You are on page 1of 126

By Glove@Smth

Visit (Click Here) For More Resources

Fundamental Principles
of Membrane Biophysics

David Njus
Department of Biological Sciences
Wayne State University

© D. Njus, 2000
Fundamental Principles
of Membrane Biophysics

David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000
Table of Contents

Chapter 1. Biological Membranes

Chapter 2. Thermodynamics of Micelle Formation

Chapter 3. The Fluid Mosaic Membrane

Chapter 4. Membrane Electrostatics

Chapter 5. Specific and Non-Specific Binding

Chapter 6. Permeability and Conductance

Chapter 7. Permeability and Conductance of Electrolytes

Chapter 8. Channels and Excitable Membranes

Chapter 9. Active Transport

Chapter 10. Facilitated Diffusion

Chapter 11. Coupled Transport

Chapter 12. Energy Coupling

Chapter 13. Epithelial Transport


I. Glossary of Symbols

II. Abbreviations

III. Fundamental Constants

IV. Conversion Factors

V. Mathematical Formulae
Glossary of Symbols

A area x distance
C molar concentration, capacitance Z collision factor
c velocity of light z valence
cmc critical micelle concentration
D diffusion coefficient
d derivative γ activity coefficient
E energy, reduction potential ∆ difference
E electric field δ dipole moment
e electronic charge ε dielectric constant
F force εo permittivity of vacuum
F Faraday constant
f frictional coefficient η viscosity, efficiency
f fugacity λ wavelength of light
G Gibbs free energy µ electrochemical potential
g conductance, acceleration of gravity ρ density
H enthalpy σ reflection coefficient
h Planck’s constant ψ electrical potential
I current
J flow
Ki equilibrium constant for reaction i
Kp partition coefficient
k Boltzmann constant
ki rate constant for reaction i
m mass, aggregation number
N Avogadro's number
n amount in moles
P permeability coefficient, pressure, power
Q heat
q charge
R gas constant, resistance
r radius
S entropy
T absolute temperature
t time
u mobility
V volume
v velocity
W work
X mole fraction

Å Angstroms
atm atmospheres k kilo 103
°C degrees centigrade c centi 10-2
cal calories
m milli 10-3
coul coulombs
D Debyes µ micro 10-6
Da Daltons n nano 10-9
eq equivalents p pico 10-12
esu electrostatic units f femto 10-15
F farads (coul/volt) a atto 10-18
g grams
j joules
°K degrees Kelvin
l liters
M moles/liter
m meters
min minutes
mol moles
S Siemens
sec seconds
V volts
Fundamental Constants

Gas constant R = 8.3144 j.°K-1.mol-1

1.9872 cal.°K-1.mol-1
8.3144 x 107 ergs.°K-1.mol-1
0.082054 l.atm.°K-1.mol-1
Boltzmann constant k = 1.38044 x 10-16 erg.°K-1
Avogadro's Number N = 6.0230 x 1023 molecules.mole-1
Ice point To = 273.15 °K
Faraday constant F = 96,490 coul.eq-1
Permittivity of vacuum εo = 8.854 x 10-12 coul.m-1.volt-1
Molar volume, ideal gas, 0°C, 1 atm Vo = 22.4138 l.mol-1
Electronic charge e = 4.80286 x 10-10 esu
1.602 x 10-19 coul
Electron mass me = 9.1083 x 10-28 g
Velocity of light c = 2.997930 x 1010 cm.sec-1
Standard acceleration of gravity g = 980.665 cm.sec-2
Planck's constant h = 6.6252 x 10-27 erg.sec
Volume conversion factor α = 103 l.m-3
Collision factor Z = 1011 M-1.sec-1

Conversion Factors

1 atm = 760 mm = 1.01325 x 106

1 cal = 4.184000 j
1 j = 107 ergs = 1 volt.coul
1 erg = 1
1 ev = 1.60206 x 10-12 erg
1 l.atm = 24.22 cal
1 Å = 10-10 m
1 Debye = 10-18 = 2 x 10-6 coul.m.mol-1
1 kcal/eq = 0.043362 volts
Mathematical Formulae

sinh x = 1/2 (ex - e-x)

sinh x = x + x3/3! + x5/5! + x7/7! + ...

ex = 1 + x + x2/2! + x3/3! + ...

Surface Area
Cylinder (minus ends) A = 2πrh
Sphere A = 4πr2

Cylinder V = πr2h
Sphere V = (4/3) πr3
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 1.1. Biological Membranes

Biological membranes maintain the spatial organization of life. Membranes
defined the boundaries of the first living cells and still work to shield cellular metabolism
from changes in the environment. Membranes prevent undesirable agents from entering
cells and keep needed molecules on the inside. They also organize the interior of
eukaryotic cells by separating compartments for specialized purposes. Membranes are not
static barriers, but active structures. To function effectively, they must selectively pass
molecules, ions, and signals from one side to the other.
The strategy underlying biological membrane function is that the best barrier
between aqueous compartments is a hydrophobic layer. The water-soluble compounds
present within cells and in their environments are not soluble in the lipid milieu of the
membrane and pass slowly or not at all through even a very thin lipid layer. This
mechanism has a number of advantages which life has exploited. First, the lipid bilayer is
a natural structure and assembles spontaneously. Second, the structure is flexible and
allows for growth and movement as well as for the insertion and operation of protein
machinery. Finally, the structure has a low dielectric constant giving the membrane
electrical properties which are used in signalling, transport and energy transduction.
To understand how biological membranes function, we will begin by analyzing
their structure. The structure determines the fundamental properties of fluidity,
permeability, and membrane potential. The origin and characterization of these properties
will be analyzed next. Finally, we will discuss how these properties contribute to the
various functions of biological membranes: signal transduction, energy transduction, and
Life, like all other processes in our universe, obeys the laws of physics and
chemistry. Consequently, the theoretical framework of physical chemistry provides
powerful tools for understanding living systems. Especially important is the requirement
imposed by the second law of thermodynamics: processes must result in a net decrease in
free energy in order to occur spontaneously. Free energy changes govern all metabolic
processes, but they are particularly apparent in common phenomena of biological
membranes. For example, membrane structure is governed by the distribution of
compounds between the hydrophobic interior of the membrane and the aqueous spaces on
either side. Free energy also determines the movement of molecules and ions across
membranes in response to concentration gradients and membrane potentials. Because
membranes have a well defined planar geometry, the mathematics is simpler than it might
otherwise be. Thus, to understand in depth the structure and function of biological
membranes, it is essential to understand and apply principles of physical chemistry. The
purpose of this course is to construct a coherent framework to do that.

Section 1.2. The Fluid Mosaic Model of Membrane Structure

Early on, it was recognized that hydrophobic compounds passed more readily than
water-soluble compounds through biological membranes. This, coupled with the
identification of lipids as a major component, led to the notion that biological membranes
have a hydrophobic character. The calculation (erroneous, as it turns out) that the lipid

Page 1.1
content was twice that needed for a single layer of lipid led to the concept (correct, as luck
would have it) of the lipid bilayer (Gorter and Grendel, 1925; Danielli and Davson, 1935).
Lipids are amphiphilic compounds with a small hydrophilic headgroup attached to long
hydrocarbon chains. In the lipid bilayer, lipids are aligned with the headgroups facing the
water on either surface of the membrane and the hydrophobic hydrocarbons sandwiched in
The lipid bilayer concept did not establish the location of the protein components of
the membrane. Originally, for lack of a better site, the proteins were stuck on to the
membrane surface. This was not tenable, of course, because proteins are responsible for
moving molecules and messages across membranes and they could not perform those
functions without being a integral part of the membrane itself. This realization gave rise to
the concept of integral and peripheral proteins. Peripheral proteins are loosely associated
with the membrane and located on the surface of the lipid bilayer. Integral proteins are
inserted into the membrane and pass all the way (or much of the way) across the
membrane. Originally, the integral proteins were thought to form a well defined matrix
with the lipid bilayer filling in the spaces in between. In the late 1960's, however, it
became clear that many proteins are not rigidly fixed in the membrane, but can diffuse
across the surfaces of cells relatively easily and independently. Membranes came to be
viewed as fluid structures with proteins and lipids arranged in their thermodynamically
most favorable structure. Lipids exist in a bilayer and provide the milieu in which the
integral membrane proteins float. The proteins are oriented so that their hydrophobic
surfaces are immersed in the hydrophobic interior of the lipid bilayer. Hydrophilic amino
acids are exposed only in the aqueous regions on either side of the membrane. The
organization of the membrane is a direct consequence of the partitioning of its components,
both lipid and protein, so that hydrophobic regions are kept within the membrane and
hydrophilic parts are exposed to the water on either side. Because the components are not
held together by bonds, they are free to diffuse and move independently within the plane of
the membrane. Singer and Nicolson (1972) captured this view in the fluid mosaic model.
The fluid mosaic model persists as the accepted view of membrane structure. The
recognition of linkages between membrane proteins and components of the cytoskeletal
system has modified the concept somewhat. The cytoskeleton does impose some
constraints on the distribution of membrane proteins. As we shall see, phase separation
also can create separate domains with different characteristics within membranes.
Consequently, the fluid mosaic model does not imply that all the components of a
particular membrane are randomly and homogeneously distributed.

Section 1.3. Classes of Biological Membranes

Biological membranes perform many different functions in all kinds of cells. They
can be divided, however, into four classes based on differences in their fundamental
energetics. These differences arose during the course of evolution as different organisms
adopted different strategies to cope with their environments.
The first class includes prokaryotic cell membranes, the inner mitochondrial
membrane and the thylakoid membrane of chloroplasts. These membranes, which share a
common evolutionary origin, do not contain cholesterol. A proton gradient drives the
functions of these membranes. The proton gradient is generated by a variety of

Page 1.2
mechanisms, but a redox chain is common. These membranes can use the proton gradient
to generate ATP using an ATP synthase of the F1F0 type.
The second class consists of plasma membranes of animal cells. These have a
Na+/K+ ATPase which pumps Na+ out of and K+ into the cells. The Na+ and K+
concentration gradients created thereby participate in many functions of the membrane
including transport, excitability, and signalling.
The third class of membranes includes the plasma membranes of plant and fungal
cells. These differ from animal cell membranes in that they lack a Na+/K+ ATPase and
instead have a proton-translocating ATPase of the P-type. The proton gradient created by
this ATPase drives transport and other functions of the plant plasma membrane. This
fundamental difference between plant and animal cell membranes reflects a basic
difference between plant and animal lifestyles. Unlike animal cells, plant cells cannot rely
on seawater or a circulatory fluid to provide external Na+, so the substitution of a proton
pump for a sodium pump is necessary. Moreover, because plants are immobile, plant cells
can have a rigid cell wall to support the plasma membrane in times of osmotic imbalance.
Animal cells, by contrast, must maintain osmotic balance by regulating the concentrations
of internal osmolytes such as Na+ and K+ to balance the external salt concentration.
The fourth class of membranes includes membranes of the vacuolar system. This
comprises the membranes of Golgi-derived organelles including lysosomes, endosomes,
secretory vesicles in animal cells and peroxisomes, vacuoles and tonoplasts in plants and
fungi. These membranes have a proton-translocating ATPase of the V-type. The proton
gradient created by this ATPase drives processes in the membrane and also makes the
interior of the organelle acidic, a property frequently important in the function of the

Section 1.4. Membrane Biosynthesis and Asymmetry

Although the lipid bilayer is basically a symmetrical structure, natural membranes
are not. The two sides of the membrane differ, so the membrane has a functional polarity.
Molecules, ions and signals will be moved one way but not the other. The two sides of the
membrane differ because of the way the membrane is synthesized.
In animal and plant cells, the plasma membrane and vacuolar membranes share a
common synthetic pathway. The proteins are inserted into the endoplasmic reticulum. The
membranes are transferred to the Golgi apparatus where carbohydrates are attached and
processed. The membranes are also sorted and leave the Golgi targeted to their final
destinations. This process has a number of consequences. First, the proteins are inserted
into the membranes with a defined orientation; they are inserted from the cytoplasmic side
of the membrane. Second, the carbohydrates are attached only to the other side, the interior
of the Golgi. Thus, the carbohydrates face only the interior of organelles and the exterior
of the cell; they do not appear on the cytosolic side of a membrane. This, along with the
processes of exocytosis and endocytosis, emphasize that the interior of vacuolar organelles
is equivalent to the exterior of the cell in terms of membrane polarity.
In bacterial cells, the proteins and lipids are synthesized inside the cell and inserted
into the membrane. The biosynthesis of the mitochondrial inner membrane and the
thylakoid membrane are more complex because some of the proteins are synthesized in the
cytoplasm and then inserted into the organelle. The end result, however, is that the

Page 1.3
proteins are placed in the membrane with a defined orientation and this gives the
membrane polarity.
The structure of a biological membrane is a consequence of both spontaneous
assembly and programmed development. The fluid mosaic model emphasizes the
spontaneous assembly of the lipid bilayer with the proteins orienting to accommodate their
own hydrophobic and hydrophilic surfaces. At the same time, the membrane is
asymmetrical largely because of its history. The structure of each protein is determined by
the amino acid sequence and the direction in which that sequence was inserted into the
membrane. The further processing of that protein, particularly the attachment of
carbohydrates, is also asymmetrical since the relevant enzymes are confined to one side of
the membrane or the other. The asymmetry is obviously crucial because it gives
membranes a polarity essential for function. Transport occurs in defined directions. This
in turn creates the membrane potential, a polarity that plays a fundamental role in many
membrane processes.

Section 1.5. Summary

Lipids in biological membranes are arranged predominantly in a bilayer structure.
Proteins are oriented so that hydrophobic amino acids are buried inside the membrane and
hydrophilic residues are exposed on the aqueous surfaces. Proteins are inserted with a
defined orientation and give polarity to the membrane. Four classes of biological
membranes may be distinguished on the basis of their different primary ion pumps. All of
these considerations should be kept in mind as we proceed to analyze the structure and
function of biological membranes.


S.J. Singer and G.L. Nicolson (1972) The fluid mosaic model of the structure of cell
membranes, Science 175, 720-731.

