Hydrophobic Interactions in
Proteins
Brian W Matthews, University of Oregon, Eugene, Oregon, USA
Proteins fold spontaneously into complicated three-dimensional structures that are
essential for biological activity. Much of the driving energy for this folding process comes
from the hydrophobic effect, i.e. the removal of nonpolar amino acids from solvent and
their burial in the core of the protein.
Introduction
The hydrophobic (literally, ‘water-hating’) effect is named
for the tendency of certain oil-like substances to avoid
contact with water (proverbially, ‘oil and water don’t
mix’). It is generally understood to be the driving force
responsible for the folding of proteins.
Proteins are synthesized in the cell as polymers made up
of linked units – amino acids. In nature there are 20
possible aminoacids thatareused forthis purpose(Table1).
Any given protein is characterized by the number of amino
acids that it contains, together with the sequence in which
the amino acids are arranged. It is the sequence of amino
acids that determines the active three-dimensional shape of
the protein. How the sequence information is used to define
the structure is not understood in detail. What is clear,
however, is that substantial free energy is required to drive
the polymer chain into a well-defined structure, and to
prevent it from unravelling. It is also generally accepted
that the energy for this process comes primarily from the
hydrophobic effect.
Polar and Nonpolar Amino Acids
About a quarter of the amino acids have side-chains that
are normally charged. These are called ‘hydrophilic’ (water
loving) and prefer to be in an aqueous environment. In
contrast, about another third have side-chains that are
made up of hydrocarbon atoms. Three examples are
alanine, leucine and phenylalanine (Table 1). These sidechains
are ‘oily’ in character and prefer to be segregated in
contact with each other, and out of contact with water. The
amino acids that contain such nonpolar side-chains are
described as ‘hydrophobic’. The side-chains of the
remaining amino acids typically contain both polar and
nonpolar atoms and have properties that reflect the
characteristics of both.
When a protein folds into a well-defined three-dimensional
structure, the majority of the hydrophobic sidechains
cluster together within the core of the protein
(something like a tiny oil-drop). This removal of the
hydrophobic side-chains from contact with solvent is
highly favourable and generates sufficient free energy to
maintain the folded structure of the protein.
Forces that Stabilize Protein Structures
Even simple proteins have complicated three-dimensional
shapes that can include a helices, b strands, turns and
irregular segments (Figure1). A number of different types of
interaction help define the structure. These include
hydrogen bonds, electrostatic interactions, van der Waals
interactions and hydrophobic interactions.
Because proteins fold in an aqueous environment, the
contribution of a given interaction to the folding of the
protein depends not so much on the strength of interaction
within the protein but on the difference between the
strength of the interaction within the protein and the
strength of interaction of the same groups with water. A
hydrogen bond, for example, may occur in the folded
protein between a hydrogen bond donor and a hydrogen
bond acceptor. In the unfolded protein, however, both the
donor and acceptor will make more-or-less equivalent
hydrogen bonds to water. Thus the net energetic contribution
to protein folding from hydrogen bonding tends to be
rather weak. (Hydrogen bonds are, however, thought to be
especially important in discriminating between the correct
folded structure and incorrect ones.) Similarly, van der
Waals interactions occur throughout folded proteins, but
equivalent interactions with solvent can occur in the
unfolded form.
Hydrophobic interactions, however, are different. As
noted above, protein folding occurs in the presence of
water and the properties of water are dominated by its
propensity to form hydrogen bonds. Polar compounds
such as sugars can share hydrogen bonds with water and,
for this reason, are readily soluble. In contrast, when a
hydrophobic (nonpolar) surface is introduced into an
aqueous environment it precludes hydrogen bonding. This
preclusion of hydrogen bonding to the hydrophobic
surface forces the water molecules to adopt alternative
Article Contents
Introductory article
. Introduction
. Polar and Nonpolar Amino Acids
. Forces that Stabilize Protein Structures
. Estimation of the Strength of the Hydrophobic Effect
. Relationship between the Hydrophobic Effect and
Surface Area
. The Hydrophobic Moment
. Core Packing and the Effects of Mutations
1ENCYCLOPEDIA OF LIFE SCIENCES © 2001, John Wiley & Sons, Ltd. www.els.net
Table 1 List of 20 fundamental amino acids and their abbreviations
Amino acid Three-letter abbreviation One-letter symbol Formula
Alanine Ala A
Arginine Arg R
Asparagine Asn N
Aspartic acid Asp D
Cysteine Cys C
Glutamic acid Glu E
Glutamine Gln Q
Glycine Gly G
Histidine His H
Isoleucine Ile I
Leucine Leu L
continued
Hydrophobic Interactions in Proteins
2
arrangements that permit hydrogen bonding to other
water molecules. This imposed restriction on the alignment
of the water molecules (strictly speaking, a reduction in
their entropy) has an energetic cost and is the physical basis
of the hydrophobic effect. Because the folding of a protein
includes the removal of many nonpolar side-chains from
an aqueous environment and their sequestration from
solvent, the energy benefit can be very substantial.
