Some Thermodynamic Effects of Varying Nonpolar Surfaces in Protein-Ligand Interactions.

Abstract

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.

Graphical Abstract

Introduction and Background

Optimizing binding affinities of small molecules for protein targets is an early goal in drug discovery. Generally, binding affinities, ΔG°, of small molecules for proteins can be increased by optimizing nonbonded, bimolecular interactions to enhance binding enthalpy, ΔH°, and by making binding entropy, ΔS°, more positive by reducing unfavorable solvation and conformational parameters.^1–3 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 energetically unfavorable.⁴ Nondirectional van der Waals (vdW) contacts are easier to enhance, but dispersion forces are less specific and weaker relative to polar interactions.

Toward improving binding entropies, conformational constraints are often introduced into flexible molecules.^5–7 However, 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 enhancing binding entropy is based upon the hydrophobic effect and involves adding nonpolar groups to a lead because the desolvation of well-ordered water molecules from a nonpolar surface is typically viewed as being entropically favorable.^12–14 However, this view is not uniformly true as we¹⁵ and others^16,17have 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.

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 difficult to deconvolute the overall energetics into individual contributions. One productive strategy toward addressing this complex problem involves performing systematic studies of structure, dynamics, and energetics for a series of closely related protein-ligand interactions. However, even when such data are available, correlating variations in thermodynamic parameters with differences in structural features is often problematic because of enthalpy/entropy compensation^19,20 and because changes in ΔH° and −TΔS° often do not correspond to changes in ΔG°.²¹ Exacerbating the difficulty is that contributions of solvent reorganization²² and protein dynamics²³ to binding free energies are poorly understood. For example, recent studies of changes in protein conformational 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 correlating 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 receptor 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.

We conducted a number of studies using different sets of pseudopeptide analogs of Ac–pTyr–Val–Asn–NH₂, the tripeptide consensus sequence for potent Grb2 SH2 binders.^8,15,26–28 Relevant to the investigation of hydrophobic effects herein, we used isothermal titration calorimetry (ITC) and x-ray crystallography to probe the consequences of replacing the pTyr+1 Val residue with cyclic, α,α-disubstituted surrogates having different ring sizes.¹⁵ The incrementally more favorable binding affinities that accompanied increasing ring size arose from favorable 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 pTyr+1 binding pocket correlated with the observed favorable changes in ΔH°. However, there was no correlation of buried nonpolar surface area with ΔC_p as might have been expected.^29,30 This trend toward favorable changes in ΔG° and ΔH° was not observed in a related study where we examined the effects of increasing the length of an alkyl group at the pTyr+1 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 binding signatures in several other systems.^31–34

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 Ac–pTyr–Val–Asn–NH₂ in which hydrophobic groups were added at the pTyr+3 subsite of 1 (Figure 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 hydrophobic 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.

Figure 1.

Analogs of Ac-pTyr–Val–Asn–NH₂ with Ac₆c at pTyr+1 and nonpolar groups at pTyr+3 with variable regions shown in red.

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 pTyr+3 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 pTyr+1 of the prototypical tripeptide Ac-pTyr–Val–Asn–NH₂ with an α,α-cyclohexane amino acid to give the known parent tripeptide 2.^15,37 Having made this substitution, two variables remained: 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 3–7 via routine peptide couplings of N-substituted asparagine amides, which were prepared via amide forming reactions, with a protected Ac-pTyr–Ac₆c–OH dipeptide in accord with the bond disconnections shown in Figure 2. The binding energetics for each of these ligands 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, ΔG°, for 3–7 vary over a range of only 0.5 kcal•mol⁻¹, the binding enthalpy, ΔH°, is 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 8–10 to determine whether a linking chain having fewer methylene groups might be needed for larger 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 experiments were not designed to optimize the linker length for all possible R groups, but they enabled us to select a three-carbon linker as a standard spacer.

Figure 2.

Determining standard linker length, n, for pTyr+3 substituted analogs of Ac-pTyr–Ac₆c–Asn–NH(CH₂)_n–R.

Table 1.

ITC data for Ac-pTyr–Ac₆c–Asn–NH(CH₂)_n–R analogs wherein linker length, n, is varied.^{[a], [b]}

Ligand	Ka (× 107 M−1)	ΔG° (kcal•mol−1)	ΔH° (kcal•mol−1)	−TΔS° (kcal•mol−1)
28	0.7 ± 0.1	−9.3 ± 0.1	−8.5 ± 0.4	−0.8 ± 0.4
3	1.2 ± 0.1	−9.6 ± 0.1	−7.5 ± 0.3	−2.1 ± 0.8
4	1.5 ± 0.2	−9.8 ± 0.1	−7.7 ± 0.1	−2.1 ± 0.3
5	2.6 ± 0.7	−10.1 ± 0.2	−8.8 ± 0.1	−1.3 ± 0.8
6	1.5 ± 0.1	−9.8 ± 0.1	−7.7 ± 0.2	−2.1 ± 0.2
7	2.8 ± 0.1	−10.2 ± 0.1	−7.9 ± 0.1	−2.3 ± 0.1
8	1.1 ± 0.1	−95 ± 0.1	−6.7 ± 0.1	−2.8 ± 0.1
9	2.4 ± 0.1	−10.1 ± 0.1	−8.1 ± 0.2	−2.0 ± 0.1
10	4.9 ± 1.3	−10.4 ± 0.2	−10.2 ± 0.3	−0.2 ± 0.7

^[a]ITC experiments were conducted at 25 °C 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

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 11–19 (Figure 3) were prepared in analogy with the syntheses of 3–10. The binding energetics for each of these compounds with the Grb2 SH2 domain was determined by ITC (Table 2) as before. The Ac-pTyr-Ac₆c–Asn–NH(CH₂)₃–R analogs 11–13 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, −TΔS°, that overcome slightly less favorable binding enthalpies, ΔH°, which are identical within experimental error.

Figure 3.

Varying nonpolar groups R of pTyr+3 substituted analogs of Ac-pTyr–Ac₆c–Asn–NH(CH₂)₃–R.

Table 2.

ITC data for Ac-pTyr–Ac₆c–Asn–NH(CH₂)₃–R analogs.^[a]–[c]

Ligand	Ka (× 107 M−1)	ΔG° (kcal•mol−1)	ΔH° (kcal•mol−1)	−TΔS° (kcal•mol−1)	ΔCp (cal•mol−1•K−1)
2	0.7 ± 0.1	−9.3 ± 0.1	−8.5 ± 0.4	−0.8 ± 0.4	−179 ± 10
11	1.6 ± 0.1	−9.8 ± 0.1	−7.9 ± 0.2	−1.9 ± 0.2	–
12	1.7 ± 0.3	−9.8 ± 0.1	−7.8 ± 0.1	−2.0 ± 0.4	–
13	1.9 ± 0.3	−9.9 ± 0.1	−8.0 ± 0.4	−1.9 ± 0.4	−215 ± 8
5	2.6 ± 0.7	−10.1 ± 0.2	−8.8 ± 0.1	−1.3 ± 0.8	−208 ± 12
14	4.0 ± 0.8	−10.3 ± 0.1	−8.6 ± 0.5	−1.7 ± 0.6	–
15	5.0 ± 0.8	−10.5 ± 0.1	−9.1 ± 0.8	−1.4 ± 0.4	–
16	4.7 ± 0.4	−10.5 ± 0.1	−9.1 ± 0.5	−1.4 ± 0.2	–
17	6.0 ± 0.3	−10.6 ± 0.1	−9.6 ± 0.1	−1.0 ± 0.1	−224 ± 8
10	4.9 ± 1.3	−10.4 ± 0.2	−10.2 ± 0.3	−0.2± 0.7	–
18	3.5 ± 0.3	−10.3 ± 0.1	−9.1 ± 0.1	−1.2 ± 0.2	–
19	2.7 ± 0.5	−10.1 ± 0.1	−9.1 ± 0.2	−1.0 ± 0.6	−217 ± 10

^[a]ITC experiments were conducted at 25 °C 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

^[c]Values for ΔC_p were determined from slopes of plots of values of ΔH° at 15, 25, and 35 °C, and the error was determined from the standard error in the slope.