Page 1.4
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 2.1. Properties of Water

Water is a remarkable substance in many ways, and its unusual properties are
crucial in the functioning of biological membranes. First, water molecules dissociate into
H+ and OH-. This property allows H+ to equilibrate among protonatable groups in all of the
molecules in a solution. As we shall see, it also makes H+ a convenient ion to use for
creating electrical potential differences across biological membranes.
A second important property of water is its polarity. The dipole moment of the OH
bond is 1.51 Debyes and the water molecule itself has a dipole moment of 1.84 Debyes.
As a consequence, water has a high dielectric constant (80.37 at 20°C) and polarizes to
neutralize electric fields. For this reason, electric fields and associated differences in
electrical potential exist primarily across biological membranes rather than across aqueous
regions of cells.
Finally, water molecules participate in hydrogen bonding. The hydrogen bond is
primarily electrostatic (a dipole-dipole interaction) with an energy of 4.5-6 kcal/mol. The
bond is linear with the hydrogen atom situated directly between two electronegative atoms
(two oxygen atoms in the case of the H2O-H2O bond). Hydrogen bonding accounts for the
relatively low freezing and boiling points of water relative to other compounds with
comparable molecular weights. Most importantly, hydrogen bonding is responsible for the
hydrophobic effect.
Because hydrogen bonds are fairly strong, water molecules will orient so as to
hydrogen bond even if this orientation restricts their mobility. For example, water
molecules on the surface (an air-water interface) will have fewer other water molecules
with which to interact than will molecules in the interior of the solution. Molecules on the
surface will nevertheless hydrogen bond to other water molecules, but the smaller number
of possible bonding partners means that they will have a lower entropy than molecules in
the interior. To increase the entropy of the system, water will minimize its surface area.
This effect is responsible for the high surface tension of water.
Because nonpolar molecules will not hydrogen bond, they reduce the bonding
possibilities of adjacent water molecules. Thus, just as increasing the surface area
decreases the entropy of an aqueous solution, so does introducing nonpolar molecules.
Exclusion of these nonpolar substances from the aqueous solution increases the entropy
and decreases the free energy. It is this entropy-driven effect that causes nonpolar
compounds to be excluded from aqueous solutions. It is important to recognize that
nonpolar molecules do not attract each other; they are pushed together because they are
mutually excluded from water. This phenomenon is known as the hydrophobic effect.

Page 2.1
Figure 2.1

Page 2.2
Section 2.2. Structures Formed by Amphiphilic Molecules
Amphiphiles are those molecules that are polar on one end and nonpolar on the
other. These promiscuous molecules have affinities for both aqueous and nonpolar phases.
At low concentrations, they dissolve in water. At some critical concentration, however,
they reach their solubility limit and begin to aggregate into micellar structures. The micelle
structure allows the molecule to keep its polar region in the aqueous phase on the surface
of the micelle and the nonpolar portion in the nonpolar interior of the micelle. The limiting
monomer solubility is called the critical micelle concentration (cmc). At concentrations
below the cmc, the amphiphile will exist as monomers. At concentrations above this level,
the excess amphiphile will aggregate to form micelles.
Two factors are involved in the spontaneous formation of micelles. First, the
hydrophobic effect causes the nonpolar portion of the molecule to be separated from water
and sequestered in the interior of the structure. Second, interactions between the head
groups determine how closely the molecules may be packed. Amphiphiles with a single
hydrocarbon chain, such as dodecyl sulfate, must pack a number of head groups around a
relatively small volume of hydrocarbon. This large surface area to volume ratio is achieved
by forming a spherical micelle structure. By contrast, amphiphiles with two hydrocarbon
chains, such as phospholipids, must pack the same number of headgroups around twice as
large a volume of hydrocarbon. This smaller surface area to volume ratio is achieved by
forming the bilayer structure (Figure 2.2).

Figure 2.2

It should be recognized that the spherical micelle and the planar bilayer are really
two extremes of a continuum. Micelles in the shape of oblate spheroids will exhibit
intermediate ratios of surface area to volume. Under a particular set of conditions, an
amphiphile will form micelles with a particular surface area to volume ratio and thus will
form micelles of a particular size. This characteristic of the micelle is described by the
aggregation number m, the average number of amphiphile molecules in a single micelle.
The critical micelle concentration and the aggregation number together characterize
the micelle that a particular amphiphile will form under a given set of conditions. We can
make some intuitive generalizations about how these parameters should respond to a
variety of changes. Increasing the chain length of the amphiphile should lower the aqueous
solubility and decrease the critical micelle concentration. For similar reasons, amphiphiles
with two hydrocarbon chains (phospholipids) should have a lower cmc than those with a
single chain (detergents). Ionic detergents should have a greater water solubility than
nonionic detergents and therefore should have a higher critical micelle concentration.
Repulsive forces between polar groups should be less for nonionic detergents than for ionic
detergents. Therefore, nonionic detergents should form micelles with smaller surface areas
per amphiphile (larger aggregation numbers). Increasing the ionic strength should diminish

Page 2.3
repulsive forces between polar groups of ionic detergents thereby increasing the
aggregation number. These characteristics of the cmc and the aggregation number are
illustrated by the data in Table I. Deoxycholate, cholate and sodium dodecyl sulfate are
ionic detergents; Lubrol WX and Triton X-100 are nonionic.


Detergent cmc (mM) m

Sodium dodecylsulfate
50 mM NaCl 2.3 72
500 mM NaCl 0.51 126
10 mM NaCl, pH 7.5 4 4
300 mM NaCl, pH 7.5 6 29
Cholate 45 2
Lubrol WX 0.125 96
Triton X-100 0.24 140

Section 2.3. The Hydrophobic Effect

The thermodynamics of micelle formation has been analyzed in an elegant fashion
by Tanford (1980). The two factors determining micelle structure, the hydrophobic effect
and head group interaction, are each assumed to contribute separately to the free energy of
the micelle. A summary of this analysis will be presented here.
To understand the contribution of the hydrophobic effect to micelle structure, let us
first consider the solubility of hydrocarbons in water. The chemical potential of the
hydrocarbon in the aqueous phase is

(2.1) µw = µw° + RT ln Xw + RT ln fw

We assume that ln fw = 0 because the hydrocarbon concentration in water is extremely

low. Now consider the chemical potential of the hydrocarbon in a pure hydrocarbon phase:

(2.2) µHC = µHC° + RT ln XHC + RT ln fHC

In pure hydrocarbon, XHC = 1 so ln XHC = ln fHC = 0. When the hydrocarbon partitions

between water and the hydrocarbon phase, equilibrium is reached when the chemical
potentials are equal (µw = µHC). Therefore,

(2.3) µHC° = µw° + RT ln Xw

Since Xw is the saturating concentration (in mole fraction) of the hydrocarbon in water, the
free energy change for transferring a hydrocarbon molecule from water into the
hydrocarbon phase can be determined from the compound's water solubility:

Page 2.4
(2.4) µHC° - µw° = RT ln Xw

For a series of n-alkanes, the following empirical relationship is found:

(2.5) µHC° - µw° = -2436 - 884 nc cal/mol

where nc is the number of carbon atoms in the molecule. Of course, as nc increases, the
water solubility of the hydrocarbon decreases. Double bonds increase the water solubility
(decrease the hydrophobicity) of hydrocarbons.
We can use a similar analysis to examine the partitioning of hydrocarbon molecules
between water and the interior of a micelle. The chemical potential of the hydrocarbon
molecule in a micelle is

(2.6) µmic = µmic° + RT ln Xmic

Setting µmic equal to the chemical potential of the hydrocarbon in water (equation 2.1)
allows us to solve for the free energy change for transfer of the hydrocarbon from water to
the interior of the micelle:

(2.7) µmic° - µw° = RT ln Xw - RT ln Xmic = RT ln (Xw/Xmic)

If Xw is the solubility of the hydrocarbon in water, then Xmic can be calculated from the
increase in solubility observed in the presence of micelles. For a series of n-alkanes and
micelles formed from sodium dodecyl sulfate, the following empirical relationship is
(2.8) µmic° - µw° = -1934 - 771 nc cal/mol

As in the case of the transfer of hydrocarbon from water to the pure hydrocarbon phase, the
free energy change is proportional to the number of carbon atoms in the compound and the
energy contribution of each carbon atom is about 800 cal/mol.

Section 2.4. Thermodynamics of Single-Component Micelles

There is a dynamic tension at work in the micelle structure. The polar groups tend
to repel each other because they have similar charges and dipole moments. Nevertheless,
they must remain close enough together to prevent water from gaining access to the
hydrophobic interior of the micelle.
Let us consider the chemical potential of an amphiphile in water (µw) and in a
micelle of size m (µmic,m).

(2.9) µw = µw° + RT ln Xw + RT ln fw
(2.10) µmic,m = µmic,m° + (RT/m) ln (Xm/m)

RT ln (Xm/m) is the contribution of the whole micelle to the free energy, so this term is
divided by m to determine the free energy contribution of each molecule of amphiphile.

Page 2.5
Since amphiphiles will equilibrate between the aqueous phase and micelles, these chemical
potentials must be equal. Therefore,

(2.11) µmic,m° - µw° = RT ln Xw + RT ln fw - (RT/m) ln (Xm/m)

If the aggregation number m is large or the mole fraction of amphiphile in micelles Xm is

small, then the final term can be ignored. This allows us to calculate the cmc knowing
µmic,m° - µw°.

(2.12) µmic,m° - µw° = RT ln Xw = RT ln cmc

Alternatively, equation 2.11 can be solved for ln Xm

(2.13) ln Xm = -(m/RT)(µmic,m° - µw°) + m ln Xw + m ln fw + ln m

This allows us to calculate the concentration of amphiphile in micelles of aggregation

number m at any given aqueous amphiphile concentration Xw. Again, we must know
µmic,m° - µw°.
To analyze µmic,m° - µw°, we will separate this energy into two parts: that
contributed by the hydrophobic effect (Um° - µw°) and that caused by head group
repulsion (Wm).

(2.14) µmic,m° - µw° = Um° - µw° + Wm

Tanford determines each of these components semiempirically, although the expressions

can be rationalized to some extent. The contribution of the hydrophobic effect is specified

(2.15) Um° - µw° = -2100 - 700 (nc -2) + 25 (A-21) + Constant

This expression attributes 700 cal/mol to each carbon atom in the hydrocarbon chain in
rough agreement with the values given in Section 2.3. The hydrophobic effect is also
assumed to depend linearly on the surface area of the micelle per molecule of amphiphile
(A). A is measured in square angstroms, so each square angstrom of hydrophobic surface
area per molecule adds about 25 cal/mol to the free energy.
This dependence of the free energy on micelle surface area per molecule is crucial
since it helps determine the size of the micelle. While the volume of the micelle increases
in proportion to the aggregation number m, the surface area increases only in proportion to
m2/3. Therefore, larger micelles will have a smaller surface area per molecule and smaller
micelles will have a larger surface area per molecule. Under a given set of conditions, an
amphiphile molecule will have an optimum surface area, and this will determine the size of
the micelle that it will form.
Having estimated the contribution of the hydrophobic effect to the free energy of
micelle formation, we must estimate the contribution of the head group interaction (Wm).
This may be calculated from pressure vs. area curves measured for monolayers of
amphiphile. If an amphiphile is added to water, it will form a monolayer on the surface of
the water with the polar end in the water and the nonpolar region extending into the air.

Page 2.6
The pressure required to compress this monolayer will depend on the repulsive forces
between the head groups (Phg) and on the intrinsic pressure exerted by ideal molecules by
virtue of their kinetic energy (Pke). The latter pressure is

(2.16) Pke = kT/A

where A is the surface area per molecule and k is the Boltzmann constant. This expression
is the two-dimensional correlate of the ideal gas law. The pressure attributable to head
group interaction is therefore

(2.17) Phg = P - (kT/A)

The work done against this head group repulsion is

(2.18) Wm = - ∫ (P-kT/A) dA

This work function can be evaluated by integrating pressure vs. area curves determined
from monolayer compression experiments and corrected for kT/A (Figure 2.2). For the
trimethylammonium and sulfate head groups, the work functions as evaluated by Tanford

(2.19) Wm = 1.51 x 105/A - 8.3 x 104/A2 - 2.4 x 107/A3

(2.20) Wm = 1.07 x 105/A + 5.4 x 105/A2 - 3.6 x 107/A3

P (erg/cm 2 )



40 60 80 100
A (Å 2 )

Figure 2.3

We now have semiempirical expressions that can be used to calculate the free
energies involved in the formation of dodecyl sulfate and cetyltrimethylammonium

Page 2.7
micelles. Table II compares experimental results to results of Tanford's semiempirical
calculations based on equations 2.12 and 2.13. If head group interactions are calculated
theoretically (Debye-Huckel theory), agreement with experimental results is not as good.

m cmc (M)
N+(CH3)3 59 0.0066
OSO3- 95 0.0015
Semiempirical (Monolayer data)
N+(CH3)3 60 0.0062
OSO3 93 0.0018
Theoretical (Debye-Huckel)
OSO3- 39 0.0042

The hydrophobic effect and the head group interaction have both been cast as
functions of the micelle surface area per molecule (equations 2.15, 2.19 and 2.20). It is
instructive to plot these energies as functions of the surface area (Figure 2.4). This shows
that there is a molecular surface area which minimizes the total free energy. The surface
area giving this minimum free energy determines the aggregation number of the micelle. A
larger surface area per molecule implies smaller micelles. A smaller surface area per
molecule implies larger micelles with the bilayer structure being the limiting case
(infinitely large micelle). The minimum free energy itself is RT ln cmc (equation 2.12).
This plot illustrates effects of changes in the energies. For example, decreasing the
work function by increasing the ionic strength will lower the cmc and increase the
aggregation number (Figure 2.5).

Page 2.8
µ (kcal/mol) 2


-4 U° mic,m - µ°w
RT ln cmc
U°m - µ°w
50 75 A min 100
larger smaller
micelles 2 micelles
A (Å )

Figure 2.4

W m (low ionic strength)

µ (kcal/mol)

0 Wm (high ionic strength)

U° mic,m - µ°w

U° m- µ° w
50 100

A (Å 2 )

Figure 2.5

Page 2.9
Lauryl Sulfate (Dodecyl sulfate)

O 3S O CH2 (CH2 )10 CH 3

Cetyltrimethylammonium (CTAB)

CH 3

H3C N+ CH 2 (CH2 )14 CH 3

CH 3

Cholic Acid: X = OH
Deoxycholic Acid: X = H

CH 3
CH 3

H CH 3



Fig 2.6. Some ionic detergents

Page 2.10
Triton X-100 (Octylphenoxypolyoxyethanol)

HO (CH2 -CH2 -O)10 (CH2 )7 CH3

Lubrol W

HO (CH2 -CH2 -O)7 (CH2 )15 CH3

Fig. 2.7. Some nonionic detergents


Stillinger, F.H. (1980) Water revisited, Science 209, 451-457.

