When analyzing a drug binding pocket it is important to assess which amino acid functional groups offer the biggest potential for increase in binding affinity. As H-bonds make up around 80% of protein-ligand interactions, we’ll start with asking the question: which amino acids form the strongest H-bonds.

charge and the twenty amino acids

The twenty common amino acids can be be grouped into four classes: lipophilic (8), polar (6), charged (5) and reactive (1) (see figure below). Before considering the H-bonding strenths of the various amino acids (and of course comparing them to water) the topic of charge needs to be addressed. This is because, whether a functional group is neutral, positively charged through protonation or negatively charged through deprotonation has a dramatic effect on it´s H-bonding strength. This is why we must first look at the concept of pKa for the amino acids.

Mathmatically, pKa is the negative base 10 logarithm of the acid dissocation constant Ka. The lower the pKa value the stronger the acid. Conversely, the higher the pKa value the stronger the base. The inverse scale pKb is sometimes used to describe bases but we will refrain from that here in order to avoid confusion and use soley the pKa scale to describe both acids and bases.

The two carboxylic acid sidechain containing amino acids aspartic acid and glutamic acid have pKas of 3.4 and 4.1 respectively. That means, at a pH 3.4 aspartic acid is 50% protonated and 50% deprotonated while at physiological pH values (eg pH 7.4) it is essentially fully deprotonated. The deprotonated form of aspartic acid is referred to as aspartate.

On the other end of the scale lie lysine (pKa = 10.7), tyrosine (pKa 11.0) and arginine (pKa 13.8) all have pKa values several log units above 7.4 and hence are fully protonated at physiological pH. While tyrosine is a neutral amino acid, lysine and arginine are positively charged under physiological conditions and are referred to as lysinium and argininium.

While the pKa values of the side chains of amino acids vary depending on their local environment with protein structures this has no impact on the five aforemention amino acids (Asp, Glu, Lys, Tyr and Arg) because their average pKa is so far away from (several orders of magnitude) physiological pH. This is in contrast to histidine which has a pKa within one order of magnitude of pH 7.4. Histidine has an average pKa of 6.5 but can vary from 5 to 8, hence we need to consider both histidine and histidinium if this amino acid occurs in your binding pocket of interest.

quantifying H-bonding strength

Quantitative scales describing H-bond acceptor (HBA) and H-bond donor (HBD) strengths have been developed by Abrahams and extended by Hunter to include influence of solvent on solution phase equilibria between H-bonded solutes. The solvent competition model, the Hunter scales of H-bond donor strength (α) and H-bond acceptor strength (β) and the two H-bond strategies (winning and losing to water) have been covered in the post In Drug Design Two Thirds of H-Bonds are Weak. It should be noted that the use these scales to semi-quantitatively estimate contributions to binding energies are limited to enthalpic contributions with entropical contributions not being assessed by this method.

H-bond donor strengths of amino acids

H-bond donor functionality of amino acids can be either stronger than water’s H-Bond donor strength, similar or weaker. The table below using model systems for each amino acid functionality highlights these three groups. Not surprisingly, it is the protonated amino acids that are the strongest H-bond donating amino acids with argininium, lysinium and histidinium being around two α units stronger than the water dimer. The phenolic group of tyrosine and amide functionalities of glutamine and asparagine are also significantly stronger donors at around one α unit greater than water. The group of amino acids with H-bond donor strengths comparable to water is surprisingly large. The indole and imidazole NH groups of tryptophan and histidine, the NH of backbone amides and the OH groups of serine and threonine have H-bond donor strengths comparable to water and thus large enthalpic contributions from H-bond formation should not be expected when forming H-bonds with these functionalities. All other amino acids having only weakly polarized CH bonds are all weaker than water’s H-Bond donor strength.

H-bond acceptor strengths of amino acids

In contrast to H-bond donors, H-bond acceptor functionality of amino acids fall into only two categories, either stronger or weaker than water’s H-bond acceptor strength (see table below). The negatively charged amino acids aspartate and glutamate are by far the strongest acceptors amongst amino acids. Histidine and the carbonyl groups of backbone amides, asparagine and glutamine coming in a distant second. A third group, somewhat weaker again but still significantly stronger than water are the alcholo accepting groups of threonine and serine. The accepting pi systems of tryptophan, tyrosine and phenylalanine are all weaker acceptors versus waters oxygen atom as are the sulfur lone pairs of methionine and cysteine. The aliphatic amino acids, having no formal acceptor functionality accumulate at the bottom of the list.

summary

The charged amino acids are both the strongest H-bond donors (argininium, lysinium and histidinium) and H-bond acceptors (aspartate and glutamate). It is interesting to note that the strongest amino acid acceptors (aspartate and glutamate) are much stronger relative to water’s acceptor strength versus the strongest amino acid donors (argininium, lysinium and histidinium) relative to water’s donor strength. There is a group of H-bond donating amino acids which have comparable donor strength to water (tryptophan, histidine, the NH of backbone amides, serine and threonine) and as such limited enthalpic gain can be expected from such H-bonds. In contrast, H-bond accepting amino acids have a clear delination between those that are stronger than water and those that are weaker. These tables can be used to select which amino acids to target in your drug target binding pocket and also which H-bond strategy to adopt (quadrants B or C see graphic above).