Examination of the binding parameters for compounds 5 and 14–16, 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 14–16 with respect to 2 results from a combination of more favorable binding entropies and enthalpies although the enthalpic contributions relative to 2 are smaller. It is difficult to assess the importance of the variations in ΔH° 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 ΔH° values of congeneric ligands related to Ac-pTyr–Val–Asn–NH₂.⁸ Comparing 5 with 14–16 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, 17–19), 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 pTyr+3 site of the parent tripeptide 2 clearly increases binding affinity. Overall, the addition of an aryl R group at the pTyr+3 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 pTyr+3 improves binding free energies compared with 2 because more favorable binding entropies dominate binding enthalpies that become less favorable. When the substituent at the pTyr+3 site contains an aromatic ring, the increased affinities relative to 2 generally arise from a combination of more favorable binding entropies and enthalpies. The trend of more favorable binding entropies for the alkylated analogs 11–13 relative those bearing an aromatic ring (e.g., 5, 10 and 14–19) is consistent with the known tendency of alkyl groups to have more favorable dehydration entropies than aryl groups.³⁹

To assess whether differences in desolvation might play a role in the binding energetics, ΔC_p values, which have been purported to be a barometer of hydrophobic effects in protein-ligand interactions,⁴⁰ for several representative ligands were determined (Table 2). The more negative ΔC_p values observed for 13, 5, 17, and 19 relative to the parent tripeptide 2 are consistent with the burial of more 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 nature 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 determined 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 (Figure 4a). Indeed, the BC loop is positioned ~1.5 Å further away from the phosphate group with which it interacts in one of the complexes. As in complexes of other Grb2 SH2 ligands, 5 adopts a β-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 pTyr+3 site (Figure 4b). As suggested by earlier modeling studies,³⁵ the phenyl group at the pTyr+3 site packs against the side chains of Lys109 and Leu111, which are oriented similarly in the two complexes (Figure 4c,d). Following alignment of the domain, the pTyr–pTyr+1–pTyr+2 portion of the bound ligands compare with an RMSD of 0.9 Å, reflecting some differences in the relative positions of the Ac–pTyr and pTyr+1 residues, but the greatest variations in atomic positions of ligand atoms are at the pTyr+3 site (RMSD = 1.6 Å) (Figure 4b).

Figure 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. Carbon 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 pTyr+3 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 pTyr+3 group with Lys109 and Leu111 (green) in complex a (c) and complex b (d).

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 = 2.2 Å), which does not interact with 11 (Figure 5a). As observed in the complexes of 5 with the domain, 11 in each complex binds in a characteristic β-turn conformation that is stabilized by an internal hydrogen bond between the pTyr carbonyl oxygen atom and the amide nitrogen atom of the pTyr+3 site (Figure 5b). Following alignment of the domain, the Ac–pTyr–pTyr+2 segments of the ligands overlay with an RMSD of 1.2 Å, reflecting differences in relative positions of backbone atoms, the pTyr–1 acetyl group, the phenyl ring and the bridging oxygen atoms of the phosphate group of the pTyr side chain, and the pTyr+1 side-chain. As observed for the complexes of 5, the greatest variations in the atomic positions in the two complexes are in the pTyr+3 side chain with the isopropyl groups occupying distinctly different spatial orientations (Figure 5b). The respective positions of the alkyl group at pTyr+3 of 11 and the side chains of Lys109 and Leu111 in the two complexes are also markedly different (Figure 5c,d). Moreover, the orientation of the Lys109 side chain in complex a of 11 (Figure 5c) is notably distinct from those observed in the two complexes of 5 (cf Figure 4c,d).

Figure 5.

Complexes in the asymmetric unit of the Grb2 SH2 domain and 11 (PDB ID: 6WO2); heteroatom labels same as in Figure 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 Figure 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 pTyr+3 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 pTyr+3 group with Lys109 and Leu111 (green) in complex a (c) and complex b (d).

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 (Figure 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 pTyr–1 carbonyl oxygen atom as in the b complex (see Supporting Information). Lacking this interaction with Arg67, the pTyr–1 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 (Figure 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 contacts 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 pTyr+2 residue and the backbone carbonyl atom of Leu120 (see Supporting Information). We inventoried the van der Waals (vdW) contacts in these complexes, and the only difference involves two vdW contacts between the cyclohexane ring at the pTyr+1 site of 5 and the domain (Table 3).

Table 3.

Thermodynamic and protein-ligand contact data for the two complexes in the asymmetric units of 5 and 11.

Ligand	ΔG°	ΔH°	−TΔS°	Polar Direct Contacts[a]	Water-Mediated Contacts	vdW contacts[b] to pTyr+1	vdW contacts[b] to pTyr+3	ΔCSAnp (Å2)
	(kcal•mol−1)
5	−10.2	−8.8	−1.4	11	4	12	8	250
				14	2	14	8	197
11	−99	−7.8	−2.1	13	1	18	3	218
				16	1	17	2	215

^[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 2.5–3.5 Å (see Supporting Information).

^[b]A hydrophobic vdW contact exists when the interatomic distance between a methylene group in the pTyr+1 residue and a carbon, nitrogen, or oxygen atom in the domain is in the range of 3.4–4.2 Å.

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 an atypical spatial orientation of the Glu87 side-chain, which is turned toward the phosphate group of 11 in the b complex, 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 Information). 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 pTyr+1 residue and the other involving the pTyr+3 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 differences 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, however, 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 pTyr+1 and pTyr+3 sites for 5 and 11. Although the differences in the number of vdW contacts at pTyr+1 are subtle, those at pTyr+3 are readily apparent by comparing the orientations and interactions of the nonpolar R groups at the pTry+3 sites with Lys109 and Leu111 in the respective complexes (cf Figures 4c,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.

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 structure of 5 with the Grb2 SH2 domain. Fixing a water molecule at a protein-ligand interface will have an unfavorable entropic consequence.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 consistent with the experimental data (Table 3). Notably, however, we have also shown in another study of complexes of similar pseudopeptide 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 orientations of the isopropyl group in the a and b complexes of 11 (see Figure 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 pTyr+3 side chain in the complexes of 11 relative to those in complexes of 5 is also reflected in the respective 2f_o-f_c electron density maps (see Supporting Information).