C. Tanford (1974) Theory of micelle formation in aqueous solutions, J. Phys. Chem. 78,
C. Tanford (1974) Thermodynamics of micelle formation: Prediction of micelle size and
size distribution, Proc. Natl. Acad. Sci. USA 71, 1811-1815.
C. Tanford (1977) The hydrophobic effect and the organization of living matter, Science
200, 1012-1018.
C. Tanford (1979) Interfacial free energy and the hydrophobic effect, Proc. Natl. Acad.
Sci. USA 76, 4175-4176.
C. Tanford (1980) The Hydrophobic Effect, Second Edition, John Wiley & Sons, New

Page 2.11

Section 3.1. Characteristics of Lipid Bilayers

Naturally occurring phospholipids have a very low critical micelle concentration.
For example, the cmc for dipalmitoyl phosphatidylcholine is 4.7 x 10-10 M. Therefore,
phospholipids will virtually always form into a bilayer structure. The lipid bilayer has a
thickness of approximately 40 Å, determined principally by the chain length of the fatty
acids in the phospholipids. Each phospholipid molecule occupies a surface area of about
70 Å2.
As in a micelle, the surface area occupied by each phospholipid molecule in the
lipid bilayer is determined by the balance between head-group interactions and the
hydrophobic effect. If the head groups favor greater separation than the hydrocarbon
chains will permit, then the hydrocarbon chains may tilt so that they are not aligned
perpendicular to the surface of the membrane. This decreases the thickness of the bilayer
and increases the cross-sectional area occupied by each fatty-acid chain. If the head-groups
tend to pack more tightly than the hydrocarbon chains, the fatty acid chains will be aligned
perpendicular to the plane of the membrane and there may also be a force on the bilayer
favoring formation a concave bend.
The hydrocarbon chains in diacyl phospholipids undergo a phase transition from an
ordered (crystalline) to a disordered (fluid) state. Some phase transition temperatures are
given in Table 3.1. Increasing unsaturation and decreasing fatty acid chain length lower
the phase transition temperature. Cholesterol generally causes the phase transition to

Table 3.1. Phase transition temperatures for the transition from ordered to disordered
hydrocarbon chains in diacyl phospholipids in hydrated multilayers

Phospholipid Transition temperature (°C)

diC22 phosphatidyl choline 75
diC18 phosphatidyl choline 54.9
diC16 phosphatidyl choline 41.4
diC14 phosphatidyl choline 23.9
diC18:1 phosphatidyl choline -22
diC16 phosphatidyl serine (low pH) 72
diC16 phosphatidyl serine (high pH) 55
diC16 phosphatidyl ethanolamine 60
diC14 phosphatidyl ethanolamine 49.5

occur over a broader range of temperatures. As described by the fluid mosaic model, the
lipids in a biological membrane are typically in a fluid state at physiological temperatures.
In the ordered state, the fatty acid chains occupy less volume than in the fluid state. In the
case of phosphatidylcholine, this means that the hydrocarbon chains must tilt to achieve
adequate separation of the head groups.

Page 3.1
Bilayers composed of mixed lipids may exhibit phase separation. If the chemical
potentials of the individual lipids are higher in the mixed phase than in homogeneous
phases, then the homogeneous phases will separate out. This can be detected as a spatial
separation of different lipids or as domains having different characteristics (e.g., fluidity).
A striking example of this is the ripple phase exhibited by phosphatidylcholine bilayers. At
low temperatures, the bilayer forms a homogeneous ordered phase and, at high
temperatures, it forms a single fluid phase. At intermediate temperatures, however, the
bilayer exhibits a periodic pattern showing an undulating pattern or ripples on its surface.
The molecular basis for this is not yet clear, but some inferences can be made. In the
intermediate temperature range, phospholipids with tilted chains coexist with
phospholipids with extended chains. The differences in chain tilting and bilayer thickness
discourage intermixing of phospholipids in different phases and causes the two groups to
segregate (Marden et al., 1984).
Lipids in natural membranes are characteristically distributed asymmetrically in the
two halves of the bilayer. For example, human red cells are mostly PC on the outside and
mostly PE on the inside (Table 3.2). This asymmetry can be maintained because
phospholipids are very slow to redistribute (flip-flop) between the two sides of a bilayer.
The original cause of the asymmetry may lie in the biosynthetic history of the membrane,
but there also appear to be ATPases that invest cellular energy in the transport of
phospholipid head groups across a variety of natural membranes (Devaux, 1992).

Table 3.2. Distribution of lipids in natural membranes

Outer Monolayer Inner Monoayer

Human RBC 40 42 10 0 8 13 45 25
Rat RBC 42 35 15 0 6 20 40 25
Bovine ROS - 10 40 40 - >80 10 10
Values are percentages of total lipid in that layer of the membrane.
RBC = red blood cell; ROS = rod outer segment; SM = sphingomyelin.

Section 3.2. Model Lipid Membranes

Because phospholipids naturally form bilayer structures (at least at low lipid:water
ratios), artificial membranes can be produced in a number of ways. Phospholipid vesicles
are easily made by sonicating lipids (Huang, 1969), by reverse phase vaporization (Deamer
and Bangham, 1976), or by dialyzing away detergent (Milsmann et al., 1978). Sonication
is convenient but produces rather small unilamellar vesicles so the membranes have a high
radius of curvature. Reverse phase vaporization produces larger vesicles but the solvent in
which the lipids are initially dissolved contaminates the preparation and may alter
permeability and other properties of the bilayer. Sonication and reverse phase vaporization
are both rather harsh treatments for proteins, so reconstitution of membrane proteins is
generally accomplished by some variation of the dialysis method.

Page 3.2
Measuring the electrical properties of membranes requires a planar membrane
separating two aqueous spaces large enough to accommodate electrodes. Planar bilayers
can be made by spreading lipid in solvent over an aperture (Mueller et al., 1963) or by
raising two monolayers past the aperture (Montal and Mueller, 1972). This produces
artificial membranes with much less total surface area than a liposome suspension, but it
does permit electrical recording of artificial membrane properties. In the solvent/aperture
method, the solvent in which the lipid is dissolved moves to the rim of the aperture and the
bilayer across the opening thins out to form a "black lipid membrane." Nevertheless, the
solvent can form lenses in the artificial membrane and there is always some question about
the influence of remaining solvent on the properties of the membrane. The monolayer
method corrects this. Even using great care, planar membranes formed by either method
are relatively unstable and their short lifetime is an experimental handicap.
To study membrane proteins, liposomes formed by dialysis have proven most
convenient. The proteins can be solubilized in the detergent and then conveniently
reconstituted. Monitoring effects of these proteins on electrical properties of the
membrane is not possible, however, because liposomes are too small to insert electrodes.
To overcome this problem a number of new approaches have been tried. First, liposomes
containing the reconstituted protein may be fused into a black lipid membrane (Miller and
Racker, 1976). Patch clamping technology has introduced a new approach. A patch
pipette can be raised through a lipid monolayer on the surface of a solution and then
lowered back down. A planar bilayer forms across the opening of the patch pipette and this
bilayer will contain proteins dispersed in the lipid monolayer (Tank et al., 1982; Suarez-
Isla et al., 1983).

Page 3.3
Section 3.3. Lipid Components
Fatty acids: Fatty acids have two characteristics that affect the physical properties
of the membrane: chain length and degree of unsaturation. Fatty acids may be identified
by a common name, by the standard nomenclature, or by the w nomenclature. According
to the standard nomenclature, fatty acids are represented as x:y, z1,z2 ... zn where x
represents the number of carbon atoms, y represents the number of double bonds, and z1,z2
... zn represent the carbon atoms preceding double bonds counting from the carboxy end.
According to the w nomenclature, fatty acids are represented as x:yωz' where x represents
the number of carbons, y represents the number of double bonds, and ωz' represents the
position of the first double bond counting from the ω carbon (methyl terminus). For
example, the structure of palmitoleic acid (16:1, 9-cis or 1ω7) is:


TABLE 3.3. Some Common Fatty acids

Saturated Unsaturated
No. of Common Common Nomenclature
Carbons Name Name Standard ω
10 Capric Palmitoleic 16:1, 9-cis 1ω7
12 Lauric Oleic 18:1, 9-trans 1ω9
14 Myristic Vaccenic 18:1, 11-cis 1ω7
16 Palmitic Linoleic 18:2, 9-cis,12-cis 2ω6
18 Stearic α-Linolenic 18:3, 9-cis,12-cis,15-cis- 3ω3
20 Arachidic γ-Linolenic 18:3, 6-cis,9-cis,12-cis- 3ω6
22 Behenic Arachidonic 20:4, 5,8,11,14 (all cis) 4ω6
24 Lignoceric

Phospholipids: Phospholipids have the general structure shown below:

Headgroup-O-P-O- CH2 O
O CH-O-C-CH2...CH3

Page 3.4
The head groups - which differ in charge, polarity and reactivity - give the phospholipids
different characteristics:
Phosphat idylcholine (PC or lecit hin)

CH3 O-

CH3 N+ CH2 CH2 O P O


Phosphat idylet hanolamine (PE)


NH3 + CH2 CH2 O P O

Phosphat idylser ine (PS)


NH3 + CH CH2 O P O


Phosphat idylinosit ol (PI)




Phosphat idic A cid (PA )



Page 3.5
Cholesterol: Cholesterol is present in plasma membranes, lysosomes and storage
granules. The cholesterol content of the Golgi complex increases on moving from the cis
to the trans cisternae. Cholesterol is not present in bacterial, inner mitochondrial or
chloroplast thylakoid membranes.

20 22

12 18
11 23
13 25
CH3 C D 16 CH3
1 19 26 27
2 9 14 CH3
10 8
HO 3 5
4 6


Sphingomyelin: The sphingomyelin content is high in lysosomes and storage

granules: rat liver lysosomes, 24%; bovine chromaffin granules,15%; serotonin granules
from pig platelets, 24.9%; bovine pituitary neurosecretory vesicles, 21.7%.

Section 3.4. Structure of Membrane Proteins

Because they are typically hydrophobic, membrane proteins have been notoriously
difficult to study using traditional protein chemistry techniques. The advent of molecular
biology has made it far easier to clone and sequence a membrane protein’s gene than to
study the protein itself. The usefulness of this approach depends to a large extent on how
much we can deduce about the structure of the protein from its amino acid sequence.
Three-dimensional structures of membrane proteins are difficult to determine, but those
that have been established seem to follow one of two patterns: the helix bundle which
includes most integral membrane proteins and the beta barrel which includes some proteins
in the outer membrane of gram negative bacteria and the outer mitochondrial membrane
(von Heijne, 1994). The helix bundle type follow the pattern set by bacteriorhodopsin; the
membrane-spanning segments are both hydrophobic and alpha-helical. These membrane-
spanning segments are separated by hydrophilic domains that are exposed to the aqueous
environments at either membrane surface. The α-helical nature of the membrane spanning
regions can be rationalized. Hydrophobic side chains will tend not to interact with each
other or with lipid components of the membrane. That means that the structure of a
hydrophobic domain will be determined primarily by the backbone hydrogen bonding
pattern. Accordingly, the a-helix is the expected pattern. The α-helix consists of 3.61
amino acids per turn spanning a distance of 1.5 Å per residue. Given a membrane
thickness of 30-40 Å, a transmembrane a-helix should contain 20-27 amino acid residues.

Page 3.6
To predict which segments in a protein are membrane-spanning regions, it has
become common to use a hydropathy index. The most widely used is that described by
Kyte and Doolittle (1982). A variety of thermodynamic parameters could be used to assign
a hydropathy value to each amino acid. The validity of any particular choice may be
debated, but the only important consideration is that some consensus index of hydropathy
is established for each amino acid side chain. Kyte and Doolittle based their hydrophathy
index on the water-vapor transfer free energies of the side chains and on the interior-
exterior distribution of amino acid side chains. The hydropathy value of a span of amino
acids is then determined by summing the hydropathy indices for each amino acid in that
span. It is convenient to choose an odd number for the span length so the hydropathy index
can be associated with the amino acid in the middle of the span. For example, if a span of
7 is chosen, the hydropathy value at amino acid 40 is the sum of the hydropathy indices of
amino acids 37-43.

Table 3.4. Kyte-Doolittle Hydropathy Scale

Amino Acid Single Kyte-Doolittle

Residue Letter Code Hydropathy Index
Isoleucine I 4.5
Valine V 4.2
Leucine L 3.8
Phenylalanine F 2.8
Cysteine/cystine C 2.5
Methionine M 1.9
Alanine A 1.8
Glycine G -0.4
Threonine T -0.7
Tryptophan W -0.9
Serine S -0.8
Tyrosine Y -1.3
Proline P -1.6
Histidine H -3.2
Glutamic acid E -3.5
Glutamine Q -3.5
Aspartic acid D -3.5
Asparagine N -3.5
Lysine K -3.9
Arginine R -4.5

Section 3.5. Solubilization and Reconstitution of Membrane Proteins

Integral membrane proteins, by nature, have hydrophobic surfaces that allow them
to penetrate into the hydrophobic center of lipid bilayers. To get these proteins out of a
membrane, therefore, these hydrophobic surfaces must be protected by detergent molecules

Page 3.7
or the protein will denature. The strategy used to solubilize membrane proteins is to add
detergent to the membrane. The detergent intercalates into the lipid bilayer forming a
mixed micelle with the phospholipids. When enough detergent has entered the membrane,
the structure changes from the bilayer favored by phospholipids to the oblate or spherical
micelles favored by the detergent. As this happens, the bilayer structure breaks down and
the proteins escape into soluble particles consisting of the protein along with detergent and
residual phospholipid. In principle, any detergent can be used to break down the lipid
bilayer and solubilize membrane proteins. In practice, however, strong detergents may also
enter into the protein itself and denature it. Dodecyl sulfate is a good example. It is useful
in gel electrophoresis because it disrupts the secondary and tertiary structure of a protein
causing it to migrate on the gel according to size alone. Dodecyl sulfate definitely
solubilizes membrane proteins, but it also denatures them destroying their structure and
Many membrane proteins (e.g., channels, transporters) have no assayable function
after they are solubilized and removed from the membrane. In order to study their
functions, they must be reconstituted into some sort of a membrane structure.
Reconstitution is typically a matter of simply reversing the solubilization process.
Phospholipids are reintroduced, detergent is removed, and the particle in which the protein
is situated changes from a micelle back into a bilayer. The objective here is to remove the
detergent but not the phospholipid. This can be accomplished by dialysis or gel filtration.
Because phospholipids have an extremely low critical micelle concentration, the rate at
which free phospholipids are removed is extremely slow. Obviously, the same will be true
for a detergent with a low cmc, making such a detergent (Triton X-100 is an example)
difficult to use in reconstitution. A detergent with a higher cmc can be removed relatively
rapidly, however, permitting reconstitution to occur successfully. The bile acids, cholic
acid and deoxycholic acid, have been particularly useful in solubilization and reconstitution
of membrane proteins. They have a high cmc, making it easy to remove them by dialysis.
Because they are ionic, their effectiveness as detergents depends on the ionic strength, and
the salt concentration can be manipulated to shift the balance between solubilization and
reconstitution. The bile acids also have a structure that is particularly suited to solubilizing
membrane proteins without denaturing them. These molecules have a generally planar
structure with a polar side and a nonpolar side. The nonpolar side protects the hydrophobic
protein surface while the other side faces the water. At the same time, this asymmetrical
polarity does not favor penetration by the detergent into the hydrophobic core of the


D. Deamer and A.D. Bangham (1976) Large volume liposomes by an ether vaporization
method, Biochim. Biophys. Acta 443, 629-634.
P.F. Devaux (1992) Protein involvement in transmembrane lipid asymmetry, Ann. Rev.
Biophys. Biomol. Struct. 21, 417-439.
C.H. Huang (1969) Studies on phosphatidylcholine vesicles. Formation and physical
characteristics, Biochemistry 8, 344-352.
J. Kyte and R.F. Doolittle (1982) A simple method for displaying the hydropathic character
of a protein, J. Mol. Biol. 157, 105-132.

