Table 1.
Cosolvent MD techniques that have been used to identify hotspots and binding sites on protein surfaces (protein abbreviations defined in the text).
| Developer (Method) | Highlights, Needed Improvements, Published Cosolvents & Proteins |
|---|---|
| Barril (MDmix) 33–36 |
Highlights: This was the first method of this type, and it laid the foundation for using occupancy grids and calculating free energies from cosolvent populations. MDmix focuses on water-miscible cosolvents. Needed Improvements: MDmix produces many “extra” hotspots that may be misleading in prospective applications. Cosolvents: Isopropanol, ethanol, acetonitrile, methanol, acetamide Proteins: Thermolysin, p53, elastase, MDM2, LFA-1/ICAM-1, PTP1B, p38 MAPK, AR, HEWL, Hsp90, HIVp |
| MacKerell (SILCS) 37–48 |
Highlights: This method has the most extensive development with significant progress in translating occupancy grids into pharmacophore models and scoring schemes. Needed Improvements: It uses high concentrations of cosolvent with artificial repulsion terms to prevent aggregation. This may unnaturally perturb any cooperative behavior between the cosolvent and create artifacts in the maps. SILCS also produces many extra hotspots that may be misleading in prospective applications. Cosolvents: Benzene, propane, water (as a hydrogen-bonding probe), acetonitrile, methanol, formamide, acetaldehyde, methylammonium, acetate, imidazole Proteins: BCL-6, trypsin, α-thrombin, HIVp, FKBP, FXa, NadD, RNase A, IL-2, p38 MAPK, DHFR, FGFr1 kinase, adenosine deaminase, ERα, AmpC β-lactamase, T4-L99A, AR, PPARγ, mGluR5, β2AR |
| Carlson (MixMD) 49–53 |
Highlights: Very careful development has lead to clean maps with a significantly reduced number of extra hotspots. MixMD focuses on very low concentrations of miscible solvents to avoid artificial repulsion terms. Needed Improvements: At this point, MixMD is qualitative in its identification of hotspots, and a quantitative scoring scheme is needed. Cosolvents: Acetonitrile, isopropanol, pyrimidine, imidazole, N-methylacetamide, acetate, methylammonium Proteins: HEWL, elastase, p53, RNase A, thermolysin, HIVp, ABL kinase, AR, CHK1 kinase, glucokinase, PDK1 kinase, PTP1B, farnesyl pyrophosphate synthase |
| Yang and Wang 54–58 |
Highlights: The authors have used more rigorous free energy calculations to estimate binding affinities. Other applications have focused on qualitatively identifying differences in PPI that might help provide specificity for designed ligands. Needed Improvements: More development is needed. Cosolvents: Isopropanol, phenol, trimethylamine N-oxide Proteins: Thermolysin, Bcl-xL, Mcl-1, IL-1R1 |
| GlaxoSmithKline and Bahar 59 |
Highlights: The method is specifically developed for assessing druggability of individual binding sites. They use their grids in a slightly different way, and they have very interesting rules for combining hotspots into druggability estimates. Needed Improvements: More development is needed. Cosolvents: Isopropanol, isopropylamine, acetic acid, acetamide Proteins: MDM2, PTP1B, LFA-1, kinesin Eg5, p38 MAPK |
| Caflisch 60,61 |
Highlights: This method estimates kinetic on/off rates and binding affinities of the cosolvents based on the MD, but only a few applications are published. Cosolvents: Dimethylsulfoxide, methanol, ethanol Proteins: FKBP, BAZ2B, CREBBP |
| Tan and Abell 62–64 |
Highlights: This method proposes low concentrations of hyrdrophobic cosolvents to reduce aggregation, but only a few applications are published. Cosolvents: Benzene, chlorobenzene Proteins: Polo-box domain of polo-like kinase 1, MDM2, MDMX, IL-2, Mcl-1, Bcl-xL, Aurora-A, RAD51, ERα, ERβ |
| Fersht 65 |
Highlights: The application focuses on cryptic binding sites, and more work is needed. Cosolvent: Isopropanol Protein: p53-Y220C |
| Gorfe (pMD) 66,67 |
Highlights: This method is also developed to map proteins embedded in membranes, but more applications are needed. Cosolvent: Isopropanol Protein: K-ras |