Another significant energetic component of protein-ligand associations 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 ΔCSA_np that occur upon binding for 5 and 11 (Table 3). Because there is no crystal structure of the monomeric Grb2 SH2 domain without a bound ligand,¹⁵⁴⁴ the calculated values for ΔCSA_np are based solely upon structures of the complexes, thereby leading some uncertainties. Further complicating the analysis, there are notable dissimilarities in the values of ΔCSA_np that are calculated for the individual complexes of 5 in the crystallographic asymmetric unit. These caveats notwithstanding, there are good correlations between ΔCSA_np and ΔH° and ΔG°, although none with either ΔC_p or −TΔS°. Namely, the dependence of ΔH° on calculated average values of ΔCSA_np for 5 and 11 is −38 ± 2 cal•mol⁻¹Å⁻², whereas the dependence of ΔG° on average values of ΔCSA_np corresponds to a contribution of −46 ± 0.3 cal•mol⁻¹Å⁻². It is noteworthy that this later value is significantly larger than the value of −12 cal•mol⁻¹Å⁻² that is obtained from the analysis of a large number of protein-ligand complexes.¹⁸ However, because crystallographic data were only obtained for 5 and 11, these average values 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 20 was coupled with Cbz-Tyr-OH using 1–(3–dimethylamino)propyl]-3-ethylcarbodiimide•HCl (EDCI) and 1–hydroxybenzotriazole (HOBt) in the presence of N-methylmorpholine (NMM) in DMF to give 21 in 73% yield. Global removal of the protecting groups by catalytic hydrogenolysis, followed by N-acetylation of the intermediate amino acid 22 provided 23 in 91% overall yield for the two steps. Phosphorylation of the tyrosine hydroxyl group then gave the dipeptide fragment 24 in 70% yield. Subsequent coupling of 24 with the N-substituted asparagines 25–41, 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 3–19 in generally good overall yields. These compounds were purified by reverse phase HPLC prior to the ITC studies.

Scheme 1.

Synthesis of Ac-pTyr–Ac₆c–Asn–NH(CH₂)_n–R analogs 3–19.

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 pTyr+3 position. The first set of compounds 2–10 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 experiments, we selected a three-methylene linker as the standard spacer and prepared another set of compounds 11–19 (i.e., 1, n = 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 11–19 differed by 0.8 kcal•mol⁻¹, the binding enthalpies varied over a range of 2.4 kcal•mol⁻¹, 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 contacts 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, ΔΔH°. Likewise, even though there are notable differences in the number of vdW contacts at the pTyr+1 and pTyr+3 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 ΔΔH°. 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 pTyr+3 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 the domain. Differences in ΔCSA_np for 5 and 11 correlate with the measured values of ΔG° and ΔH° but not with ΔC_p or −TΔS°.

Collectively, these studies show that adding nonpolar groups to the pTyr+3 position of 2 leads to analogs with higher binding affinities, 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 enthalpies. 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 the electrostatic nature and size of the R group. Detailed experimental investigations such as these in combination with computational 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.

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 25–30 °C (bath temperature). Flash chromatography was performed with the indicated solvents and Merck 250–400 mesh silica gel. Purification by reverse phase high performance liquid chromatography (RP HPLC) was conducted using a binary solvent system, 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 × 250 mm (flow rate of 8 mL/min), being used for preparative work and a C18 column (10 mm particle size, 300 Å pore size), 4.6 mm diameter × 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 million (ppm, δ) 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 = singlet; d = doublet; dd = doublet of doublet; ddd = doublet of doublet of doublets; t = triplet; q = quartet; p = pentuplet; hep = heptet; m = multiplet; comp = overlapping multiplets of magnetically nonequivalent protons; br = broad; app = 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 Cbz–Tyr (630 mg, 2.0 mmol) and benzyl 1-aminocyclohexanecarboxylate (20) (513 mg, 2.2 mmol) in DMF (10 mL) at −20 °C 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 °C 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 chromatography (hexanes-EtOAc 1:1) to yield 770 mg (73%) of 21 as an amorphous white solid. ¹H NMR (400 MHz, CDCl₃) δ 7.38–7.27 (comp, 10 H), 7.00 (d, J = 7.6 Hz, 2 H), 6.70 (d, J = 7.6 Hz, 2 H), 6.20 (br s, 1 H), 6.05 (br s, 1H), 5.41 (br s, 1 H), 5.16–5.03 (comp, 4 H), 4.33 (d, J = 6.4 Hz, 1 H), 2.95 (dd, J = 13.6, 5.6 Hz, 1 H), 2.85 (dd, J = 13.6, 5.6 Hz, 1 H), 1.90 (d, J = 13.2 Hz, 2 H), 1.76 (t, J = 11.6 Hz, 2 H), 1.58–1.42 (comp, 3 H), 1.28–1.05 (comp, 3 H); ¹³C NMR (100 MHz, CDCl₃) δ 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+H) requires 531.2495].

H₂N-Tyr-Ac₆c-OH (22).

The benzyl carbamate 21 (625 mg; 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) δ 7.13 (d, J = 8.4 Hz, 2 H), 6.76 (d, J = 8.4 Hz, 2 H), 4.04 (t, J = 7.6 Hz, 1 H), 3.05 (dd, J = 13.8, 8.2 Hz, 1 H), 2.96 (dd, J = 13.8, 8.2 Hz, 1 H), 2.23 (d, J = 13.6 Hz, 1 H), 1.85–1.76 (comp, 2 H), 1.67–1.63 (m, 1 H), 1.52–1.12 (comp, 5 H), 1.28–1.05 (q, J = 12.8 Hz, 1 H); ¹³C NMR (100 MHz, CD₃OD) δ 181.3, 169.6, 158.0, 131,6, 126.9,116.7, 62.9, 56.2, 38.1, 35.4, 31.7, 26.6, 22.8, 22.7; HRMS (ESI) m/z 307.1652 [C₁₆H₂₃N₂O₄ (M+H)+ requires 307.1657].

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 dryness 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) δ 7.08 (d, J = 8.4 Hz, 2 H), 6.79 (d, J = 8.4 Hz, 2 H), 4.64 (dd, J = 8.6, 6.2 Hz, 1 H), 3.00 (dd, J = 13.8, 6.2 Hz, 1 H), 2.76 (dd, J = 13.8, 6.2 Hz, 1 H), 2.08–1.97 (comp, 2 H), 1.90 (s, 3 H), 1.76 (td, J = 12.8, 3.6 Hz, 2 H), 1.61–1.21 (comp, 6 H); ¹³C NMR (100 MHz, CD₃OD) δ 177.8, 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+H) requires 349.1763].

(S)-1-{2-Acetamido-3-[4-{bis(benzyloxy)phosphoryloxy}phenyl]propanamid o}cyclohexane carboxylic acid (24).

NMM (42 μL, 0.38 mmol) and TBSCl (57.1 mg, 0.38 mmol) were added 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 temperature, the reaction was cooled to 0 °C, 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 °C for 1 h and then 14.5 h at room temperature. The reaction was cooled to 0 °C, and 6 M tert-butyl hydroperoxide in decane (1 mL) was added. After stirring for 30 min at 0 °C and 4 h at room temperature, the reaction was cooled to 0 °C, and 10% aqueous NaHSO₃ (5 mL) was added. The solution was stirred at 0 °C 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 × 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 (CH₂Cl₂-MeOH 10:1) to afford 160 mg (70%) 24 as an amorphous white solid. ¹H NMR (400 MHz, CDCl₃) δ 7.69 (br s, 1 H), 7.34–7.27 (comp, 10 H), 7.18 (d, J = 8.4 Hz, 2 H), 7.06 (d, J = 8.4 Hz, 2 H), 5.08 (app d, J = 6.4 Hz, 4 H), 4.83 (dd, J = 14.4, 6.8 Hz, 1 H), 3.11–3.97 (comp, 2 H), 1.98–1.90 (comp, 2 H), 1.93 (s, 3 H), 1.77–1.65 (comp, 2 H), 1.56–1.02 (comp, 6 H); ¹³C NMR (100 MHz, CDCl₃) δ 176.7, 171.5, 171.4, 149.3, 149.2, 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+H) requires 607.2215].