Page 3.8
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000
M. Marder, H.L. Frisch, J.S. Langer, and H.M. McConnell (1984) Theory of the
intermediate rippled phase of phospholipid bilayers, Proc. Natl. Acad. Sci. USA 81,
C. Miller and E. Racker (1976) Ca2+-induced fusion of fragmented sarcoplasmic reticulum
with artificial planar bilayers, J. Memb. Biol. 30, 283-300.
M.H.W. Milsmann, R.A. Schwendener and H.G. Weder (1978) The preparation of large
single bilayer liposomes by a fast and controlled dialysis, Biochim. Biophys. Acta
512, 147-155.
M. Montal and P. Mueller (1972) Formation of bimolecular membranes from lipid
monolayers and a study of their electrical properties, Proc. Natl. Acad. Sci. USA 69,
P. Mueller, D.O. Rudin, H.T. Tien, and W.C. Wescott (1963) Methods for the formation of
single bimolecular lipid membranes in aqueous solution, J. Phys. Chem 67, 534-
J.R. Silvius (1992) Solubilization and functional reconstitution of biomembrane
components, Ann. Rev. Biophys. Biomol. Struct. 21, 323-348.
B.A. Suarez-Isla, K. Wan, J. Lindstrom, and M. Montal (1983) Single-channel recordings
from purified acetylcholine receptors reconstituted in bilayers formed at the tip of
patch pipets, Biochemistry 22, 2319-2323.
D.W. Tank, C. Miller, and W. Webb (1982) Isolated-patch recording from liposomes
containing functionally reconstituted chloride channels from Torpedo electroplax,
Proc. Natl. Acad. Sci. USA 79, 7749-7753.
G. von Heijne (1994) Membrane proteins: From sequence to structure, Ann. Rev. Biophys.
Biomol. Struct. 23, 167-192.

Page 3.9
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 4.1. Electrostatics in One Dimension

To understand the electrostatics of biological membranes, one must understand the
concepts of charge, electric field and electrical potential. This section, therefore, reviews
some basic concepts of electrostatic theory. The units to be used in this discussion require
some attention because the choice of units determines the nature of the constants appearing
in the equations. Physics textbooks commonly employ cgs units (statcoulombs, ergs, etc.).
Most of the practical measurements made in biological research, however, are in
rationalized MKS units (coulombs, volts, etc.). For that reason, rationalized MKS units
will be used throughout this discussion.
Electrostatic forces are so strong that positive and negative charges are generally
paired and overall electroneutrality is strictly observed in nature. If a positive and negative
charge are separated, there will be an attractive force between them. According to
Coulomb's Law, this force will be proportional to the absolute magnitudes of the two
charges (q1 and q2) and to the inverse square of the distance between them (x):

(4.1) F = (q1q2)/(4πεox2)

εo, the permittivity of a vacuum, is a constant equal to 8.854 x 10-12 coul/m.volt.

Increasing the distance between the two charges will require us to do work, which
(following the usual physical definition) is

(4.2) W = - ∫ F dx

If a charge is located in the vicinity of many other charges, the force on it will be the sum
of the forces exerted by the other charges. In such a case, it is often more convenient to
define the electric field at a point x as the force exerted by all of the charges in the
neighborhood on a unit charge at that point. Thus, the electric field created by a point
charge q at a distance x is

(4.3) E(x) = q/4πεox2

The electrical potential ψ(x) is the work that must be done to move a unit charge from a
reference point to a point x.
(4.4) ψ(x) = - ∫ ref
E dx

Because we are interested in the electrical properties of biological membranes, we

can take advantage of their planar symmetry. All points the same distance from the surface
of the membrane will have the same electric field and the same electrical potential, so the
dimension perpendicular to the membrane surface is the only significant one. We will
consider charge not to be distributed as points in space but to be spread smoothly on planes
parallel to the membrane. If positive charge is smoothly distributed on a plane of charge,

Page 4.1
as in Figure 4.1A, it will create an electric field perpendicular to the plane. The electric
field will be proportional to the charge density on the plane (qs):

(4.5) E = qs/2εo

If we have two parallel planes of charge, one with a positive surface charge density
(+qs) and one with an equal density of negative charge (-qs), their electric fields will be
additive in the space between them and will cancel in the spaces on either side (Figure
4.1B). Between the planes, therefore, the total electric field will be qs/εo; outside of that
space, the total electric field will be zero. This arrangement is the common textbook
example, the parallel plate capacitor. Because there is an electric field between the two
plates of the parallel plate capacitor, there will be a difference in electrical potential
x x
(4.6) ∆ψ = - ∫0 E dx = - ∫0 (qs/εo) dx = - qsx/εo

This is the potential energy that must be expended to move a unit charge from plate 1 (at 0)
to plate 2 (at x).

+q s +q s -q s

-q s +q s qs
E = E = E =0 E = E =0
2 ε 2 ε ε
ο ο ο

0 x

Single plane of charge Parallel plate capacitor

Figure 4.1. Electric fields associated with planes of charge

The preceding equations apply to charges arranged in a vacuum. In the real world,
charges will be separated by a medium composed of polarizable molecules. These
molecules will tend to align with the electric field and cancel it. The electric field will be
diminished in proportion to the polarizability of the medium, expressed as the dielectric
constant ε. Therefore, if the plates of a parallel plate capacitor are separated by a medium
of dielectric constant ε, the electric field between the plates will be

(4.7) E = qs/εεo

Page 4.2
Of course, the electrical potential difference between the plates will be correspondingly
smaller as well:

(4.8) ∆ψ = - qsx/εεo

Water, which has a dielectric constant of 80, is very polarizable. For that reason, charge
separation across water will tend to create a relatively small electric field and small
electrical potential differences. By contrast, hydrocarbons such as hexane (ε = 1.89) have
a very low dielectric constant. For that reason, charge separation across organic phases
(such as the interior of biological membranes) will create large electric fields. It is this low
dielectric constant that makes it possible to create significant electrical potential
differences across biological membranes.

Section 4.2. Membrane Potential

Separation of electrical charges by biological membranes creates an electrical
potential difference across the membrane. This total difference in electrical potential,
commonly called the membrane potential, plays a crucial role in many membrane
functions. It is the force that drives ions across the membrane. Because it defines the
electrical energy lost when an ion crosses the membrane, the membrane potential partly
determines the energy stored in ion concentration gradients. Finally, the electric field
associated with the membrane potential acts on dipolar groups in membrane proteins and
may regulate the activities of these proteins. Voltage-dependent channels, for example,
open and close in response to changes in the membrane potential.
It is conceptually helpful to divide the charge separation created by biological
membranes into three components. First, charges may be separated by moving ions all the
way across the membrane from the aqueous medium on one side to the aqueous medium
on the other. This separation of capacitative charge is probably the most important
component of the membrane potential. Second, fixed charge bound to the membrane
surface will be neutralized by counterions present in the adjacent aqueous solution.
Because these counterions will tend to diffuse away from the membrane surface, there will
be a consequent charge separation. This leads to the surface potential. Finally, because the
ester linkages between fatty acids and the glycerol backbones of the membrane lipids are
dipolar in character, alignment of these dipoles creates a charge separation which gives rise
to the dipole potential.

Page 4.3
Outside Membrane Inside

created by

potential (outside) dipole membrane
potential potential

potential (inside)

Figure 4.2. Components of the membrane potential

When charge is moved from one side of a biological membrane to the other, we
assume that the net charge on one side is equal and opposite to the net charge on the other.
This is strictly required to maintain overall electroneutrality. These equal and opposite
charges separated by the low dielectric medium of the lipid membrane form an electrical
arrangement like the parallel plate capacitor. We will call the charge transferred across the
membrane the capacitative charge qc. The separation of this capacitative charge will create
an electric field in the membrane

(4.9) E = qc/Aεεo

and an electrical potential difference across the membrane

(4.10) ∆ψ = qcx/Aεεo

The capacitance C is the ratio of the capacitative charge to the potential difference:

(4.11) C = qc/∆ψ = Aεεo/x

The capacitance, therefore, is proportional to the area of the membrane and inversely
proportional to the thickness. For a biological membrane, the capacitance is typically

Page 4.4
about 1 µF/cm2. This corresponds to a membrane 35 Å thick with a dielectric constant of
The total membrane potential will include contributions from the capacitative
charge, from the surface charge on each side of the membrane, and from the surface dipole
on each side of the membrane (Figure 4.2). Because the contributions of surface charge
and surface dipole on one side of the membrane will tend to cancel the contributions of
those components on the other side, the capacitative charge is generally the primary
determinant of the total membrane potential.

Section 4.3. Surface Potential

When fixed charges are bound to a surface, and the counter ions are dissolved in the
adjacent solution, the counter ions will tend to move away from the surface because of
diffusion. This creates a charge separation and consequently an electrical potential
difference called the surface potential. Whereas the relationship between the capacitative
charge and the membrane potential is a simple proportionality (the capacitance), the
relationship between the surface charge and the surface potential is quite complex. The
magnitude of the surface potential depends on the amount of fixed surface charge and also
on the separation between the fixed charge and the diffusable charge. The charge
separation depends on a dynamic tension with diffusion pushing the counterions away from
the surface and electrical attraction pulling them toward the surface. The theoretical
analysis of surface potentials and surface charge was originally developed by Gouy and
Chapman. Just as the DeBye-Hückel theory describes ion distributions as a function of
distance from a fixed point charge, the Gouy-Chapman theory describes ion distributions as
a function of distance from the membrane surface. The analysis rests on three basic
principles: the Boltzmann distribution, the Poisson equation, and electroneutrality.

qs q v (x)

ψ (∞) = 0

ψ (x)


0 x

Figure 4.3. The electrical potential as a function of distance

from the surface.

Page 4.5
First, we imagine a plane of fixed charge with a surface charge density qs. Adjacent
to it is a medium with a space charge density qv(x) where x is the distance from the surface
(Figure 4.3). Overall electroneutrality requires that the total space charge be equal and
opposite to the sum of the fixed charge qs and the capacitative charge qc where the
capacitative charge is the net charge on the other side of the membrane. As we shall see,
the capacitative charge is usually small compared to the surface charge. For simplicity, we
shall simply include it as part of qs:

(4.12) - ∫0 qv(x) dx = qs

The space charge at any point x is obtained by summing over all of the ionic species
present in the solution:

(4.13) qv(x) = Fα Σi ziCi(x)

For each ion species i, zi is the valence and Ci is the concentration in moles/liter. F is the
Faraday constant and α is the constant 103 l/m3. Since the ion concentrations should
depend on the electrical potential according to the Boltzmann distribution:

(4.14) Ci(x) = Cio exp(-ziF ψ(x)/RT)

where Cio is the concentration of ion species i far from the membrane (x=∞). The
Boltzmann distribution (eq. 4.14) together with equation 4.13 relates qv(x) to ψ(x).

(4.15) qv(x) = Fα Σi ziCio exp(-ziF ψ(x)/RT)

Another relationship between qv(x) and ψ(x) is provided by Poisson's equation:

(4.16) qv(x)/εεo = -d2ψ/dx2

Poisson's equation can be rationalized by considering the single plane of charge (Figure
4.1A). If the plane is assumed to have a thickness dx, the change in the electric field dE/dx
upon crossing the plane is qs/εεo. If equation 4.4 is differentiated twice, it is apparent that

(4.17) d2ψ/dx2 = -dE/dx = -qs/εεo

Poisson's equation (eq. 4.16) can be integrated in two ways. First, using eq. 4.12,
∞ ∞
(4.18) qs = - ∫0 qv(x) dx = εεo ∫0 (d2ψ/dx2) dx

We define ψ(∞) = 0, so

(4.19) qs = εεo (dψo/dx)

Page 4.6
Second, using eq. 4.15 and 4.16,

(4.20) d2ψ/dx2 = - (Fα/εεo) Σi ziCio exp(-ziFψ/RT)

To integrate equation 4.20, we multiply both sides by 2dψ/dx. We can then integrate this
equation to obtain

(4.21) (dψ/dx)2 = (2RTα/εεo) Σi Cio exp(-ziFψ/RT) + Constant

At x = ∞, ψ = 0 and dψ/dx = 0, so Constant = -(2RTα/εεo) Σi Cio and equation 4.21


(4.22) (dψ/dx)2 = (2RTα/εεo) Σi Cio [exp(-ziFψ/RT) - 1]

Combining this with equation 4.19 gives

(4.23) qs2 = 2RTαεεo (Σi Cio [exp(-ziFψo/RT) - 1])

Equation 4.23 is the general form of the Gouy-Chapman equation relating the
surface charge to the surface potential. A more convenient form is obtained by assuming
that all of the ions in the aqueous phase are univalent. We will let Cio = Co for both anions
and cations. Then equation 4.23 simplifies to

(4.24) qs = (8RTCoαεεo)1/2 [sinh (Fψo/2RT)]

It is apparent that the surface potential depends on the magnitude of the surface
charge (Figure 4.4) and also on the ionic strength of the aqueous medium (Figure 4.5).
Clearly, the counterions necessary to neutralize the fixed charge will have a much greater
effect on the ion concentration near the membrane surface if the ion concentration is low.
At low ionic strength, therefore, the concentration gradient will be greater and diffusion
forces will be stronger. This will drive the counterions farther from the surface leading to a
greater charge separation and a larger potential difference.
Another way of visualizing this is to simplify equation 4.24 by expanding the sinh
as a power series and dropping higher order terms (assume ψo is small). Equation 4.24
then reduces to

(4.25) qs = (8RTCoαεεo)1/2(Fψo/2RT)

This may be rewritten as

(4.26) qs/εεoψo = (2CoαF2/RTεεo)1/2

The left-hand side of this equation is equal to the inverse of the distance x between plates
of a parallel plate capacitor (equation 4.8). A constant κ with units of m-1 is customarily
defined as

Page 4.7
(4.27) κ2 = (αF2/RTεεo) Σi Cio zi2

If all of the ions in the aqueous phase are univalent, then κ is equal to the right-hand side of
equation 4.26. Therefore,

(4.28) 1/x = κ

One may think of 1/κ as a measure of the thickness of the diffuse double layer. That is, the
same surface potential ψo would result if all of the diffusable charge were placed at a
distance 1/κ from the surface of the membrane.
In the foregoing analysis, we have incorporated the capacitative charge into the
surface charge. In actual practice, the capacitative charge is usually negligible compared to
the fixed surface charge and contributes little to the surface potential. To illustrate this, we
may note that the capacitative charge will rarely be greater than 10-7 coul/cm2 since that
charge will create a 100 mV membrane potential given the usual membrane capacitance of
1 µF/cm2. Typical surface charge densities are much larger than this (chromaffin granules
have a surface charge density of -1.38 x 10-6 coul/cm2).

Figure 4.4. The surface potential as a function of the surface charge density, according to
the Gouy-Chapman equation. The curve is for a univalent electrolyte at 10 mM
concentration. The dielectric constant has been taken as 80 and the temperature as 20°C.

