
C. Perazzolo, S.E.V. Phillips, K. Seshadri, C.H. Trinh, W.B. Turnbull,
G. Bodenhausen, S.W. Homans, Thermodynamics of binding of 2-Methoxy-3-
isopropylpyrazine and 2-Methoxy-3-isobutylpyrazine to the major urinary
protein, J. Am. Chem. Soc. 126 (2004) 1675e1681;
References
(c) T.A. Pertinhez, E. Ferari, E. Casali, J.A. Patel, A. Spisni, L.J. Smith, The binding
cavity of mouse major urinary protein is optimised for a variety of ligand
binding modes, Biochem. Biophys. Res. Commun. 390 (2009) 1266e1271.
[25] For a review of SH2 domains, see J.M. Bradshaw, G. Waksman, Molecular
recognition by SH2 domains adv, Protein Chem. 61 (2002) 161e210.
[26] J.E. DeLorbe, J.H. Clements, B.B. Whiddon, S.F. Martin, Thermodynamic and
structural effects of macrocyclic constraints in protein-ligand interactions,
ACS Med. Chem. Lett. 1 (2010) 448e452.
[1] S.F. Martin, J.H. Clements, Correlating structure and energetics in protein-
ligand interactions: paradigms and paradoxes, Annu. Rev. Biochem. 82
(2013) 267e293.
[2] G. Klebe, Applying thermodynamic profiling in lead finding and optimization,
Nat. Rev. Drug Discov. 14 (2015) 95e110.
[3] G. Klebe, Broad-scale analysis of thermodynamic signatures in medicinal
chemistry: are enthalpy-favored binders the better development option?
Drug Discov. Today 24 (2019) 943e948.
[27] J.M. Myslinski, J.H. Clements, J.E. DeLorbe, S.F. Martin, Protein-ligand in-
teractions: thermodynamic effects associated with increasing the length of an
alkyl chain, ACS Med. Chem. Lett. 4 (2013) 1048e1053.
[4] T.V. Chalikian, Hydrophobic tendencies of polar groups as a major force in
molecular recognition, Biopolymers 70 (2003) 492e496.
[5] For a lead reference, see A.R. Khan, J.C. Parrish, M.E. Fraser, W.W. Smith,
P.A. Bartlett, M.N.G. James, Lowering the entropic barrier for binding con-
formationally flexible inhibitors to enzymes, Biochemistry 37 (1998)
16839e16845.
[28] J.M. Myslinski, J.H. Clements, S.F. Martin, Protein-ligand interactions: probing
the energetics of a putative cation-pi interaction, Bioorg. Med. Chem. Lett 24
(2014) 3164e3167.
́
[29] J.M. Sturtevant, Heat Capacity and Entropy Changes in processes involving
[6] E.H. Rühmann, M. Rupp, M. Betz, A. Heine, G. Klebe, Boosting affinity by

14
D.L. Cramer et al. / European Journal of Medicinal Chemistry 208 (2020) 112771
proteins, Proc. Natl. Acad. Sci. U.S.A. 74 (1977) 2236e2240.
sarcosine chains with bovine carbonic anhydrase II: an unexpected manifes-
tation of enthalpy/entropy compensation, J. Am. Chem. Soc. 128 (2006)
5802e5812.
[30] V.V. Loladze, D.N. Ermolenko, G.I. Makhatadze, Heat capacity changes upon
burial of polar and nonpolar groups in proteins, Protein Sci. 10 (2001)
1343e1352.
[31] R. Talhout, A. Villa, A.E. Mark, J.B.F.N. Engberts, Understanding binding Af-
finity: A combined isothermal titration calorimetry/molecular dynamics study
of the binding of a series of hydrophobically modified benzamidinium chlo-
ride inhibitors to trypsin, J. Am. Chem. Soc. 125 (2003) 10570e10579.
[32] S.D. Sharrow, M.V. Novotny, M.J. Stone, Thermodynamics of binding between
mouse major urinary protein-I and the pheromone 2-sec-Butyl-4,5-
dihydrothiazole, Biochemistry 42 (2003) 6302e6309.