General procedure for the synthesis of aspargine amides.

N-Methylmorpholine (NMM) (0.22 mL, 2.0 mmol), 1–hydroxybenzotriazole hydrate (HOBt·H₂O) (200 mg, 1.3 mmol) and 1–(3–dimethylamino)propyl]-3-ethylcarbodiimide•HCl (EDCI) (150 mg, 0.78 mmol) were added sequentially to a solution of Cbz-Asn–H (190 mg, 0.71 mmol) and the appropriate amine, RNH₂, (0.65 mmol) in DMF (5 mL) at −30 °C. The mixture was allowed to warm slowly to room temperature and then stirred for 16 h. The reaction mixture was concentrated under reduced pressure, 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 Cbz-NH-Asn-NHR, which was purified when necessary by silica gel chromatography using a suitable mixture of hexanes/EtOAc. A solution of the crude Cbz-NH-Asn-NHR in MeOH (40 mL) containing 10% Pd/C (40 mg) was stirred for 6–12 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.

H-Asn-NH-CH₂Ph (25).

Prepared in 64% yield as a white solid; mp 143–144 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.31–7.29 (comp, 3 H), 7.25–7.21 (comp, 2 H), 3.70 (dd, J = 8.0, 5.1 Hz, 1 H), 3.31–3.30 (comp, 2 H), 2.61 (dd, J = 15.5, 8.0 Hz, 1 H), 2.47 (dd, J = 15.5, 8.0 Hz, 1 H), 1.82 (p, J = 7.5 Hz, 2 H); ¹³C NMR (150 MHz) δ 175.1, 174.7, 138.6, 128.4, 127.4, 127.0, 52.2, 42.9, 39.9; HRMS (ESI) m/z 222.1237; C₁₁H₁₆N₃O₂ (M+H) requires 222.1240.

H-Asn-NH-(CH₂)₂Ph (26).

Prepared in 94% yield as a white solid; mp 143–144 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.30–7.26 (comp, 3 H), 7.24–7.18 (comp, 2 H), 3.68 (dd, J = 8.0, 5.1 Hz, 1 H), 3.46–3.40 (comp, 2 H), 2.61 (dd, J = 15.4, 8.0 Hz, 1 H), 2.47 (dd, J = 15.2, 8.0 Hz, 1 H); ¹³C NMR (150 MHz) δ 175.1, 174.5, 139.3, 128.7, 128.4, 126.2, 52.0, 40.8, 39.5, 35.3; HRMS (ESI) m/z 236.1394; C₁₂H₁₈N₃O₂ (M+H) requires 236.1394.

H-Asn-NH-(CH₂)₃Ph (27).

Prepared in 24% yield as a white solid; mp 170–172 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.27–7.14 (comp, 5 H), 3.64 (dd, J = 8.0, 5.5 Hz, 1 H), 3.23 (dd, J = 10.0, 5.5 Hz, 2 H), 2.66–2.58 (comp, 3 H), 2.46 (dd, J = 15.0, 8.0 Hz, 1 H), 1.82 (p, J = 7.52 Hz, 2 H); ¹³C NMR (150 MHz) δ 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; C₁₃H₂₀N₃O₂ (M+H) requires 250.1304.

H-Asn-NH-(CH₂)₄Ph (28).

Prepared in 50% yield as a white solid; mp 119–121 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.26–7.13 (comp, 5 H), 3.62 (dd, J = 8.0, 5.5 Hz, 1 H), 3.23–3.19 (comp, 2 H), 2.66–2.58 (comp, 3 H), 2.46 (dd, J = 15.0, 8.0 Hz, 1 H), 1.82 (p, J = 7.5 Hz, 2 H), 1.56 (p, J = 7.5 Hz, 2 H); ¹³C NMR (150 MHz) δ 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+Na) requires 286.1531.

H-Asn-NH-(CH₂)₅Ph (29)

Prepared in 70% yield as a glass. ¹H NMR (400 MHz, CD₃OD) δ 7.26–7.11 (comp, 5 H), 3.63 (dd, J = 8.0, 5.2 Hz, 1 H), 3.22 (dd, J = 14.2, 6.8 Hz, 1 H), 3.15 (dd, J = 14.2, 7.2 Hz, 1 H), 2.63–2.57 (comp, 3 H), 2.43 (dd, J = 15.2, 8.0 Hz, 1 H), 1.63 (p, J = 7.6, 2 H), 1.54 (p, J = 7.2 Hz, 2 H), 1.40–1.32 (comp, 2 H); ¹³C NMR (100 MHz, CD₃OD) δ 176.3, 175.9, 143.8, 129.4 (2C), 129.3 (2C), 126.7, 53.4, 41.3, 40.4, 36.8, 32.4, 30.3, 27.5; HRMS (ESI) m/z 278.1863; C₁₅H₂₄N₃O₂ (M+H)⁺ 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) δ 7.58 (d, J = 8.0 Hz, 1 H), 7.30 (d, J = 8.4 Hz, 1 H), 7.16 (app td, J = 7.6, 1.0 Hz, 1 H), 7.10 (s, 1 H), 7.04 (app td, J = 7.6, 1.0 Hz, 1 H), 4.52 (dd, J = 19.2, 14.8 Hz, 2 H), 3.70 (s, 3 H), 3.64 (dd, J = 8.2, 5 Hz, 1 H), 2.61 (dd, J = 15.2, 4.8 Hz, 1 H), 2.42 (dd, J = 15.2, 8.4 Hz, 1 H); ¹³C NMR (100 MHz, CD₃OD) δ 175.9 (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+H)⁺ requires 275.1508.

H-Asn-NH-(CH₂)₂(1-Me-1H-indol-3-yl) (31).

Prepared in 85% yield as a glass. ¹H NMR (400 MHz, CD₃OD) δ 7.57 (d, J = 6.8 Hz, 1 H), 7.29 (d, J = 8.4 Hz, 1 H), 7.15 (ddd, J = 8.2, 7.2, 1.2 Hz, 1 H), 7.03 (ddd, J = 7.8, 6.8, 0.8 Hz, 1 H), 6.99 (s, 1 H), 3.72 (s, 3 H), 3.62 (dd, J = 8.0, 4.8 Hz, 1 H), 3.48 (td, J = 7.4, 4.0 Hz, 2 H), 2.93 (t, J = 7.2 Hz, 2 H), 2.58 (dd, J = 15.2, 4.8 Hz, 1 H), 2.38 (dd, J = 15.2, 8.4 Hz, 1 H); ¹³C NMR (100 MHz, CD₃OD) δ 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+Na)⁺ requires 311.1484.

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) δ 7.52 (d, J = 8.0 Hz, 1H), 7.26 (d, J = 8.4 Hz, 1 H), 7.13 (t, J = 7.6 Hz, 1 H), 7.01 (t, J = 7.6 Hz, 1 H), 6.91 (s, 1H), 3.67 (s, 3 H), 3.61 (dd, J = 8.0, 5.2 Hz, 1 H), 3.28–3.18 (comp, 2 H), 2.75 (t, J = 7.4 Hz, 2 H), 2.59 (dd, J = 15.2, 4.8 Hz, 1 H), 2.41 (dd, J = 15.4, 8.2 Hz, 1 H), 1.88 (p, J = 7.4 Hz, 2 H); ¹³C NMR (100 MHz, CD₃OD) δ 176.3, 175.9, 138.7, 129.2, 127.6, 122.3, 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+H) requires 303.1822.