Page 4.8
Figure 4.5. The decay of potential from a surface.
The surface charge density has been assumed to be 0.0158 coul/m2. The electrolyte is
univalent and the dielectric constant and temperature have been taken as 80 and 20°C

Surface charge has a number of interesting effects. Because a negative surface

charge attracts cations to the membrane surface, it increases the conductance of the
membrane to cations (McLaughlin et al., 1970) and enhances the binding of cations to the
membrane surface (McLaughlin and Harary, l976). It should be noted that the surface
potential has a greater affect on the distributions of divalent ions than on distributions of
monovalent ions. For this reason, high Ca2+ concentrations at the membrane surface,
sometimes attributed to binding, may actually be caused by the surface potential. The
surface potential also changes the pH near the surface of the membrane; this may shift the
apparent pK values of protonatable groups and the apparent pH optima of membrane-
bound enzymes. Finally, the surface potential may act to repel or attract other surfaces
thereby functioning in phenomena such as exocytosis and intercellular communication.

Section 4.4. Dipole Potential

A third kind of charge separation that can be created by biological membranes is
that of the surface dipole. Phospholipids are dipolar in character. These dipoles, all
oriented in the same direction, lead to a charge separation, which creates the dipole
potential. The dipole potential may not be a significant component of the membrane
potential because the dipoles on opposite surfaces of the membrane are oriented in opposite

Page 4.9
directions and tend to cancel each other. However, as we shall see, the dipole potential
may be a major factor in determining the ionic permeability of the lipid bilayer.
The dipolar character of phospholipids seems to arise from the ester linkage
between the fatty acid groups and the glycerol backbone. The head groups of the
phospholipids are certainly polar. However, they seem not to contribute appreciably to the
dipole potential. This is partly attributable to the fact that the head groups lie on the
aqueous surface of the membrane where the dielectric constant of the medium tends to
neutralize the dipoles. Moreover, the head groups seem to lie flat on the surface of the
membrane so the dipole moment in the direction perpendicular to the membrane surface is
The dipole potential is

(4.29) ∆ψ = D/εεo

where D is the surface dipole density in coul/m. To estimate the potential magnitude of
surface dipole effects, let us assume that a typical phospholipid has a dipole moment of 1.5
Debyes (1.5 x 10-18 or 5 x 10-30 coul.m) and occupies a membrane area of 60 Å2.
This yields a surface dipole density of 8.34 x 10-12 coul/m and implies a dipole potential of
(1000/ε) mV. If we assume that the dielectric constant of membrane in the region of the
dipole layer membrane is between 4 and 10, the surface dipole should create a potential
difference of 100-250 mV.
The electric field set up by the capacitative charge may cause the dipoles to orient in a
direction perpendicular to the membrane. Thus, the surface dipole may change in response
to a membrane potential. In this way, the surface dipole will affect the electric field near
the membrane surface and may be important in modulating the response of membrane-
bound enzymes to the membrane potential.


R.G. Ashcroft, H.G.L. Coster, and J.R. Smith (1981) The molecular organisation of
bimolecular lipid membranes. The dielectric structure of the
hydrophilic/hydrophobic interface, Biochim. Biophys. Acta 643, 191-204.
R. Aveyard and D.A. Haydon (1973) An Introduction to the Principles of Surface
Chemistry, Cambridge University Press, Cambridge.
R.F. Flewelling and W.L. Hubbell (1986) Hydrophobic ion interactions within membranes,
Biophys. J. 49, 531-540.
R.F. Flewelling and W.L. Hubbell (1986) The membrane dipole potential in a total
membrane potential model, Biophys. J. 49, 541-552.
D.A. Haydon and S.B. Hladky (1972) Ion transport across thin lipid membranes: A critical
discussion of mechanisms in selected systems, Quart. Rev. Biophys. 5, 187-282.
S. McLaughlin (1977) Electrostatic potentials at membrane-solution interfaces, Current
Topics in Membrane Transport 9, 71-144.
S. McLaughlin and H. Harary (1976) The hydrophobic adsorption of charged molecules to
bilayer membranes: A test of the applicability of the Stern equation, Biochemistry
15, 1941-1948.

Page 4.10
P.L. Yeagle (1979) Effect of transmembrane electrical potential and micelle geometry on
phospholipid head group conformation, Arch. Biochem. Biophys. 198, 501-505.

Page 4.11
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 5.1. The Langmuir Adsorption Isotherm

Small molecules may bind to membranes either non-specifically or specifically.
Non-specific binding occurs when small amphiphilic molecules adsorb to the surface of the
lipid bilayer or to the equivalent hydrophobic/hydrophilic interfaces of membrane proteins.
Specific binding occurs when ligands bind to selective binding sites of specific receptors.
In either case, similar analyses can be used to describe the binding process.
The simplest description of the relationship between bound and free ligand assumes
an equilibrium defined by the law of mass action:

(5.1) L + M = LM

where L is the free ligand, M signifies empty binding sites on the membrane, and LM
represents ligand bound to the membrane. The equilibrium constant K is

(5.2) K = [L][M]/[LM]

Let us define Lm as the concentration of bound ligand ([LM]) and Lmax as the total number
of binding sites on the membrane ([M]+[LM]). Then the equilibrium constant becomes

(5.3) K = [L](Lmax-Lm)/Lm

Equation 5.3 may be solved to yield an expression for the amount of bound ligand

(5.4) Lm = Lmax [L] / ([L] + K)

The amount of bound ligand Lm depends upon the binding constant K, the number of
binding sites Lmax, and the concentration of ligand [L] at the surface of the membrane
(Figure 5.1). As described in Section 5.2, if the ligand is charged, its concentration at the
surface of the membrane [L] will not necessarily equal its concentration in the bulk of the
aqueous phase but will be related to it by the surface potential. As described in Section
5.3, the binding constant K reflects the change in free energy occurring upon binding.
Equation 5.3 may also be rearranged to yield the Scatchard equation:

(5.5) Lm/[L] = Lmax/K - Lm/K

This form is particularly useful because a plot of Lm/[L] vs. Lm is linear and may be used to
define the parameters Lmax and K (Figure 5.2).

Page 5.1
Lmax K

Lmax [L]

K Lmax
[L] Lm

Figure 5.1 Figure 5.2

Section 5.2. Effect of Surface Potential

[L] is the concentration of ligand at the surface of the membrane. As described
earlier, this is related to [L]∞, the concentration far from the membrane, by the Boltzmann

(5.6) [L] = [L]∞ exp (-ziFψo/RT)

Obviously, the known ligand concentration in the bulk of the aqueous phase [L]∞ may be
used in place of the ligand concentration at the membrane surface [L] if the ligand is
uncharged or if the surface potential ψo is negligible. Since the surface potential depends
upon both the ionic strength of the solution and upon the surface charge density, it can be
considered negligible if the ionic strength is high or if the surface charge density is low. If
the surface potential is greater than a few millivolts, however, [L] will be significantly
different from [L]∞. For example, if the surface potential is -18 mV and the ligand is a
univalent anion, [L] will be only 50% of [L]∞, and using [L]∞ in place of [L] will
significantly change the appearance of the binding curve (Figure 5.3) and the Scatchard
plot (Figure 5.4).
A special problem occurs when a charged ligand binds to the membrane to such an
extent that the bound ligand itself affects the membrane surface charge. This is generally
not a problem for ligands that bind to receptors since the receptor density (Lmax) is typically
insignificant compared to the surface charge density of the membrane. Ligands that adsorb
to the membrane, however, may have a significant effect on the surface charge density.
This situation, first considered by Stern, has been analyzed more recently by McLaughlin
and Harary (1976). If we assume that the surface charge density of the membrane is
initially zero, then [L] = [L]∞ for low values of Lm. As Lm increases, however, the surface
potential will increase and binding will deviate in the direction expected in the presence of
a surface potential. Thus, the Scatchard plot will appear to curve (Figure 5.4).

Page 5.2
Section 5.3. The Binding Constant
The binding constant K reflects the standard free energy change upon binding. This
can be seen by recognizing that the free energy change for the reaction in equation 1 is

(5.7) ∆G = ∆G° + RT ln ([LM]/[L][M])

At equilibrium, ∆G = 0 and [LM]/([L][M]) = 1/K. Therefore,

(5.8) ∆G° = RT ln K

For adsorption, the free energy change represents the free energy change for the transfer of
the ligand from water into the membrane. For receptors, the free energy change carries
additional significance. In this case, ligand binding not only changes the location of the
ligand, but it changes the conformation of the receptor to which the ligand binds. Thus, the
free energy of binding includes not just an affinity between the binding site and the protein
but a change in the conformation of the protein. The conformational change is crucial, of
course, because it activates the receptor thereby transmitting the ligand-binding signal.

Section 5.4. Specific vs. Non-Specific Binding

Often, it is desirable to study the binding of a ligand to a receptor or other
membrane-bound protein, but the ligand (because it is hydrophilic or amphiphilic) exhibits
considerable non-specific binding to the membrane. This may be analyzed as follows:
Total binding to the membrane (from eq. 5.4) will be the sum of the specific (sp) and non-
specific (ns) contributions:

(5.9) Lm = Lmsp + Lmns = Lmaxsp [L]/([L] + Ksp) + Lmaxns [L]/([L] + Kns)

Let us assume that the affinity of the non-specific binding sites is very weak compared to
the affinity of the specific sites (Kns >> Ksp). To measure specific binding, one uses ligand

Page 5.3
concentrations that range about Ksp, so [L] will be very small relative to the non-specific
binding constant [L] << Kns). Equation 5.9 then reduces to

(5.10) Lm = Lmaxsp [L]/([L] + Ksp) + Lmaxns [L]/Kns

so non-specific binding will be directly proportional to the ligand concentration over this
range. This non-specific binding can be measured by adding a high concentration of ligand
(Kns >> [L] >> Ksp). Under these conditions, equation 5.9 reduces to

(5.11) Lm = Lmaxsp + Lmaxns [L] / Kns

Because Lmaxsp << Lmaxns, the measured binding Lm will all be non-specific. This gives the
slope of the non-specific binding line (Lmaxns / Kns). Then specific binding can be
calculated by subtracting non-specific binding from total binding as in Figure 5.5.


S. McLaughlin and H. Harary (1976) The hydrophobic adsorption of charged molecules to

bilayer membranes: A test of the applicability of the Stern equation, Biochemistry
15, 1941-1948.
T.M. DeLorey and R.W. Olsen (1992) γ-Aminobutyric acidA receptor structure and
function, J. Biol. Chem. 267, 16747-16750.
C.D. Strader, T.M. Fong, M.R. Tota, D. Underwood, and R.A. F. Dixon (1994) Structure
and function of G-protein-coupled receptors, Ann. Rev. Biochem. 63, 101-132.
M. Hollmann and S. Heinemann (1994) Cloned glutamate receptors, Ann. Rev. Neurosci.

Page 5.4
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 6.1. Formal Analysis of Permeability and Conductance

The passage of ions and molecules across membranes is a phenomenon often
divided into two parts: permeability and conductance. Conductance describes the
movement of electrolytes (ions and charged molecules) across the membrane in response to
a membrane potential. Permeability describes the movement of uncharged molecules
(nonelectrolytes) as well as the movement of electrolytes in the absence of a membrane
Three different approaches may be used to analyze permeability and conductance:
classical thermodynamics, nonequilibrium thermodynamics and statistical mechanics.
Strictly speaking, classical thermodynamics and statistical mechanics apply only to
equilibrium situations. Nonequilibrium thermodynamics, by contrast, was developed to
describe non-equilibrium phenomena such as permeability and conductance. Many purists,
therefore, prefer to discuss permeability and conductance in terms of nonequilibrium
thermodynamics. Unfortunately, this approach generally provides little insight into the
physical events occurring as molecules or ions cross biological membranes. For that
reason, we will focus here on the other two approaches: We will see that they both yield
the same results and provide different insights into the processes involved.
The analysis of permeability and conductance using classical thermodynamics is
based on the concept of the electrochemical potential:

(6.1) µ = µ° + RT ln C + zFψ

At equilibrium, a given ion or molecule must have the same electrochemical potential on
both sides of the membrane. If we use the subscript i to denote one side of the membrane
(inside) and the subscript o to denote the other side (outside), then

(6.2) µi° + RT ln Ci + zF ψi = µo° + RT ln Co + zF ψo

Since µi° = µo°, equation 6.2 can be rearranged to yield

(6.3) Ci/Co = exp [-zF(ψi - ψo)/RT]

This is the famous Nernst equation, which gives the equilibrium concentration gradient of
an electrolyte in the presence of a membrane potential. If the molecule is a non-electrolyte
(z=0) or if the membrane potential is zero, then equation 6.3 simplifies to

(6.4) Ci = C o

The analysis of permeability and conductance using statistical mechanics is based

on two concepts: the concept of unidirectional fluxes and transition state theory. The
notion that net flux across the membrane is simply the sum of two individual unidirectional
fluxes was first introduced by Ussing. Transition state theory, championed by Eyring,
argues that the rate at which a given transition will occur depends on the probability that
the ion or molecule has enough energy to cross the transition barrier. The probability of

Page 6.1
having this activation energy is given by the Boltzmann distribution. Let us imagine that
the activation energy for a particular ion or molecule crossing the membrane is Ea (Figure
6.1). Then the rate of influx (Joi) is

(6.5) Joi = Co exp [-(Ea-zF ψo)/RT] x constant

The rate of efflux (Jio) is

(6.6) Jio = Ci exp [-(Ea-zF ψi)/RT] x constant

These two rates must be equal at equilibrium, so

(6.7) Ci exp [-Ea/RT + zF ψi/RT] = Co exp [-Ea/RT + zF ψo/RT]

Equation 6.7 reduces to the Nernst equation (equation 6.3), so the statistical mechanics
approach predicts the same equilibrium as does the thermodynamics approach.

Energy Outside Inside



ψ zF

ψ zF

Figure 6.1. Transition-state energy for membrane permeation.

Section 6.2. Equilibria

Permeation and conduction are passive processes and will proceed in the direction
of equilibrium. For electrolytes in the presence of a membrane potential, we have shown
that the equilibrium concentration gradient is given by the Nernst equation (equation 6.3).
For non-electrolytes and for electrolytes when the membrane potential is zero, the
equilibrium reduces to Co = Ci (equation 6.4). An interesting exception to these rules is
found in the case of weak acids and weak bases. At neutral pH, weak acids and weak bases
are predominantly in their charged forms (A- and BH+). These charged species do not
permeate across the hydrophobic barrier presented by biological membranes. The charged
species, however, are in equilibrium with uncharged species that will permeate the
membrane. Permeation of the uncharged species causes the charged species to reach the
following equilibria:

(6.8) [BH+]i/[BH+]o = [H+]i/[H+]o

Page 6.2
(6.9) [A-]i/[A-]o = [H+]o/[H+]i

To understand the logic behind this, let us consider the case of a weak base (Figure 6.2).
The uncharged species (B) will reach the equilibrium expected of nonelectrolytes (Bo = Bi).
On either side of the membrane, this unprotonated species will be in equilibrium with the
protonated form:

(6.10) K = [B]o[H+] o /[BH+] o = [B]i[H+] i /[BH+] i

Since [B]o = [B]i, equation 6.10 reduces to equation 6.8. A similar analysis may be applied
to weak acids to establish equation 6.9.