Estimation of the Strength of the
Hydrophobic Effect
The classical way to estimate the magnitude of the
hydrophobic effect for a given compound is to measure
the free energy of transfer, DGtr, of the compound from the
gas, liquid or solid phase to water. A positive value for DGtr
means that the molecule prefers a nonaqueous environTable
1 – continued
Amino acid Three-letter abbreviation One-letter symbol Formula
Lysine Lys K
Methionine Met M
Phenylalanine Phe F
Proline Pro P
Serine Ser S
Threonine Thr T
Tryptophan Trp W
Tyrosine Tyr Y
Valine Val V
Hydrophobic Interactions in Proteins
3
ment. In the case of the amino acids, measurements can be
made with the free amino acid or with variants modified to
better represent the amino acids incorporated within the
protein chain.
As well as choosing the particular compounds to be
investigated, one also has to decide which transfer medium
best mimics the interior of a protein. For this reason many
different hydrophobicity scales are available. Table 2 shows
hydrophobicities of the 20 amino acids based on transfer
between water and octanol. It is the relative value of the
hydrophobicity that is relevant. For this reason it is often
convenient to define the hydrophobicity of glycine to be
zero and to quote values for the other amino acids relative
to this reference.
As can be seen in Table 2, amino acids with large,
nonpolar or largely nonpolar side-chains such as leucine
and tryptophan (Table 1) are most hydrophobic. The least
hydrophobic amino acids are those that are charged and
those such as asparagine that are largely polar.
Relationship between the Hydrophobic
Effect and Surface Area
It was noted above that the hydrophobic effect is due to the
effect of nonpolar atoms on surrounding water molecules.
As such, one would expect the magnitude of the hydrophobic
effect for a given amino acid to be proportional to
the surface area of the nonpolar atoms that it contains.
That this is the case is illustrated in Figure 2. The amino
acids alanine, valine, leucine and phenylalanine have sidechains
that are made up of non-polar hydrocarbon atoms
(Table 1). These lie on one straight line. The other amino
acids shown in Figure 2 have side-chains that are partly
made up of hydrocarbon atoms but also include some
polar atoms as well. These lie on a second line.
The Hydrophobic Moment
As shown in Table 2, the hydrophobicities of the given
amino acids vary substantially. Some are strongly hydrophobic
while others are strongly hydrophilic. A molecule
N
C
Figure 1 Sketch of the backbone structure of the protein methionine
aminopeptidase from Escherichia coli. The chain begins at the aminoterminus
(N), includes 264 amino acids, and ends at the carboxy-terminus
(C). The protein includes a helices and b sheet strands as well as two metal
ions (spheres) at the active site. Reprinted from Bazan JF et al. (1994)
Proceedings of the National Academy of Sciences of the USA 91: 2473–2477,
figure 2.
Table 2 Hydrophobicitiesa of the 20 naturally occurring
amino acids
a
The hydrophobicities are based on the solvent transfer free energies
from octanol to water.
Hydrophobicity
Amino acid (kJ mol–1) (kcal mol–1)
Tryptophan 9.41 2.25
Phenylalanine 7.49 1.79
Isoleucine 7.53 1.80
Leucine 7.11 1.70
Cysteine 6.44 1.54
Methionine 5.14 1.23
Valine 5.10 1.22
Tyrosine 4.02 0.96
Proline 3.01 0.72
Alanine 1.30 0.31
Threonine 1.09 0.26
Glycine 0.00 0.00
Serine –0.17 –0.04
Histidine 0.54 0.13
Glutamine –0.92 –0.22
Asparagine –2.51 –0.60
Glutamic acid –2.68 –0.64
Aspartic acid –3.22 –0.77
Lysine –4.14 –0.99
Arginine –4.23 –1.01
Hydrophobic Interactions in Proteins
4
that includes both hydrophobic and hydrophilic parts is
called amphiphilic. For such amphiphilic molecules it is
sometimes useful to define a hydrophobic moment, which
is analogous to a dipole moment. For a single amino acid,
the hydrophobic moment can be defined as a line that
points from the Ca atom to the middle of the side-chain,
and whose length is proportional to the hydrophobicity of
the side-chain. For a protein or part of a protein, the dipole
moment is obtained by summing the individual vectors (in
magnitude and direction) corresponding to the constituent
amino acids. As an example, an a helix located on the
surface of a protein will have one side of the helix exposed
to solvent and the other side facing the interior of the
protein. The amino acids that comprise the buried side of
the a helix will, in general, be much more hydrophobic than
those on the solvent-exposed side of the helix. Because of
this asymmetry the a helix will have a large hydrophobic
moment directed towards the centre of the protein.