H-Asn-NH-(4-methylpentyl) (33).

Prepared in 67% yield as a glass. ¹H NMR (400 MHz, CD₃OD) δ 3.41 (dd, J = 7.1, 5.5 Hz, 1 H), 3.24–3.12 (comp, 2 H), 2.62 (dd, J = 15.2, 7.1 Hz, 1 H), 2.53–2.19 (comp, 2 H), 1.59–1.47 (comp, 3 H), 1.24–1.20 (m, 1 H), 0.89 (d, J = 6.5 Hz, 6 H); HRMS (ESI) m/z 216.1706; C₁₀H₂₂N₃O₂ (M+H) requires 216.1705.

H-Asn-NH-(5-methylhexyl) (34).

Prepared in 98% yield as a white solid; mp 210 °C (dec). ¹H NMR (400 MHz, CD₃OD) δ 3.61 (dd, J = 8.2, 5.1 Hz, 1 H), 3.24–3.12 (comp, 2 H), 2.62 (dd, J = 15.2, 8.2 Hz, 1 H), 2.42 (dd, J = 15.2, 8.2 Hz, 2 H), 1.59–1.47 (comp, 3 H), 1.24–1.20 (m, 1 H), 0.89 (d, J = 6.5 Hz, 6 H); ¹³C NMR (150 MHz) δ 221.5, 39.3, 39.1, 29.5, 27.9, 24.6, 21.8; HRMS (ESI) m/z 252.1682; C₁₁H₂₃N₃O₂Na (M+Na) requires 252.1688.

H-Asn-NH-(CH₂)₃-cyclohexyl (35).

Prepared in 72% yield as a white solid; mp 167–168 °C. ¹H NMR (400 MHz, CD₃OD) δ 4.13 (dd, J = 8.5, 4.8 Hz, 1 H), 3.25–3.12 (comp, 2 H), 2.84 (dd, J = 17.2, 4.8 Hz, 1 H), 2.74 (dd, J = 17.2, 4.8 Hz, 1 H), 1.79–1.62 (comp, 5 H), 1.55–1.51 (comp, 2 H), 1.27–1.18 (comp, 6 H), 0.94–0.84 (comp, 2 H); ¹³C NMR (150 MHz) δ 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+H) requires 256.2020.

H-Asn-NH-(CH₂)₃-(p-tolyl) (36)

Prepared in 40% overall yield as a white solid; mp 123–125 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.06 (br s, 4 H), 4.88 (comp, 5 H), 3.62 (dd, J = 5.2, 8.0 Hz, 1 H), 3.19 (m, 2 H), 2.62–2.57 (comp, 3 H), 2.43, (dd, J = 8.0, 15.2 Hz, 1 H), 2.27 (s, 3 H), 1.77–1.74 (comp, 2 H); ¹³C NMR (100 MHz, CD₃OD) δ 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+H) requires 264.1712.

H-Asn-NH-(CH₂)₃-(4-methoxyphenyl) (37).

Prepared in 99% yield as a white solid; mp 204–205 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.10 (d, J = 6.5 Hz, 2 H), 6.8 (d, J = 6.5 Hz, 2 H), 3.66–3.62 (m, 1 H), 3.23–3.19 (comp, 2 H), 2.66–2.44 (comp, 4 H), 2.15, (s, 3 H), 1.78 (p, J = 7.1 Hz, 2 H); ¹³C NMR (150 MHz) δ 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 280.1658; C₁₄H₂₂N₃O₃ (M+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 × 5 mL) to afford the desired product; mp 166–168 °C. ¹H NMR (400 MHz, CD₃OD) δ 7.26 (d, J = 8.8 Hz, 2 H), 7.19 (d, J = 8.0 Hz, 2 H), 4.12 (dd, J = 5.2, 8.4 Hz, 1 H), 3.23 (comp, 2 H), 2.84 (dd, J = 4.4, 16.8 Hz, 1 H), 2.75 (dd, J = 8.4, 17.2 Hz, 1 H), 2.63 (t, J = 7.6 Hz, 2 H), 1.82 (p, J = 6.8 Hz, 2 H); ¹³C NMR (100 MHz, CD₃OD) δ 171.8, 167.9, 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+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 (400 MHz, CD₃OD) δ 8.07 (d, J = 8.3 Hz, 1 H), 7.89–7.82 (dd, J = 7.6, 1.1 Hz,1 H), 7.72 (d, J = 7.8 Hz, 1 H), 7.56–7.43 (comp, 2 H), 7.43–7.32 (comp, 2 H), 4.16 (dd, J = 4.9, 8.5 Hz, 1 H), 3.44–3.31 (comp, 2 H), 3.17–3.08 (comp, 2 H), 2.92–2.72 (comp, 2 H), 2.00–1.93 (comp, 2 H); HRMS (ESI) m/z 300.1712; C₁₇H₂₂N₃O₂ (M+H) requires 300.1707.

H-Asn-NH-(CH₂)₃-(1-H-1H-indol-3-yl) (40).

Prepared in 47% yield as a yellow oil. ¹H NMR (400 MHz, CD₃OD) δ 7.52 (d, J = 8.0 Hz, 1 H), 7.32 (d, J = 8.0 Hz, 1 H), 7.07 (t, J = 7.6 Hz, 1 H), 7.02 (s, 1 H), 6.98 (t, J = 7.6 Hz, 1 H), 3.61 (dd, J = 8.0, 5.2 Hz, 1 H), 3.28–3.22 (m, 2 H), 2.77 (t, J = 7.4 Hz, 2 H), 2.55 (ddd, J = 29.8, 15.2, 5.4 Hz, 1 H), 2.41 (dd, J = 15.2, 8.2 Hz, 1 H), 1.90 (p, J = 7.2 Hz, 2 H). ¹³C NMR (100 MHz, CD₃OD) δ 176.0, 175.9, 138.2, 128.7, 123.0, 122.2, 119.5, 119.4, 115.6, 112.2, 53.3, 41.1, 40.3, 30.9, 23.6; HRMS (ESI) m/z 289.1658; C₁₅H₂₁N₄O₂ (M+H)⁺ requires 289.1663].

H-Asn-NH-(CH₂)₃-(1H-indol-1-yl) (41).

Prepared in 58% yield as a yellow oil after purification by silica gel chromatography eluting with CH₂Cl₂/MeOH (9:1) then CH₂Cl₂/MeOH (2:1). ¹H NMR (400 MHz, CD₃OD) δ 7.53 (d, J = 8.0 Hz, 1 H), 7.32 (dd, J = 8.4, 0.8 Hz, 1 H), 7.21 (d, J = 3.2 Hz, 1 H), 7.14 (td, J = 7.6, 1.0 Hz, 1 H), 7.01 (td, J = 7.4, 0.8 Hz, 1 H), 6.42 (dd, J = 3.2, 0.8 Hz, 1 H), 4.19 (t, J = 7.0 Hz, 2 H), 3.59 (dd, J = 7.6, 5.2 Hz, 1 H), 3.17 (td, J = 6.8, 1.6 Hz, 2 H), 2.59 (dd, J = 15.4, 5.4 Hz, 1 H), 2.44 (dd, J = 15.4, 7.8 Hz, 1 H), 2.00 (p, J = 6.8 Hz, 2 H); ¹³C NMR (100 MHz, CD₃OD) δ 175.7, 175.5, 137.4, 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+H) requires 289.1668.

General procedure for the synthesis of tripeptides 3–19.