Outside Membrane Inside

BH + BH +

H++ B B + H +

Figure 6.2. Permeation of a weak base.

Section 6.3. Permeation of Non-Electrolytes

The velocity (v) at which a molecule or ion will diffuse is proportional to the
gradient of its electrochemical potential (dµ/dx):

(6.11) v = -(1/Nf) (dµ/dx)

N is Avogadro's number and f is the frictional coefficient. The total rate of flow J is the
velocity multiplied by the concentration C:

(6.12) J = vC = -(C/Nf)(dµ/dx)

For a non-electrolyte, dµ/dx = d(RT ln C)/dx, so

(6.13) J = (-C/Nf)(RT/C)(dC/dx) = -(RT/Nf)(dC/dx)

If we define RT/Nf as the diffusion coefficient D, equation 6.13 reduces to Fick's Law of

Page 6.3
Membrane permeation involves diffusion across the membrane. To analyze this,
we must integrate Fick's Law (equation 6.13) from one side of the membrane to the other
(Figure 3).
d d
(6.14) ∫0 J dx = - ∫0 (RT/Nf)(dC/dx)dx

Outside Membrane Inside




C mo

0 d

Figure 6.3. Concentration profile for steady-state flow across a membrane

This integration, which allows us to determine how the flow depends on the overall
concentration gradient (Ci - Co) across the membrane, is relatively simple. As we shall see
(Section 7.1), the corresponding integration for the case of non-electrolytes is much more
To integrate equation 6.14, we assume that the flow of non-electrolyte across the
membrane reaches a steady state. This means that the flow J must be the same at all points
within the membrane so that the concentration profile across the membrane does not
change with time. Steady-state is a natural assumption because it is a stable condition.
Suppose that J varied within the membrane so that the flow into a particular region was
greater than the flow out of that region. The concentration of non-electrolyte in that region
would then rise. This, in turn, would cause the flow out of the region to increase and the
flow into the region to decrease. Thus, the concentration would spontaneously change to
maintain equality between the flows into and out of the region. This assures steady state.
We may then integrate equation 6.14 assuming that J is independent of x (constant).

(6.15) Jd = -(RT/Nf)(Cmi-Cmo)

Cmi is the concentration in the membrane at the inner surface and Cmo is the concentration
in the membrane at the outer surface.
We imagine that, at each surface of the membrane, molecules in the aqueous phase
are in equilibrium with molecules in the membrane phase. Therefore, the chemical
potential in the water phase (µw) must equal the chemical potential in the membrane (µm):

(6.16) µw = µw° + RT ln Cw = µm = µm° + RT ln Cm

Page 6.4
The concentration at the surface of the membrane (Cm) is then

(6.17) Cm = Cw exp [(µw°- µm°)/RT]

An expression for J is obtained by substituting equation 6.17 into equation 6.15 and

(6.18) J = -(RT/Nfd){exp [(µw° - µm°)/RT]}(Ci - Co)

The flux is proportional to the concentration difference across the membrane (Ci - Co). It is
also proportional to the permeability coefficient P defined as

(6.19) P = (RT/Nfd){exp [(µw° - µm°)/RT]}

The permeability coefficient includes the term RT/Nf, which is the diffusion coefficient of
the molecule in the lipid phase of the membrane. This term will depend on the size of the
molecule. The term exp[(µw°-µm°)/RT] describes the partitioning of the molecule between
water and the membrane and will depend on the lipid solubility of the molecule. Since
Cm/Cw is the membrane:water partition coefficient (Kp), equation 6.17 implies that

(6.20) Kp = exp [(µw° - µm°)/RT]

A comparison of equations 6.19 and 6.20 shows that P and Kp should be linearly related.
Walter and Gutknecht (1984) tested this prediction in a study of the permeability of lipid
bilayers to a series of carboxylic acids. After correcting for unstirred layer effects and
assuming that only the protonated form of the carboxylic acid permeates, they found that
the permeability coefficient is related to the hexadecane:water partition coefficient (Kp') as

(6.21) log P = 0.90 log Kp' + 0.87

The observed slope of 0.90 is close to the predicted slope of 1.0. Moreover, the free energy
change for the transfer of the carboxylic acid from water into the membrane (µm°-µw°) can
be determined from either the permeability coefficient P or the partition coefficient Kp.
The incremental change in the free energy per methylene group for the series of carboxylic
acids (acetic, propionic, butyric, hexanoic) is -898 ± 159 cal/mole determined from the
partition coefficient and -764 ± 54 cal/mole determined from the permeability coefficient.
These numbers agree very well with the energies described in the discussion on micelle
formation (Section 2.4).

Section 6.4. Unstirred Layers

When a compound diffuses from one side of a membrane to the other, the
membrane may be the principal barrier to flow but not the only barrier. Passage of the
molecule across the membrane may also be slowed by diffusion across the aqueous layers
adjacent to either surface of the membrane. These so-called unstirred layers may range in
thickness from 1 µm to 500 µm (Remember that the membrane itself is only 4 x 10-3 µm

Page 6.5
thick). The unstirred layer effect will generally be most prominent for relatively nonpolar
compounds. For these compounds, the permeability coefficient will be large and diffusion
across the membrane itself will be relatively fast. Diffusion across the aqueous layers,
therefore, may be partially rate-limiting. For compounds that are quite water soluble,
permeation across the membrane will be slow, and diffusion across the unstirred layers will
have relatively less effect.
To consider the effect of unstirred layers, imagine that a compound has a membrane
permeability coefficient P and an aqueous diffusion constant D. Let the unstirred layers
have thicknesses, di and do (Figure 6.4), let the bulk concentrations of the compound be Ci
(inside) and Co (outside) and let the concentrations at the surface of the membrane be Cmi
and Cmo. The flow through the membrane then is

(6.22) Jm = P (Cmi - Cmo)

The flow through the unstirred layers will be

(6.23) Ji = (D/di) (Ci - Cmi)

(6.24) Jo = (D/do) (Cmo - Co)

At steady-state, the flows must all be the same (Jm = Ji = Jo = J). Therefore, equations 6.22,
6.23 and 6.24 can be transformed to

(6.25) J/P = Cmi - Cmo

(6.26) Jdi/D = Ci - Cmi

(6.27) Jdo/D = Cmo - Co

By summing these three equations, we obtain

(6.28) J (1/P + di/D + do/D) = Ci - Co

Unstirred Unstirred
Layer Membrane Layer

do d
C mi

C mo

Page 6.6
Figure 6.4. Permeation and diffusion through unstirred layers

Therefore, the effect of unstirred layers is to decrease the permeability so the apparent
permeability coefficient (Papp) is smaller than P:

(6.29) 1/Papp = 1/P + di/D + do/D


A. Finkelstein (1976) Water and nonelectrolyte permeability of lipid bilayer membranes, J.

Gen. Physiol. 68, 127-135.
A. Finkelstein and A. Cass (1968) Permeability and electrical properties of thin lipid
membranes, J. Gen. Physiol.52, 145s-172s.
A.R. Koch (1970) Transport equations and criteria for active transport, Am. Zool. 10, 331-
S. G. Schultz (1980) Basic Principles of Membrane Transport, Cambridge University
Press, Cambridge.
A. Walter and J. Gutknecht (1984) Monocarboxylic acid permeation through lipid bilayer
membranes, J. Membrane Biol. 77, 255-264.

Page 6.7
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 7.1. Permeation of Electrolytes

The permeation of electrolytes may be analyzed using the same approach as is used
for nonelectrolytes (Section 6.3). The flux J is assumed to be proportional to the
thermodynamic driving force, the derivative of the electrochemical potential:

(7.1) J = - ( ) C
dx [ µ o + RT ln C + zF ψ ] = - RT
Nf [ dC
dx + ( zFC
RT ) ]

As before, we must integrate this from one side of the membrane to the other. In
this case, however, we have two parameters, C and ψ, that will vary as we cross the
membrane. We may use the steady-state assumption to define one of these parameters in
terms of the other, but we will need another equation to define both. We will return to this
problem later.
First, note that

zFC d ψ ψ/RT)
(zFψ /RT)
] [ dx ] e(zF
(7.2) = +
dx RT dx

A comparison of equations 7.1 and 7.2 reveals that exp (zFψ/RT) may be used as an
integrating factor so that

(7.3) J exp (zFψ/RT) = - (RT/Nf) d/dx[C exp (zFψ/RT)]

If we make the steady-state assumption (J is constant across the membrane), then we may
partially integrate equation 7.3:
zFψ/RT zFψmi/RT zFψmo/RT
(7.4) J ∫0 e dx = - (RT/Nf) [Cmi e - Cmo e ]

Cmi and Cmo are the concentrations in the membrane at d and 0 respectively (Figure 7.1).
They are related to the concentrations Ci and Co in the bulk aqueous phases by the
membrane:water partition coefficient (Kp) and by the surface potential:

(7.5) Cmi = Ci Kp exp[zF(ψi- ψmi)/RT]

(7.6) Cmo = Co Kp exp[zF(ψo- ψmo)/RT]

ψ ψ
(7.7) J ∫
d zF ψ /RT RTK p ( C e zF i /RT - C e zF o /RT )
dx = -
0 e Nf i o

If we define the electrical potential as zero on the outside (ψo = 0), then

Page 7.1
zF ψ m /RT
( )
(7.8) J = RTKp
Nfd d
∫ exp (zFψ/RT) dx
[ Co - C i e ]
where ψm is the membrane potential. We may note that RTKp/Nfd is the permeability
coefficient P (equations 6.19 and 6.20). Therefore,

(7.9) J = PQ[Co - Ci exp (zFψm/RT)]

where Q is defined as
(7.10) Q = d/∫0 exp (zFψ/RT) dx

Outside Membrane Inside

C mi

C mo ψi
Co ψ mi

0 d

Figure 7.1. Concentration profile for steady-state electrolyte flow

The problem now is to evaluate Q. To integrate exp (zFψ/RT), we need to define

ψ(x). We will use the constant field approximation first proposed by Goldman. A second
possibility is the assumption of electrical neutrality in the membrane, an approach first
explored by Planck. Physically, the two approaches are similar. A constant field within
the membrane implies that there must be electrical neutrality. In terms of formalism,
however, the Goldman approach is simpler. The equations obtained by assuming a
constant field (or a linear gradient in electrical potential) are simpler than those obtained by
summing anion and cation concentrations within the membrane and setting the total charge
equal to zero.
The constant field assumption states that the electric field (E = -dψ/dx) is constant
throughout the membrane. Therefore,

(7.11) ψ(x) = - Ex + ψ(0)

If we assume that surface potentials are negligible, then ψ(0) = 0 and

ψm = ψ(d) = - Ed. Equation 7.10 may then be integrated to give

Page 7.2
zF ψ m/RT
(7.12) Q =
exp (zF ψ /RT) - 1
Introducing equation 7.12 into equation 7.9 yields a final expression for the flow J:

zF ψm /RT zF ψ m /RT
(7.13) J = P ( exp (zF ψm /RT) - 1 ) (Co - C i e )
Equations 7.9 and 7.13 provide us with alternative expressions for determining the
flow of electrolytes across biological membranes with the latter equation including the
assumption of a constant field. These expressions allow us to calculate the flow of
electrolyte driven across the membrane by a membrane potential, by a concentration
gradient, or by a combination of these two forces. In Figure 7.2, the membrane potential
and the concentration gradient are used as the axes of a two-dimensional graph. Equations
7.9 and 7.13, therefore, allow us to calculate the flow at any point on this graph. Let us
examine some special cases. First, note that both equations reduce to the Nernst equation
if the flow J is zero:

(7.14) Ci/Co = exp (-zFψm/RT)

Second, if ψm is zero, Q = 1 and equation 7.9 reduces to Fick's Law:

(7.15) J = P (Co - Ci)

Finally, if there is no concentration gradient (Ci/Co = 1), equation 7.13 reduces to:

(7.16) J = - PCzFψm/RT

Since I = -JFz, equation 7.16 is equivalent to Ohm's Law (I = g ψm) where the conductance
g is PCz2F2/RT.

TABLE I. Permeability Coefficients of Ions

H2O 2 x 10-5 - 2 x 10-2 cm/sec

Na+ 8 x 10-9 cm/sec (squid axon)
K+ 6 x 10-7 cm/sec (squid axon)
H+ 10-5 cm/sec (chromaffin vesicle)
Na+ and K+ 10-10 - 10-11 cm/sec (lipid bilayer)
Cl- 10-10 cm/sec (phospholipid vesicle)

Page 7.3
zψ Ohm's Law zψ Ohm's Law

Fick's Law Fick's Law

C i /C o ln C i /C o


0 1 0

Figure 7.2. Regions of validity of the permeation and conductance equations

The above analysis of electrolyte permeation was developed using the

electrochemical potential. We may also analyze the permeation of electrolytes using the
transition state approach. Let us imagine that the energy of the electrolyte is Eo on the
outside of the membrane, Ei on the inside, and Em in the membrane (Figure 7.3). The
unidirectional flux in the inward direction will be

(7.17) roi = Co exp [-(Em- Eo)/RT] x constant

The unidirectional flux in the outward direction will be

(7.18) rio = Ci exp [-(Em- Ei)/RT] x constant

The net flow J will be the difference between these two fluxes:

(7.19) J = roi - rio

= {Co exp [-(Em- Eo)/RT] - Ci exp [-(Em- Ei)/RT]} x constant

Page 7.4
Energy Outside Membrane Inside




Figure 7.3. Energy barrier for permeation of electrolytes

Following our usual convention, we will define the electrical potential as zero on the
outside. Then Eo will simply be the standard free energy of the electrolyte in water (Eo =
µw°) and Ei will differ from this by the membrane potential (Ei = µw° + zFψm). Upon
introducing these values for the energies, equation 7.19 reduces to

(7.20) J = exp [-(Em- µw°)/RT] [Co - Ci exp (zFψm/RT)] x constant

If we let PQ = exp [-(Em- µw°)/RT] x constant, then equation 7.20 is the same as the flux
equation derived using the electrochemical potential approach (equation 7.9).

Section 7.2. The Born Charging Equation

Several factors determine the energy Em of an ion in a membrane. These include 1)
hydrophobic interactions, 2) electrostatic potentials (both surface and dipole), 3) short
range forces (steric effects), and 4) the Born charging energy. Hydrophobic interactions
and short range forces apply to non-electrolytes as well. Surface and dipole potential
effects were considered earlier. In this section, we will consider the effect of the Born
charging energy.
The Born charging energy is the energy required to assemble a given amount of
charge on a particle of a given size. Because this energy is lower in a medium with a high
dielectric constant, the Born charging energy is much smaller for an ion in water than for
an ion in a hydrocarbon medium. This means that an ion requires much more energy to
enter a hydrocarbon phase than to enter an aqueous phase. The Born charging energy,
therefore, accounts for the insolubility of ions in hydrocarbon phases and for the
impermeability of biological membranes to ions. Because the Born charging energy is
greater for a localized charge than for an equivalent delocalized charge, ions with
delocalized charge will permeate through biological membranes more easily.