[33] J. Mecinovic, P.A. Snyder, K.A. Mirica, S. Bai, E.T. Mack, R.L. Kwant,
D.T. Moustakas, A. Heroux, G.M. Whitesides, Fluoroalkyl and alkyl chains have
similar hydrophobicities in binding to the "hydrophobic wall" of carbonic
anhydrase, J. Am. Chem. Soc. 133 (2011) 14017e14026.
[34] S.G. Krimmer, J. Cramer, M. Betz, V. Fridh, R. Karlsson, A. Heine, G. Klebe,
Rational design of thermodynamic and kinetic binding profiles by optimizing
surface water networks coating protein-bound ligands, J. Med. Chem. 59
(2016) 10530e10548.
[39] A.V. Plyasunov, E.L. Shock, Thermodynamic functions of hydration of hydro-
carbons at 298.15 K and 0.1 MPa, Geochem. Cosmochim. Acta 64 (2000)
439e468.
[40] J.M. Sturtevant, Heat capacity and entropy changes in processes involving
proteins, Proc. Natl. Acad. Sci. U.S.A. 74 (1977) 2236e2240.
[41] J.D. Dunitz, The entropic cost of bound water in crystals and biomolecules.",
Science 264 (1994) 670.
[42] C. Gerlach, M. Smolinski, H. Steuber, C.A. Sotriffer, A. Heine, D.G. Hangauer,
G. Klebe, “Thermodynamic inhibition profile of a cyclopentyl and a cyclohexyl
derivative toward thombin: the same but for different reasons” angew. Chem.
Int. Ed 46 (2007) 8511e8514.
[43] T. Brandt, N. Holzmann, L. Muley, M. Khayat, C. Wegscheid-Gerlach, B. Baum,
A. Heine, D. Hangauer, G. Klebe, Congeneric but still distinct: how closely
related trypsin ligands exhibit different thermodynamic and structural
properties, J. Mol. Biol. 405 (2011) 1170e1187.
[44] A.P. Benfield, B.B. Whiddon, J.H. Clements, S.F. Martin, Structural and energetic
aspects of grb2-SH2 domain swapping, Arch. Biochem. Biophys. 462 (2007)
47e53.
[45] Instructions 52-1782-00 AG Suprose 6 prep grade, Superose 12 prep grade. GE
Healthcare (2006).
[35] H. Fretz, P. Furet, C. Garcia-Echeverria, J. Rahuel, J. Schoepfer, Structure-based
design of compounds inhibiting Grb2-SH2 mediated protein-protein in-
teractions in signal transduction pathways, Curr. Pharmaceut. Des. 6 (2000)
1777e1796.
[36] P. Furet, B. Gay, G. Caravatti, C. García-Echeverría, J. Rahuel, J. Schoepfer,
H. Fretz, Structure-based design and synthesis of high affinity tripeptide li-
gands of the Grb2-SH2 domain, J. Med. Chem. 41 (1998) 3442.
[37] C. Garcia-Echeverria, B. Gay, J. Rahuel, P. Furet, Mapping the Xþ1 binding site
[46] Z. Otwinowski, W. Minor, Methods Enzymol. 276 (1997) 307e326.
[47] Collaborative Computational Project 4, Acta Crystallogr. D50 (1994) 760e763.
[48] P. Emsley, B. Lohkamp, W. Scott, K. Cowtan, Acta Crystallogr. D66 (2010)
486e501.
[49] A.W. Schuettelkopf, D.M.F. van Aalten, Acta Crystallogr. D60 (2004)
1355e1363.
of the Grb2 SH2 domain with
a
,
a
-disubstituted cyclic amino acids, Bioorg.
Med. Chem. Lett 9 (1999) 2915e2920.
[38] V.M. Krishnamurthy, B.R. Bohall, V. Semetey, G.M. Whitesides, The paradox-
ical thermodynamic basis for the interaction of ethylene glycol, Glycine, and
[50] F.M. Richards, Annu. Rev. Biophys. Bioeng. 6 (1977) 151e176.
[51] MacroModel, Schrodinger, LLC, New York, 2007. NY, version 9.1.
€