NMM (31 μL, 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 °C. 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 6–12 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.

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 10% B to 95% B over 30 min. ¹H NMR (400 MHz, CD₃OD) δ 7.31–7.24 (comp, 5 H), 7.19–7.14 (comp, 4 H), 4.60 (dd, J = 13.3, 4.8 Hz, 2 H), 4.47 (dd, J = 15.2, 5.1 Hz, 1 H), 4.35 (dd, J = 13.3 Hz, 4.8 Hz, 1 H), 2.97 (dd, J = 14.4, 5.1 Hz, 1 H), 2.86 (dd, J = 14.4, 5.1 Hz, 1 H), 2.78–2.71 (comp, 2 H), 2.00–1.92 (comp, 2 H), 1.82 (s, 3 H), 1.76–1.69 (comp, 2 H), 1.56–1.46 (comp, 3 H), 1.27–1.18 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 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+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 10% B to 95% B over 30 min. ¹H NMR (400 MHz, CD₃OD) δ 7.27 – 7.24 (comp, 6 H), 7.17–7.14 (comp, 3 H), 4.65 (dd, J = 9.5, 6.3 Hz, 1 H), 4.53 (dd, J = 6.9, 5.1 Hz, 1 H), 3.42–3.38 (comp, 2 H), 3.13–3.09 (m, 1 H), 2.94–2.86 (m, 1 H), 2.84–2.68 (comp, 4 H), 1.98–1.95 (comp, 2 H), 1.88 (s, 3 H), 1.8–1.71 (comp, 3 H), 1.54 (comp, 3 H), 1.40–1.20 (comp, 4 H); ¹³C NMR (150 MHz, CD₃OD) δ 176.6, 175.7, 174.2, 173.6, 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; C₃₀H₃₉N₅O₉P (M–H) requires 668.2464.

Tripeptide 5.

Prepared from 27 in 28% 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 10% B to 95% B over 30 min. ¹H NMR (400 MHz, CD₃OD) δ 7.24 – 7.19 (comp, 6 H), 7.14–7.12 (comp, 3 H), 4.65 (dd, J = 9.5, 6.3 Hz, 1 H), 4.53 (dd, J = 6.9, 5.1 Hz, 1 H), 3.26–3.22 (comp, 2 H), 3.13–3.11 (m, 1 H), 2.92–2.71 (comp, 3 H) 2.66–2.62 (t, J = 7.5 Hz, 2 H), 2.05–1.93 (comp, 2 H), 1.88 (s, 3 H), 1.80–1.69 (comp, 3 H), 1.62–1.50 (comp, 3 H), 1.38–1.19 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 175.5, 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+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 10% B to 95% B over 30 min.¹H NMR (400 MHz, CD₃OD) δ 7.23 – 7.19 (comp, 6 H), 7.14–7.12 (comp, 3 H), 4.66 (dd, J = 9.5, 6.1 Hz, 1 H), 4.53 (dd, J = 6.9, 5.1 Hz, 1 H), 3.26–3.08 (comp, 3 H), 2.92–2.76 (comp, 3 H), 2.66–2.62 (t, J = 7.5 H, 2 H), 2.05–1.93 (br s, 2 H), 1.88 (s, 3 H), 1.80–1.71 (comp, 3 H), 1.61–1.50 (comp, 3 H), 1.37–1.18 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 176.6, 175.6, 174.2, 173.6, 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.

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 white solid; mp 136–138 °C. ¹H NMR (500 MHz, CD₃OD) δ 8.06 (br s, 0.9 H), 7.27 (d, J = 10.2 Hz, 2 H), 7.24–7.21 (comp, 2 H), 7.15–7.10 (comp, 5 H), 4.66 (dd, J = 11.1, 7.5 Hz, 1 H), 4.51 (dd, J = 8.4, 6.0 Hz, 1 H), 3.19 (td, J = 8.4, 3.6 Hz, 2 H), 3.13 (dd, J = 17.0, 7.5 Hz, 1 H), 2.91 (dd, J = 17.0, 10.8 Hz, 1 H), 2.81 (dd, J = 18.6, 8.4 Hz, 1 H), 2.72 (dd, J = 18.6, 6.0 Hz, 1 H), 2.59 (t, J = 9.3 Hz, 2 H), 2.02–1.94 (comp, 2 H), 1.90 (s, 3 H), 1.80–1.69 (comp, 2 H), 1.65–1.52 (comp, 7 H), 1.41–1.33 (comp, 3 H), 1.23 (app t, J = 11.4 Hz, 2 H); ¹³C NMR (125 MHz, CD₃OD) δ 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 according 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 162–164 °C. ¹H NMR (600 MHz, CD₃OD) δ 8.98 (br s, 0.8 H), 7.51 (d, J = 5.6 Hz, 1 H), 7.25 (d, J = 5.6 Hz, 1 H), 7.15–7.13 (m, 2 H), 7.08 (d, J = 5.6 Hz, 2 H),7.05–7.02 (m, 3 H), 4.66 (d, J = 9.6 Hz, 1 H), 4.62 (dd, J = 4.6, 3.4 Hz, 1 H), 4.51 (dd, J = 6.2, 4.2 Hz, 1 H), 4.47 (d, J = 9.6 Hz, 1 H), 3.65 (s, 3 H), 2.84–2.75 (m, 3 H), 2.62 (dd, J = 9.6, 6.0 Hz, 1 H), 1.94–1.89 (comp, 2 H), 1.76 (s, 3 H), 1.74–1.64 (comp, 2 H), 1.64–1.50 (comp, 3 H), 1.32–1.17 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 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+Na)⁺ requires 707.2565.

Tripeptide 9.

Prepared from 31 in 88% crude overall yield according 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 157–159 °C. ¹H NMR (600 MHz, CD₃OD) δ 8.02 (br s, 0.65 H), 7.57 (d, J = 5.2 Hz, 1 H), 7.28 (d, J = 5.6 Hz, 1 H), 7.22 (d, J = 5.6 Hz, 2 H), 7.15–7.12 (m, 3 H), 7.02–7.00 (m, 2 H), 4.63 (dd, J = 6.0, 4.0 Hz, 1 H), 4.56 (dd, J = 4.6, 3.4 Hz, 1 H), 3.73 (s, 3 H), 3.49 (t, J = 4.8 Hz, 2 H), 3.07 (dd, J = 9.4, 4.2 Hz, 1 H), 2.98 (t, J = 5.0 Hz, 2 H), 2.86 (dd, J = 9.6, 6.0 Hz, 1 H), 2.81 (dd, J = 10.4, 4.8 Hz, 1 H), 2.72 (dd, J = 10.4, 3.6 Hz, 1 H), 1.96–1.94 (m, 2 H), 1.97 (s, 3 H), 1.74–1.68 (comp, 2 H), 1.57–1.52 (comp, 3 H), 1.34–1.29 (m, 1 H), 1.22 (app t, J = 6.0 Hz, 2 H); ¹³C NMR (150 MHz, CD₃OD) δ 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.

Tripeptide 10.