Page 7.5
x a

Figure 7.4. Charging a conducting sphere

Imagine that an ion is a conducting sphere of radius a (Figure 7.4). If q is the

charge placed on the sphere and ε is the dielectric constant of the medium, the Born
charging energy is

(7.21) W = q2/8πεεoa

This equation can be derived as follows. Outside of a conducting sphere, the electric field
created by the sphere is the same as the electric field created by a point charge (of equal
charge) located at the center of the sphere. Therefore, the force between a sphere of charge
q' and a charge dq' is given by Coulomb's Law:

(7.22) F = q'dq'/4πεεox2

The work required to move the charge dq onto the sphere from an infinite distance away is
a a
(7.23) dW = - ∫∞ F dx = - ∫∞ (q' dq'/4πεεox2) dx
= q' dq'/4πεεoa

The work required to place the entire charge q on the sphere is then
(7.24) W = ∫ δW = ∫0 q' dq'/4πεεoa = q2/8πεεoa

The change in the charging energy upon moving the ion from water into a
hydrocarbon phase is

(7.25) W = (q2/8πεoa) (1/εhc - 1/εw)

The membrane is not an infinite hydrocarbon phase, but is a thin layer of hydrocarbon with
water on both sides. Therefore, the change in charging energy upon moving the ion from
water into a membrane is somewhat smaller than the change shown in equation 7.25. The
work done in moving a charge q a distance x into a membrane of thickness d has been
approximated by Flewelling and Hubbell (1986):

Page 7.6
( )( )[ ( )( ) ]
2 1 a - 1.2 x 2
q 1 - 2x a
(7.26) W = 8aπε o ε hc d d

Section 7.3. The Goldman-Hodgkin-Katz Equation

If ions are in equilibrium across a membrane, then the membrane potential will be
given by the Nernst equation. This is rarely the case, however. Generally, ions are in
constant flux (transport and permeation) and the capacitative charge is determined by the
steady-state distribution of ions. The membrane potential can nevertheless be determined
from this steady-state distribution using the Goldman-Hodgkin-Katz equation. At steady-
state, the net charge flux will be zero.

(7.27) 0 = Σcations zjFJj + Σanions zjFJj

The fluxes Jj are defined by equation 7.13. If we impose the simplifying assumption that
all of the ions are univalent, then

(7.28) 0 = ΣcationsFPj{(Fψm/RT)/[exp(Fψm/RT)-1]}{Coj - Cijexp(Fψm/RT)}

- ΣanionsFPj{(-Fψm/RT)/[exp(-Fψm/RT)-1]}{Coj - Cijexp(-Fψm/RT)}

Dividing by Fψm/RT and rearranging the exponentials in the anion term yields

(7.29) 0 = Σcations FPj{1/[exp (Fψm/RT) - 1] }{ Coj - Cij exp (Fψm/RT) }

- Σanions FPj{ (-1/[1 - exp (Fψm/RT)] }{ Coj exp (Fψm/RT) - Cij }

Upon dividing by F{1/[exp (Fψm/RT) - 1] }, we obtain

(7.30) 0 = Σcations Pj{Coj - Cij exp (Fψm/RT)}

- Σanions Pj{Coj exp (Fψm/RT) - Cij}

Then solving for the exponential term gives

Σ cations P jC oj + Σ anions P jC ij
(7.31) exp (Fψ /RT) =
Σ cations P jC ij + Σ anions P C oj

Therefore, the membrane potential is defined by the ion concentrations and permeabilities
as follows:

Σ cations P jC oj + Σ anions P jC ij
(7.32) ψ = (RT/F) ln
m Σ cations P jC ij + Σ anions P C oj

This is the Goldman-Hodgkin-Katz equation.

Page 7.7

R.F. Flewelling and W.L. Hubbell (1986) Hydrophobic ion interactions within membranes,
Biophys. J. 49, 531-540.
R.F. Flewelling and W.L. Hubbell (1986) The membrane dipole potential in a total
membrane potential model, Biophys. J. 49, 541-552.
A.L. Hodgkin and A.F. Huxley (1952) A quantitative description of membrane current and
its application to conduction and excitation in nerve, J. Physiol. 117, 500-544.
A.R. Koch (1970) Transport equations and criteria for active transport, Am. Zoologist 10,
R.I. Macey (1978) Mathematical models of membrane transport processes, in Membrane
Physiology (T.E. Andreoli, J.F. Hoffman, and D.D. Fanestil, eds.), Plenum, New
York, pp. 125-146.
A. Parsegian (1969) Energy of an ion crossing a low dielectric membrane: Solutions to
four relevant electrostatic problems, Nature 221, 844-846.

Page 7.8
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 8.1. Channel-forming Antibiotics

Ion channels are needed to conduct ions across biological membranes because ions,
particularly cations, do not permeate readily across the lipid bilayer. The structural
simplicity required to create an ion-conducting channel across a biological membrane is
exemplified by channel-forming antibiotics, such as gramicidin and amphotericin B. These
antibiotics form ungated channels, so they dissipate the ion gradients needed for proper
membrane function and cause cells to spend energy in futile ion pumping. Since these
ungated channels are lethal, regulation of channel opening or “gating” is obviously an
important feature of natural ion channels.
Gramicidin is a pentadecapeptide consisting of alternating L and D amino acids:


The conductance through gramicidin channels varies as the square of the gramicidin
concentration indicating that the compound functions as a dimer. Indeed, two peptides can
be linked at the formyl groups on their amino terminal ends, and the coupled structure will
function as a channel. The channel formed by gramicidin is not very selective and has a
unitary (single channel) conductance of 5 pS. It is thought that gramicidin forms a helix
with the hydrophobic side chains on the outside and the carbonyl oxygens oriented to the
inside. This forms a channel 2 Å in diameter.
Amphotericin B forms a larger channel and increases the permeability of lipid
bilayers to water and small electrolytes as well as ions. The conductance depends on the
4th-12th power of the concentration suggesting that a number of molecules are required to
form the channel. Because amphotericin B is an elongated molecule with a hydrophobic
side and a hydrophilic side, it is thought to line the sides of the channel like the staves on a
barrel. Amphotericin B requires a sterol for activity, and thus makes channels in
membranes that contain cholesterol.

Section 8.2. Voltage-Gated Channels

In recent years, molecular biological techniques have yielded a wealth of
information about the membrane-spanning proteins that form ion channels. A
generalization that may be emerging is that channels with similar gating mechanisms have
similar structures. The voltage-gated channels, in particular, have common structural
features despite the fact that they have different ion selectivities and conductances.
Voltage-gated channels comprise the S4 superfamily, so named because the proteins
function as tetramers having either four subunits or four homologous segments. The K+
channel, for example, is a tetramer with six membrane-spanning regions in each subunit.
The Na+ channel is a single large peptide (~260 kDa), but that peptide has four
homologous segments with 6 membrane-spanning regions in each. Looking down at the
membrane, the four segments are arranged at the corners of a square with the ion channel
itself passing down through the center between them.
As mentioned, the core of the Na+ channel is formed by a single large α subunit. In
the eel electroplax, that is the only subunit. The sodium channel from mammalian skeletal

Page 8.1
muscle also contains a β1 subunit (38 kDa), while the sodium channel in mammalian brain
contains both a β1 (36 kDa) and a β2 subunit (33 kDa) along with the α subunit.
Voltage-gated potassium channels include the delayed rectifier (DR) channel,
which functions in actions potentials in excitable membranes, and the CaK channel, which
is activated by Ca2+ as well as by depolarization. Both are blocked by barium. The CaK
channel has a very high unitary conductance and is specifically blocked by charybdotoxin.
Calcium channels have been classified into a variety of types based on functional
characteristics and pharmacology. The well characterized L-type channel is a high-
threshold, slow inactivating channel, which is blocked by dihydropyridines. Low-
threshold, fast inactivating Ca2+ channels are classed as T-type. Two other high-threshold
channels (N and P) are distinguished by their sensitivity to peptide toxins. N channels are
blocked by ω-conotoxin GVIA while P channels are blocked by ω-agatoxin IVA.
Structurally, these channels are thought to be similar. The L-type Ca2+ channel from
skeletal muscle has five subunits: α1 (170 kDa), α2δ (175 kDa), β (52 kDa) and γ (32
kDa). The α1 subunit contains binding sites for Ca2+ channel antagonists and is thought to
be the subunit forming the functional Ca2+ channel. The α2δ piece consists of a large α2
subunit linked to the δ subunit by disulfide bonds.

Section 8.3. Ligand-Gated Channels

Extracellularly activated ligand-gated ion channels seem to fall into three major
groups based on the number of subunits. The nicotinicoid group, represented by the
nicotinic acetylcholine receptor, has five homologous subunits arranged in a pentagonal
structure around a central channel. This group includes the cation-conducting nicotinic and
serotonin (5HT) receptors and the anion-conducting GABAA, GABAC and glycine
receptors. The best characterized member of the nicotinicoid receptor group is the
nicotinic acetylcholine receptor. Structurally, the acetylcholine receptor consists of five
homologous subunits: 2α, 1β, 1γ and 1 δ. Each subunit has at least four transmembrane
segments. The acetylcholine binding sites are located on the α subunits.
The second group of extracellularly activated ligand-gated ion channels, the
glutamate-activated cation channels, has four homologous subunits. This group includes
the AMPA, Kainate and NMDA receptors named for ligands (agonists) that specifically
activate each type. The third group of ligand-gated ion channels, the ATP-gated channels,
have three homologous subunits and include the ATP2x and ATP2z receptors.
As ion channels, the ligand-gated channels seem to exhibit less specificity than the
voltage-gated channels. The cation channels do not discriminate between Na+ and K+, so
they drive the membrane potential toward zero (midway between the Na+ and K+
equilibrium potentials). Because this depolarizes the membrane, these receptors are often
called excitatory. The anion channels drive the membrane potential toward the Cl-
equilibrium potential (negative inside). Thus, they tend to restore the resting membrane
potential and are often called inhibitory.

Section 8.4 Ryanodine and Inositol Tris Phosphate Receptors

The ryanodine and inositol trisphosphate receptors are related proteins that form
Ca2+ channels in intracellular membranes. The ryanodine receptor is a very large protein
(565 kDa) and is responsible for releasing Ca2+ from the sarcoplasmic reticulum (SR) in
muscle. Most of this protein (the amino-terminal 80%) is cytoplasmic and constitutes a

Page 8.2
"foot" structure. The remaining 20% on the carboxyl end includes 4 to 10 transmembrane
segments and presumably creates the ion channel structure. Ryanodine, a plant product,
opens this channel. In vivo, however, the channel is gated by cytoplasmic Ca2+. Thus,
when L-channels in the T-tubule membranes open and allow Ca2+ to enter the muscle cell,
the ryanodine receptor channel in the sarcoplasmic reticulum membrane responds by
releasing more Ca2+ from the SR. Thus, the ryanodine receptor mediates Ca2+-induced
Ca2+ release.
A related protein, the inositol-1,4,5-trisphosphate receptor releases Ca2+ from the
endoplasmic reticulum in response to the intracellular messenger inositol-1,4,5-
trisphosphate (IP3). It has a molecular mass of 260 kDa and, like the ryanodine receptor,
consists of a large amino-terminal foot and a smaller carboxyl portion containing 8-10
transmembrane segments. Both proteins appear to associate into homotetramers.

TABLE 8.1. Characteristics of Channels

Type Gating Unitary Blockers

Na+ Depolarization 10 pS Tetrodotoxin
K+ (DR) Depolarization 55 pS Ba2+
K+ (Ca2+) Depolarization/Ca2+ 240 pS Charybdotoxin/Ba2+
Ca2+ (L) High-threshold 9 pS (Ca2+) Dihydropyridines
(slow inact.) depolarization 25 pS (Ba2+)
Ca (T) Low-threshold 8 pS (Ba2+)
(fast inact.) depolarization 8 pS (Ca2+)
Ca (N) High-threshold 12 pS (Ba2+) ω-conotoxin GVIA
Ca2+ (P) Depolarization ω-agatoxin IVA
Nicotinic Acetylcholine 90 pS Bungarotoxin
Ryanodine Ca2+
InsP3 Receptor Inositol trisphosphate

Section 8.5. Channel Conductance

As described in Section 7.1, the flow of an electrolyte across a membrane
(expressed as a current I = -JFz) depends on the membrane potential and the concentration
gradient as described by equation 8.1:

zF ψm /RT zF ψm /RT
(8.1) I = -FzP ( exp (zFψm /RT) - 1 )( Co - C i e )
The current carried by ion channels is commonly expressed using a variation on Ohm's

(8.2) Ii = gi (ψm - ψi)

Page 8.3
Ii is the current carried by ion i, gi is the conductance of the membrane to that ion, ψm is the
membrane potential and ψi is the equilibrium potential for that ion as defined by the Nernst
equation (equation 6.3). Adjusting the membrane potential by subtracting ψi insures that
there will be no current when the membrane potential equals the ion’s equilibrium
potential. Equation 8.2 predicts that a plot of current (Ii) against voltage (ψm), also known
as an IV curve, will be linear. The slope is the conductance gi and the X-intercept is the
equilibrium potential ψi.
If a channel is perfectly selective for a particular ion, that channel will exhibit an IV
curve as defined by equation 8.2. Often, however, a channel is not absolutely selective and
will pass different kinds of ions with different conductances. In this case, the current
through the channel must be summed over the different ions:

(8.3) I = ∑ [gi (ψm - ψi)] = ψm ∑ gi - ∑ (gi ψ i) = ∑ gi {ψm - ∑ (gi ψ i) / ∑ gi }

For a non-selective channel, Ohm's Law still holds, but the conductance is the sum of the
conductances for all of the ions, and the X-intercept is a weighted average of the
equilibrium potentials. The X-intercept is called the reversal potential because the current
through the channel reverses direction as the membrane potential crosses this value. If the
channel is highly selective for a particular ion, the reversal potential will be close to the
equilibrium potential for that ion. Thus, the reversal potential is an indicator of a channel's
ion selectivity. Note that the reversal potential of a channel depends on the channel's
selectivity and also on the equilibrium potentials (concentration gradients) of the ions it
The functioning of individual channels has been illuminated in recent years by the
development of the patch clamp technique. This technique, which permits the observation
of currents through single channels, has revealed the opening and closing times for
different channel types. It has also enabled measurement of unitary conductances. The
unitary conductance is important because it reflects the restriction imposed by the channel's
selectivity filter. For example, the sodium channel, which has a relatively low unitary
conductance, probably has a relatively long narrow tunnel. It is estimated that this is a pore
3 Å x 5 Å in cross section and 10-12 Å in length. The high conductance K channel, by
contrast, probably has a relatively short narrow tunnel. In both cases, it is imagined that
the channel has large vestibules on one or both sides of the selectivity filter. This permits
free (and rapid) diffusion through most of the membrane and limits conductance only to the
extent necessary to maintain selectivity.

