An Affordable Topography-Based Protocol for Assigning a Residue’s Character on a Hydropathy (PARCH) Scale

The hydropathy of proteins or quantitative assessment of protein–water interactions has been a topic of interest for decades. Most hydropathy scales use a residue-based or atom-based approach to assign fixed numerical values to the 20 amino acids and categorize them as hydrophilic, hydroneutral, or hydrophobic. These scales overlook the protein’s nanoscale topography, such as bumps, crevices, cavities, clefts, pockets, and channels, in calculating the hydropathy of the residues. Some recent studies have included protein topography in determining hydrophobic patches on protein surfaces, but these methods do not provide a hydropathy scale. To overcome the limitations in the existing methods, we have developed a Protocol for Assigning a Residue’s Character on the Hydropathy (PARCH) scale that adopts a holistic approach to assigning the hydropathy of a residue. The parch scale evaluates the collective response of the water molecules in the protein’s first hydration shell to increasing temperatures. We performed the parch analysis of a set of well-studied proteins that include the following—enzymes, immune proteins, and integral membrane proteins, as well as fungal and virus capsid proteins. Since the parch scale evaluates every residue based on its location, a residue may have very different parch values inside a crevice versus a surface bump. Thus, a residue can have a range of parch values (or hydropathies) dictated by the local geometry. The parch scale calculations are computationally inexpensive and can compare hydropathies of different proteins. The parch analysis can affordably and reliably aid in designing nanostructured surfaces, identifying hydrophilic and hydrophobic patches, and drug discovery.


Contents
Table S1.Input parameter definitions and setup ..           S2.PARCH values of the 20 single amino acids.
The N-and C-termini of the single amino were charged as well as the side chains (when applicable).

Figure S1. Radial distribution function of water and proteins
We evaluated the hydration of the equilibrated proteins.Figure S1 shows the radial distribution functions of the oxygen atom in water (OH2) relative to the protein backbone for the proteins.In all cases, water was structured around the proteins in the first, second, and third hydration shells at an average distance of 0.315 nm, 0.415 nm, and 0.515 nm, respectively.The d shell = 0.415 nm was selected.cases.An increase in the d ion and d b values from 3.0 to 5.0 nm affected the parch values, but the changes were small (Figure S2).The cutoff did not change the parch value trends; the residues with high parch values remained higher than others when d ion and d b values were changed.
In terms of the computational cost, increasing d ion and d b values had a substantial effect (Figure S3).For example, the wall-clock time for BNS at d ion = d b =5.0 nm was 4x larger than d b =d ion =3.0, and for TS, the exact change led to a 6x higher cost.We expect that for larger proteins than BNS and TS, the computational cost will be much larger with d b =di on =5.0 nm for a small change in the residues' parch values.To ensure the affordability of the PARCH scale calculations, we selected d b =d ion =3.0 nm.Annealing rate

S12
We studied the effect of annealing rates on parch value calculations for BNS and TS as test cases.For both proteins, three annealing rates-1 K/1 ps (high), 1 K/10 ps (medium), and 1 K/100 ps (low)-were investigated.We found that high annealing rates led to higher parch values than medium and low annealing rates (Figure S4).A view into the annealing simulation trajectories showed that water could not establish a network around the protein at a high annealing rate, leading to an overestimation of the parch values.In contrast, water can establish a network at medium and low annealing rates.Also, the parch value differences between medium and low annealing rates are minor.
The computational costs for the three annealing rates are remarkably different (Figure S5).Although the high annealing rate is the least expensive, it is unfit due to overestimating the parch values.The cost of the low annealing rate is 6 to 9 times higher than the medium annealing rate.Therefore, based on accuracy and efficiency, we selected the medium annealing rate (1 K/10 ps) for our parch calculations.

Figure S2 .
Figure S2.Parch values of (a) BNS and (b) TS residues for different d ion and d b values.

Figure S4 .
Figure S4.Parch values of (a) BNS and (b) TS residues for different d water values

Figure
Figure S7.PARCH values of the (a) BNS and (b) TS residues for different force constants

Table S1 . Input parameter definitions and setup
S4Table