Prepared from 32 in 95% crude overall yield according 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 139–143 °C. ¹H NMR (600 MHz, CD₃OD) δ 8.04 (s, 1 H), 7.54 (dt, J = 7.8, 1.2 Hz, 1 H), 7.27 (d, J = 8.4 Hz, 1 H), 7.16 (d, J = 8.4 Hz, 2 H), 7.13–7.10 (comp, 3 H), 6.98 (ddd, J = 7.8, 7.2, 1.2 Hz, 1 H), 6.94 (s, 1 H), 4.63 (dd, J = 9.3, 6.3 Hz, 1 H), 4.55 (dd, J = 6.9, 5.1 Hz, 1 H), 3.68 (s, 3 H), 3.33–3.23 (m, 2 H), 3.09 (dd, J = 14.4, 6.0 Hz, 1 H), 2.87–2.73 (comp, 5 H), 2.04–1.93 (comp, 4 H), 1.86 (s, 3 H), 1.82–1.70 (comp, 2 H), 1.59–1.52 (comp, 3 H), 1.37–1.19 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 176.7, 175.7, 174.3, 173.7, 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; C₃₄H₄₄N₆O₉P (M–H) requires 711.2903.

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) δ 7.26 (d, J = 8.2 Hz, 2 H), 7.14 (d, J = 8.2 Hz, 2 H), 4.62–4.68 (m, 1 H), 4.47–4.54 (m, 1 H), 3.08–3.20 (comp, 3 H), 2.91 (dd, J = 9.2 Hz, 7.2 Hz, 1 H), 2.80 (dd, J = 12.3 Hz, 7.0 Hz, 1 H), 2.73 (dd, J = 12.3 Hz, 7.0 Hz, 1 H), 2.00–1.92 (comp, 2 H), 1.91 (s, 3 H), 1.74–1.61 (comp, 2 H), 1.52–1.49 (comp, 5 H), 1.38 (m, 1 H) 1.22–1.17 (comp, 4 H), 0.87 (d, J = 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.

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) δ 7.19 (d, J = 8.2 Hz, 2 H), 7.05 (d, J = 8.2 Hz, 2 H), 4.57–4.55 (dd, 1 H, J = 8.9, 6.2 Hz, 1 H), 4.54–4.48 (dd, J = 7.2, 5.5 Hz, 1 H), 3.10–3.01 (comp, 3 H), 2.85–2.79 (m, 1 H), 2.74–2.61 (comp, 2 H), 2.00–1.88 (comp, 5 H), 1.74–1.61 (comp, 2 H), 1.52–1.49 (comp, 6 H), 1.29–1.12 (comp, 7 H) 0.78–0.76 (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.

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 10% B to 95% B over 30 min. ¹H NMR (400 MHz, CD₃OD) δ 7.28–7.26 (comp, 2 H), 7.15–7.13 (comp, 2 H), 4.68–4.64 (m, 1 H), 4.54–4.48 (m, 1 H), 3.19–3.07 (comp, 3 H), 2.92–2.89 (dd, J = 13.7 Hz, 9.6 Hz, 1 H), 2.83–2.79 (comp, 2 H), 2.00–1.92 (br d, 2 H), 1.88 (s, 3 H), 1.74–1.61 (comp, 7 H), 1.59–1.49 (comp, 5 H), 0.87–0.85 (comp, 2 H), 1.24–1.08 (comp, 9 H) 0.95–0.81 (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; C₃₁H₄₇N₅O₉P (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 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 156–158 °C. ¹H NMR (600 MHz, CD₃OD) δ 7.22 (d, J = 8.4 Hz, 2 H), 7.12 (dd, J = 1.2, 9.0 Hz, 2 H), 7.06 (d, J = 7.8 Hz, 2 H), 7.01 (d, J = 7.8 Hz, 2 H), 4.64 (dd, J = 6.6, 9.0 Hz, 1 H), 4.51 (dd, J = 4.8, 6.6 Hz, 1 H), 3.21 (comp, 2 H), 3.10 (dd, J = 6.0, 13.8 Hz, 1 H), 2.88 (dd, J = 9.0, 14.4, 1 H), 2.82 (dd, J = 7.2, 15.6 Hz, 1 H), 2.73 (dd, J = 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), 1.56–1.51 (comp, 3 H), 1.34–1.21 (comp, 3 H); ¹³C NMR (150 Hz, CD₃OD) δ 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 + Na)⁺ 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) δ 7.22 (d, J = 8.2 Hz, 2 H), 7.14–7.10 (comp, 4 H), 6.77 (d, J = 8.2 Hz, 2 H), 4.65 (dd, 1.0 H, J = 9.2, 6.2 Hz, 1 H), 4.54–4.50 (m, 1 H), 3.73 (s, 3 H), 3.25–3.16 (comp, 2 H), 3.11 (dd, J = 14.0, 6.2 Hz, 1 H), 2.91–2.75 (comp, 3 H), 2.57 (br t, J = 7.52 Hz, 2 H), 2.05–1.88 (comp, 2 H), 1.88 (s, 3 H), 1.85–1.68 (comp, 5 H), 1.62–1.48 (comp, 4 H), 1.40–1.18 (comp, 4 H); ¹³C NMR (125 MHz, CD₃OD) δ 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) requires 689.2784.

Tripeptide 16.

The protected tripeptide was prepared from 38 in 16% overall yield according to the general procedure for the synthesis 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 165–167 °C. ¹H NMR (600 MHz, CD₃OD) δ 7.22–7.18 (comp, 6 H), 7.13 (d, J = 8.4 Hz, 2 H), 4.64 (dd, J = 6.6, 9.6 Hz, 1 H), 4.50 (dd, J = 4.8, 6.6 Hz, 1 H), 3.20 (m, 2 H), 3.10 (dd, J = 6.6, 14.4 Hz, 1 H), 2.87 (dd, J = 9.0, 14.4 Hz, 1 H), 2.81 (dd, J = 7.2, 15.6 Hz, 1 H), 2.74 (dd, J = 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 = 6.6 Hz, 2 H), 1.75 (comp, 2 H), 1.56–1.52 (comp, 3 H), 1.36–1.22 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 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.

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) δ 8.09 (d, J = 8.5 Hz, 1 H), 7.82 (d, J = 8.0 Hz, 1 H), 7.68 (d, J = 7.9 Hz, 1 H), 7.48–7.41 (comp, 2 H), 7.39–7.32 (comp, 2 H), 7.16 (d, J = 8.5, 2 H), 7.09 (d, J = 8.5 Hz, 2 H), 4.63 (dd, J = 6.3, 9.2 Hz, 1 H), 4.55 (dd, J = 5.0, 6.9 Hz, 1 H), 3.37–3.31 (comp, 3 H), 3.12 (app t, J= 7.8 Hz, 2 H), 3.07 (dd, J = 6.3, 14.7 Hz, 1 H), 2.87–2.81 (comp, 2 H), 2.75 (dd, J= 5.0, 15.6 Hz, 1 H), 1.84 (s, 3 H), 2.03–1.93 (comp, 3 H), 1.81–1.68 (comp, 2 H), 1.58–1.50 (comp, 3 H), 11.36–1.20 (comp, 3 H); ¹³C NMR (150 MHz, CD₃OD) δ 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; C₃₅H₄₅N₅O₉P (M+H) requires 710.2949.

Tripeptide 18.