Core Packing and the Effects of
Mutations
Attempts have been made to estimate the magnitude of the
hydrophobic effect by substituting one nonpolar amino
acid for another within the core of a protein and measuring
the resultant change in stability of the protein. One
difficulty, however, is that the cores of proteins are tightly
packed and the substitution of amino acids of different
shapes and sizes tends to introduce steric clashes that
complicate the interpretation of the experiment. A possible
way to avoid such steric interference is to only use
replacements in which a larger nonpolar residue is replaced
Table 3 Changes in protein stability resulting from the replacement of larger hydrophobic amino
acids with smaller ones
∆∆G (kJ mol–1
)
Substitution Number of examples Low High Average
Ile → Val 9 2.1 7.6 5.4±1.7
Ile → Ala 9 4.6 21.3 15.9±2.9
Leu → Ala 17 7.1 25.9 14.6±4.6
Val → Ala 11 0.0 19.7 10.5±3.8
Met → Ala 4 8.8 19.2 12.5±3.8
Phe → Ala 4 14.6 18.4 15.9±1.2
0 2500 5000 7500 10000 12500 15000–400
Hydrophobicity (J mol–1)
20
40
60
80
100
120
140
160
180
Trp
Phe
Tyr
Met
His
Thr
Leu
Val
Ala
Accessiblesurfacearea(Å2
)
Ser
Figure 2 Relationship between the hydrophobicities of the amino acids
and the solvent-accessible areas of their side-chains. Reprinted and
adopted from Chothia C (1974) Nature 248: 338–339, figure 1.
25
0
Decreaseinproteinstability
∆∆G(kJmol–1)
0 25 50 75 100 125 150
Increase in cavity volume (Å3)
L121A
L46A
L118A
L133A
L99A
Figure 3 Loss of protein stability, DDG, for a series of leucine-to-alanine
substitutions within T4 lysozyme. ‘L99A’, for example, denotes the mutant
in which leucine 99 is replaced by alanine. Mutations that result in cavities
of the largest volume cause the greatest loss of stability. At a cavity volume
of zero, the loss of stability can be attributed to the difference between the
hydrophobicity of leucine and alanine. Reprinted and adapted from Xu J
et al. (1998) Protein Science 7: 158–177, figure 17A.
Hydrophobic Interactions in Proteins
5
with a smaller one. Table3 summarizes the results of a series
of experiments of this type. If we consider, for example, the
leucine-to-alanine substitutions included in the table, the
average loss in stability is 14.6 + 4.6 kJ mol2 1
. As can be
seen, however, there is a very large spread in the individual
measurements, ranging from 7.1 to 25.9 kJ mol2 1
. This
variation occurs because a given protein may respond in
different ways to leucine-to-alanine substitutions at
different sites. Sometimes the protein structure surrounding
the replacement will hardly change at all, with the result
that a cavity will be formed. In other cases the atoms
surrounding the smaller replacement will collapse or partly
collapse to occupy the space vacated by the large sidechain.
Mutations that lead to the creation of larger cavities
tend to be more destabilizing than those that lead to small
cavities (Figure 3). This in turn suggests that the loss in
stability resulting from large-to-small substitutions such as
leucine-to-alanine is due to two different factors: (1) the
hydrophobic effect and (2) van der Waals interactions
between the large side-chain and the atoms it contacts. If
the protein structure remains the same, the replacement of
the large side-chain with a smaller one will result in the loss
of some of these favourable van der Waals contacts.
However, if the structure relaxes in response to the
mutation, it will regenerate new van der Waals contacts
that will tend to offset those present in the parent structure.
This van der Waals term, which varies from one mutation
site to another, explains why the loss of stability depends
on the size of the cavity that is created (Figure 3).
If the straight line shown in Figure 3 is extrapolated to
zero cavity volume, one can define this intercept as DG0.
Here the protein structure relaxes to completely fill any
space created by the mutation. The intercept value is the
difference between the energy of burying an alanine rather
than a leucine in the core of the protein. In other words, the
intercept measures the difference between the hydrophobic
stabilization of leucine and alanine. The value of
8.8 kJ mol2 1
obtained from Figure 3 corresponds only
moderately well to the difference between the octanol-towater
transfer free energies of the same two amino acids
(5.81 kJ mol2 1
; Table 2). While the approach illustrated in
Figure 3 illustrates, in principle, how amino acid hydrophobicities
can be determined from mutant proteins, its use
is limited in practice by a number of technical factors. For
this reason the amino acid hydrophobicities obtained from
solvent transfer experiments (such as in Table 2) are
recommended for everyday use.
Further Reading
Creighton TE (1992) Protein Folding. New York: WH Freeman.
Richards FM (1991) The protein folding problem. Scientific American
(January 1991), pp. 54–63.
Matthews BW (1996) Structural and genetic analysis of the folding and
function of T4 lysozyme. FASEB Journal 10: 35–41.
Pace CN, Shirley BA, McNutt M and Gajiwala K (1996) Forces
contributing to the conformational stability of proteins. FASEB
Journal 10: 75–83.
Eriksson AE, Baase WA, Zhang X-J et al. (1992) Response of a protein
structure to cavity-creating mutations and its relation to the
hydrophobic effect. Science 255: 178–183.
Xu J, Baase WA, Baldwin E and Matthews BW (1998) The response of
T4 lysozyme to large-to-small substitutions within the core and its
relation to the hydrophobic effect. Protein Science 7: 158–177.
Hydrophobic Interactions in Proteins
6