Which is the amino acids classification when analyzing the solvent accessibility of a protein?

Question

I came across an article which has defined 3 categories: Buried [A, L, F, C, G, I, V, W], Exposed [ R, K, Q, E, N, D] and Intermediate [M, S, P, T, H, Y] when analyzing the solvent accessibility of a protein. But I could not find any reference validating the claim.

Another article has come up with 4 divisions:

B=completely buried (0-4% exposed)
b=partly buried (4-25% exposed)
e=partly exposed (25-50% exposed)
E=completely exposed (50+% exposed)

Which is authentic when dealing with solvent accessibility, and which should be used? Also for the second article [B,b,e,E], which amino acid corresponds to which class?

Answer 1

There is some confusion in the question.

Solvent accessibility of a residue, i.e. the area accessible to the solvent (water), depends on the 3D structure of the protein. It's not a property of an amino-acid (chemical compound such as alanine, leucine, ...).

There may be a correlation -- amino-acids with hydrophobic side chains are more likely to be buried, but in general amino-acids (chemical compounds) cannot be classified as buried or exposed.

I don't have access to the first article, but I tried running the prediction server described in the second linked article, and I got a result

GSGMQIFVKTFRDRLRNLLPQTITLEVEPSDTIENVKAKIQDKEGIPPDQQVLIFSRKRL
CCCCEEEEEECEEEEEECCCCEEEEEECCCCCHHHHHHHHHHHHCCCHHHEEEEECCEEC
EEEBeBbBebbEbEbEebEEeEbeBEBeeEeeBeeBbEeBeEeEEbEeEeBeBbbEEeeb

...

Porter (Secondary Structure):
H = Helix   (DSSP classes H, G and I)
E = Strand  (DSSP classes E and B)
C = Coil    (DSSP classes S, T and .)

PaleAle (Relative Solvent Accessibility):
B = very buried      (<=4% accessible)
b = somewhat buried  (>4% and <=25% accessible)
e = somewhat exposed (>25% and <=50% accessible)
E = very exposed     (>50% accessible)

In one place in the sequence you can see QQ, predicted as eB. So clearly glutamine is NOT assigned to a particular RSA class.

Dividing RSA (relative solvent accessibility, 0-100%) into categories is arbitrary, I don't think there is any widely used convention. The easiest way would be to just use numeric values.