Prepared from 40 in 70% crude overall yield according 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 162–165 °C. ¹H NMR (600 MHz, CD₃OD) δ 8.04 (s, 1 H), 7.54 (d, J = 7.8, 1 H), 7.30 (d, J = 7.8 Hz, 1 H), 7.18 (d, J = 8.4 Hz, 2 H), 7.11 (d, J = 8.4 Hz, 2 H), 7.05 (app td, J = 7.5, 1.2 Hz, 1 H), 7.03 (s, 1 H), 6.95 (ddd, J = 7.8, 7.2, 1.2 Hz, 1 H), 4.63 (dd, J = 9.3, 6.3 Hz, 1 H), 4.55 (dd, J = 6.9, 5.1 Hz, 1 H), 3.32–3.24 (comp, 2 H), 3.09 (dd, J = 14.1, 6.3 Hz, 1 H), 2.88–2.73 (comp, 5 H), 2.03–1.94 (comp, 4 H), 1.86 (s, 3 H), 1.81–1.70 (comp, 2 H), 1.59–1.51 (comp, 3 H), 1.38–1.28 (m, 1 H), 1.26–1.19 (comp, 2 H); ¹³C NMR (150 MHz, CD₃OD) δ 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 according 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 145–148 °C. ¹H NMR (600 MHz, CD₃OD) δ 8.11 (s, 1 H), 7.51 (dt, J = 7.8, 0.9 Hz, 1 H), 7.44 (d, J = 8.4 Hz, 1 H), 7.28 (s, 1 H), 7.22 (d, J = 8.4 Hz, 2 H), 7.13–7.09 (comp, 3 H), 6.99 (ddd, J = 7.5, 7.2, 0.6 Hz, 1 H), 6.39 (dd, J = 3.0, 1.2 Hz, 0.1 H), 4.66 (dd, J = 9.0, 6.6 Hz, 1 H), 4.51 (dd, J = 7.2, 4.8 Hz, 1 H), 4.23 (t, J = 7.2 Hz, 2 H), 3.26–3.18 (comp, 2 H), 3.11 (dd, J = 13.8, 6.6 Hz, 1 H), 2.91–2.84 (comp, 2 H), 2.78 (dd, J = 15.6, 4.8 Hz, 1 H), 2.10–1.92 (comp, 4 H), 1.88 (s, 3 H), 1.82–1.72 (m, 2 H), 1.60–1.52 (comp, 3 H), 1.39–1.32 (m, 1 H), 1.26–1.19 (comp, 2 H); ¹³C NMR (150 MHz, CD₃OD) δ 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 solutions were prepared from distilled water that was passed through a Nanopure water purification system (Barnstead) to give a resistivity within the range 17.0–18.2 MΩcm. Chromatography columns were stored in a 20:80 v/v ethanol/water mixture at 4 °C.

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 53–163 of the Grb2 SH2 domain was obtained from the Schering-Plough Research Institute and expressed in E. coli (SG13009, Qiagen). Cultures were grown at 30 °C in LB media containing 0.1 g/L ampicillin (Acros Organics) and 0.035 g/L kanamycin (Sigma-Aldrich) to an OD600nm of 0.5–0.8 at which time expression was induced by isopropyl-β-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).

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 = 2,000), and centrifuged (3000 g) at 4 °C until the volume had been reduced to 800–1000 μL (ca 20–30 h) giving a concentration of ~ 30–100 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 approximately 0.75–1 bed volumes. The fractions containing pure monomeric Grb2 SH2 domain were combined to a final concentration of 250–500 μM 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 °C. The final HEPES dialysis buffer was retained and its pH re-adjusted to 7.50 using either HCl or NaOH. The dialyzed protein was diluted to a concentration of 65–85 μM in the final HEPES dialysis buffer prior to ITC experiments.

Solutions of ligands 2–19 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 solution. The buffer and the solutions of protein and ligands were filtered (0.20 μm PDVF) and degassed prior to the ITC experiment.

ITC experiments were performed in at least duplicate experiments employing independently prepared aliquots of ligand, and in some cases protein, using a Microcal VP-ITC equilibrated at 25 °C 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, followed by the above-mentioned rinse procedures. Typically, ~ 35 × 7–9 μL injections (200 sec intervals) of ligand were made into the sample cell containing protein, preceded by one 2 μL 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 thermodynamic parameters n (number of binding sites), Ka and ΔH°. 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. ΔG° was calculated indirectly from K_a by applying the modified Arrhenius equation, ΔG° = −RTlnK_a. ΔS° was calculated by applying the Gibbs relationship ΔG° = ΔH° − TΔS°. Prior to performing ITC titrations of flexible and constrained ligands with new batches of purified Grb2 SH2 domain, an ITC experiment using the tripeptide Ac–pTyr–Val–Asn–NH₂ was performed, and the thermodynamic values obtained were compared to those previously obtained in order to verify protein purity and folding.⁸ Representative ITC data for each ligand are given in Supporting Information.

The temperature dependence of ΔH°, ΔC_p, of binding for ligands 2, 5, 13, 17, and 19 was obtained by measuring ΔH° in duplicate at three different temperatures between 15 and 35 °C according to the procedure outlined above and determining the slope of the plot of ΔH° vs. temperature. The error in ΔC_p was obtained by determining the standard error in the slope.

Treatment of concentration error.

Errors in the thermodynamic 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 the thermodynamic relations ΔG° = −RTlnK_a and ΔG° = ΔH° − TΔS° to arrive at errors in the thermodynamic parameters K_a, ΔG°, ΔH°, and ΔS°.

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 a resistivity within the range 17.0–18.2 MΩcm. The resulting protein solution was placed in 15 mL Centriplus concentrators (Millipore, MWCO = 2,000) and centrifuged at 3000 g and 4 °C until the protein concentration was within the range 5–15 mg/mL. The ligands were each dissolved in this solution to give a protein-ligand molar ratio of 1:2. These solutions were each heated to 50 °C for 10 min to convert any Grb2 SH2 domain-swapped dimer to the monomeric domain,⁴⁴ filtered through a 0.45 micron PVDF filter disk, cooled, and stored at 4 °C. 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-drop, vapor-diffusion method, employing 7 μL drops and 350 μL 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 molar ratio of 5 to protein (ca. 10 mg/mL) was prepared. This solution (4.0 μL) was mixed with a precipitant solution containing 0.1 M HEPES, pH 7.5, and 25% w/v polyethylene glycol, MW 10,000 (3.0 μL) and allowed to equilibrate with the aforementioned precipitant well solution (350 μL) at 298 K. Useable crystals grew after 4 weeks.

Grb2 SH2 complexed with 11.

An aqueous solution containing a 1.5 molar ratio of 11 to protein (ca. 10 mg/mL) was prepared. This solution (4.0 μL) was mixed with a precipitant solution containing 0.2 M sodium citrate tribasic dihydrate, 0.1 M HEPES, pH 7.5, and 20% v/v 2-propanol (3.0 μL Hampton crystal screen I, condition no. 27) and allowed to equilibrate with the aforementioned precipitant well solution (350 μL) at 298 K. Useable crystals grew after four weeks.

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 25–35% (v/v). Crystals were allowed to equilibrate in this solution for 0.5–1 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 70 mA producing CuKα graphite-monochromatic radiation (1.5418 nm). Data frames were collected in 0.5° intervals with exposure times of 120 sec at crystal-to-detector distances of 100–120 mm.

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 COOT⁴⁸ 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 polar Connolly surface area, ΔCSA_np and ΔCSA_p,⁵⁰ were calculated using Macromodel v. 9.1⁵¹ based on the assumption that no change in the conformation of either the ligand or the protein occurs 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 ΔCSA values for the formation of each complex. A probe radius of 1.4 Å was employed.

Highlights.

• Adding nonpolar groups increases affinity

• Alkyl groups increase affinity because of more favorable changes in ΔS°

• Alkyl aryl groups increase affinity because of more favorable changes in ΔH° and ΔS°

• Binding energetics with added nonpolar groups characterized by ΔH/ΔS compensation

• Differences in ΔΔG° and ΔΔH° correlate with buried nonpolar surface area