Section 8.6. Channel Selectivity

An important property of ion channels is that they exhibit selectivity for particular
ions over other closely related ions. Eisenman examined the five monovalent metal cations
(Cs+, Rb+, K+, Na+, Li+) and noted that, although there are 120 possible selectivity
sequences, only 11 sequences are observed. He rationalized this by arranging the eleven
sequences in order from low field strength to high field strength:

Page 8.4
TABLE 8.2. The Eisenman Selectivity Series for Monovalent Cations

I Cs > Rb > K > Na > Li Low field strength

II Rb > Cs > K > Na > Li
III Rb > K > Cs > Na > Li
IV K > Rb > Cs > Na > Li
V K > Rb > Na > Cs > Li
VI K > Na > Rb > Cs > Li
VII Na > K > Rb > Cs > Li
VIII Na > K > Rb > Li > Cs
IX Na > K > Li > Rb > Cs
X Na > Li > K > Rb > Cs
XI Li > Na > K > Rb > Cs High field strength

To cross through a channel, an ion must shed its water of hydration and pass a selectivity
site in the channel. This implies that the ion must interact more favorably with the
selectivity site than with its water of hydration. At low field strength, the selectivity of the
channel is dominated by the dehydration energy of the ion. Ions that are easily dehydrated
pass through the channel more readily than do ions that interact strongly with water. At
low field strength, therefore, larger ions are favored over smaller ions and the order is in
the direction of decreasing atomic weight. At high field strength, the selectivity of the
channel is dominated by the attraction of the ion for the selectivity site. Those ions that
bind well will be favored over those that bind poorly. At high field strength, therefore,

Intermediate-field-strength site
High-field-strength site Hydration energy

Low-field-strength site

Cs+ Rb+ K+ Na+ Li +

Figure 8.1. Competition between hydration and selectivity-site binding
in ion channel selectivity.

Page 8.5
smaller ions are favored over larger ions and the order is in the direction of increasing
atomic weight. The sodium channel has selectivity characteristics typical of a high field
strength. This high field strength is attributed to one or more carboxylic acid groups that
may line the tunnel and interact with the passing cations.

TABLE 8.3. Properties of Ions

Ion Atomic Ionic Crystal ∆Hhydration

Weight Radius (Å) (kcal/mole)
Li+ 6.94 0.68 121
Na+ 23.00 0.98 95
K+ 39.10 1.33 76
Rb+ 85.47 1.48 69
Cs+ 132.91 1.67 62

TABLE 8.4. Selectivity of Channels

Type Selectivity
ACh Receptor NH4+ > Cs+ > Rb+ > Na+
Na+ Na+,Li+ > K+ > Rb+ > Cs+
K+ (DR) Tl+ > K+ > Rb+ > NH4+ > Na+ > Li+ >> Cs+
K+ (Ca2+) Tl+ > K+ > Rb+ > NH4+ > Na+, Li+, Cs+
Ca2+ (L) Ca2+ > Sr2+ > Ba2+ > Li+ > Na+ > K+ > Cs+ > Mg2+
Ca2+ (T) Ca2+ ≈ Ba2+

Section 8.7. Excitability

As defined by the Goldman-Hodgkin-Katz equation (Equation 7.32), the membrane
potential is determined by the relative permeability of the membrane to different ions.
Because animal cell membranes are relatively more permeable to K+ than to Na+, the
resting membrane potential is normally close to the K+ equilibrium potential and is
negative (inside relative to outside). When a nerve or muscle fiber is stimulated and
conducts and action potential, the membrane depolarizes (the membrane potential becomes
less negative and then positive) and then repolarizes to the resting membrane potential.
This happens because Na+ channels open first and then K+ channels open. When Na+
channels open, the membrane becomes more permeable to Na+ than to K+ and the
membrane potential shifts toward the Na+ equilibrium potential (or more accurately, the
reversal potential of the Na+ channel). When the K+ channels open, the membrane again
becomes more permeable to K+ than to Na+ and the membrane potential shifts back toward
the reversal potential of the K+ channel.
Na+ and K+ channels open and close in the course of an action potential because
their opening (gating) depends on the membrane potential. Of course, the membrane
potential in turn depends on the opening of the channels making the action potential an
autocatalytic event. To analyze this, recognize that the currents across the membrane all
contribute to a change in the capacitative charge:

(8.4) dqc/dt = - ( IK + INa + Il )

Page 8.6
Equation 8.4 includes currents of K+ and Na+ as well as a small residual current (mostly Cl-
), which is commonly known as the “leakage current.” Equation 8.4 may be expanded
knowing that the membrane potential is related to the capacitative charge by the
capacitance (Equation 4.11), and the currents are given by Ohm’s Law (Equation 8.2):

(8.5) C (dψm/dt) = - gK(ψm-ψK) – gNa(ψm-ψNa) – gl (ψm-ψl)

Equation 8.5 makes it clear that the change in the membrane potential ψm depends on the
conductances of the sodium and potassium channels. The sodium and potassium channels
in turn are gated by voltage, so their conductances change depending on the membrane
potential. To separate the interdependence of channel conductance and membrane
potential, Hodgkin and Huxley used the voltage-clamp technique, in which the membrane
potential is fixed or clamped so that Na+ and K+ currents can be recorded at that constant
membrane potential. They found that Na+ and K+ channels remain closed at the resting
membrane potential but their probability of opening increases when the membrane
potential is raised. When the membrane is depolarized, Na+ channels open quickly and
then inactivate or close. K+ channels open more slowly and remain open. This means that
depolarizing a membrane will cause Na+ channels to open first. The inward Na+ current
will drive the membrane potential in the positive direction toward the Na+ equilibrium
potential. This accelerates the opening of Na+ channels ensuring a strong depolarization.
After a brief time, however, the Na+ channels inactivate and the Na+ current stops.
Concurrently, K+ channels open, and the outward K+ current drives the membrane potential
back down toward the K+ equilibrium potential. This brings the membrane potential back
to the resting value and also causes the K+ channels to close terminating the action
From data gathered in their voltage clamp experiments, Hodgkin and Huxley found
empirical relations that express the changes in conductances as functions of the membrane
potential. These functions are quite complicated and for our purposes need only be
represented as f1(gK, ψm) and f2(gNa, ψm).

(8.6) dgK/dt = f1(gK, ψm)

(8.7) dgNa/dt = f2(gNa, ψm)

Knowing how the membrane potential depends on conductances (equation 8.5) and
how the conductances depend on membrane potential (equations 8.6 and 8.7), it is possible
to reconstruct the action potential. A membrane action potential is an action potential in
which the membrane potential changes along the whole length of the nerve fiber at the
same time. Thus, the entire nerve fiber depolarizes and repolarizes simultaneously. A
membrane action potential can be simulated by numerically integrating equations 8.5, 8.6
and 8.7. Assume that the membrane potential begins at a value somewhat above the
resting potential. Given this membrane potential, calculate the change in gNa, gK and ψm
expected to occur in a brief interval of time (say 10 µsec). After changing the values of
gNa, gK and ψm accordingly, calculate the changes expected in the next brief time interval
by repeating the process. If the initial membrane potential was set to a value above

Page 8.7
threshold, then the membrane potential will rise up in an action potential and then return to
rest. If the initial membrane potential was set to a value below threshold, then the
membrane potential will simply drift back down to its resting value.
The membrane action potential is easy to calculate, but normally an action potential
does not occur along an entire nerve fiber simultaneously. Rather, it travels from one end
to the other. To analyze this propagating action potential, recognize that the action
potential travels because charge diffuses down the nerve fiber, depolarizing the membrane
ahead of the action potential, and causing the action potential to move forward along the
fiber. The current along the length of the nerve fiber is given by Ohm’s Law:

(8.8) Ilong = - (1/r ) (dψm/dx)

The longitudinal current is proportional to the change in potential (dψm/dx) and

inversely related to the resistance of the nerve fiber r.
The current across the membrane now must equal the change in the current along
the membrane.

(8.9) Imem = (-1/πD) dIlong/dx

D is the diameter of the nerve fiber. Therefore,

(8.10) Imem = (1/πDr)(d2ψm/dx2) = (1/πDr) (d2ψm/dt2) (dt/dx)2

Recognize that dx/dt is the conduction velocity of the action potential (θ). Then,

(8.11) Imem = (1/ πDrθ2) (d2ψm/dt2) = C (dψm/dt) + gK (ψm - ψK) + gNa (ψm - ψNa +
gl(ψm - ψl)

Equation 8.11 may be numerically integrated as for the membrane action potential if the
correct value for the conduction velocity θ is chosen.


T. Begenisich (1987) Molecular properties of ion permeation through sodium channels,

Ann. Rev. Biophys. Biophys. Chem. 16, 247-263.
W.A. Catterall (1986) Molecular properties of voltage-sensitive sodium channels, Ann.
Rev. Biochem. 55, 953-985.
W.A. Catterall (1988) Structure and function of voltage-sensitive channels, Science 242,
G. Eisenman and J.A. Dani (1987) An introduction to molecular architecture and
permeability of ion channels, Ann. Rev. Biophys. Biophys. Chem. 16, 205-226.
A.L. Hodgkin and A.F. Huxley (1952) A quantitative description of membrane current and
its application to conduction and excitation in nerve, J. Physiol. 117, 500-544.
A.F. Huxley (1959) Ion movements during nerve activity, Ann. N.Y. Acad. Sci. 81, 221-

Page 8.8
H.A. Lester (1992) The permeation pathway of neurotransmitter-gated ion channels, Ann.
Rev. Biophys. Biomol. Struct. 21, 267-292.
P.S. McPherson and K.P. Campbell (1993) The ryanodine receptor/Ca2+ release channel, J.
Biol. Chem. 268, 13765-13768.
R.J. Miller (1992) Voltage-sensitive Ca2+ channels, J. Biol. Chem. 267, 1403-1406.
M. Mishina, T. Kurosaki, T. Tobimatsu, Y. Morimoto, M. Noda, T. Yamamoto, M. Terao,
J. Lindstrom, T. Takahashi, M. Kuno and S. Numa (1984) Expression of functional
acetylcholine receptor from cloned cDNA, Nature 307, 604-608.
M. Noda, T. Ikeda, T. Kayano, H. Suzuki, H. Takeshima, M. Kurasaki, H. Takahashi, and
S. Numa (1986) Existence of distinct sodium channel messenger RNAs in rat brain,
Nature 320, 188-192.
M. Noda, S. Shimizu, T. Tanabe, T. Takai, T. Kayano, T. Ikeda, H. Takahashi, H.
Nakayama, Y. Kanaoka, N. Minamino, K. Kangawa, H. Matsuo, M.A. Raftery, T.
Hirose, S. Inayama, H. Hayashid, T. Miyata and S. Numa (1984) Primary structure
of Electrophorus electricus sodium channel deduced from cDNA sequence, Nature
312, 121-127.
M. Noda, H. Takahashi, T. Tanabe, M. Toyosato, S. Kikyotani, Y. Furutani, T. Hirose, H.
Takashima, S. Inayama, T. Miyata and S. Numa, (1983) Structural homology of
Torpedo californica acetylcholine receptor subunits, Nature 302, 528-532.
B.M. Olivera, G. Miljanich, J. Ramachandran, and M.E. Adams (1994) Calcium channel
diversity and neurotransmitter release: The ω-conotoxins and ω-agatoxins, Ann.
Rev. Biochem. 63, 823-867.
R.W. Tsien, P. Hess, E.W. McCleskey and R.L. Rosenberg (1987) Calcium channels:
Mechanisms of selectivity, permeation, and block, Ann. Rev. Biophys. Biophys.
Chem. 16, 265-290.
G. Yellen (1987) Permeation in potassium channels: Implications for channel structure,
Ann. Rev. Biophys. Biophys. Chem. 16, 227-246.
C.D. Ferris, and S.H. Snyder (1992) Inositol 1,4,5-trisphosphate activated calcium
channels, Ann. Rev. Physiol. 54,

Page 8.9
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Active transport is usually defined as transport of molecules or ions from a region

of lower to a region of higher concentration (i.e., transport against a concentration
gradient). Here, we will define it in a stricter sense as transport which requires chemical or
photochemical energy. By this definition, only a small number of enzymes are capable of
catalyzing active transport. With few exceptions, these enzymes transport inorganic
cations: Na+, K+, Ca2+ and H+. Active transporters may be divided into four general
classes: 1) ion-translocating ATPases, 2) H+-transporting electron transfer chains, 3)
group-translocating enzymes, and 4) photochemically-driven transporters such as
bacteriorhodopsin. It should be recognized that these enzymes do not consume energy;
rather they transduce chemical energy into electrochemical potential gradients.

Section 9.1. Ion-translocating ATPases

Ion-translocating ATPases derive energy from the hydrolysis of ATP and use this
energy to move ions across the membrane against concentration gradients. The ATPases
fall into two broad classes: the P (or E1E2) type and the F/V/A type. The P-type include
the widely studied Na+/K+ ATPase found in the plasma membrane of animal cells, the Ca2+
ATPases found in sarcoplasmic and endoplasmic reticulum (SERCA) and in the plasma
membrane (PMCA), the H+/K+ ATPase which is involved in acid extrusion, and an H+
ATPase which is found in the plasma membranes of plant and fungal cells. The P-type
ATPases are all inhibited by vanadate, a characteristic which seems to be shared by
ATPases that are phosphorylated in the course of their catalytic cycle. The phosphate
group hydrolyzed from ATP is transiently attached to an aspartyl residue on the enzyme.
The P-type ATPases characteristically have a single, large polypeptide chain, which
mediates both ion transport and ATP hydrolysis. In general, the catalytic segment appears
to lie near the middle of this chain with 4 membrane-spanning domains on the amino end
and four to six membrane-spanning domains on the carboxyl end. Thus, the amino
terminus and the catalytic site both lie on the cytoplasmic side of the cell membrane.
The P-type ATPases hydrolyze cytosolic ATP and transport the substrate out of the
cytosol (out of the cell or into the sarcoplasmic or endoplasmic reticulum). The Na+/K+
and H+/K+ ATPases are exceptions in that they mediate an exchange of ionic substrates.
Both enzymes bring K+ into the cytosol as the other ion (Na+ or H+) is transported out.
These two enzymes are also unusual in having a second (β) subunit. They are thus
heterodimers, and the β subunit is thought to function in K+ transport.
In the catalytic cycle of the Na+/K+ ATPase, 3 Na+ ions bind from the cytosolic side
of the enzyme. ATP hydrolysis phosphorylates the enzyme and initiates a conformational
change exposing the ion binding sites to the external side of the membrane. The Na+ ions
dissociate and are replaced by 2 K+ ions. Then, the enzyme dephosphorylates and returns
to the original conformation releasing the K+ ions to the inside. It is important that ATP
hydrolysis not occur until 3 Na+ ions are bound and that dephosphorylation not occur
unless 2 K+ ions are bound. Otherwise, ATP hydrolysis and ion transport would not be
properly coupled, and ATP would be spent inefficiently on suboptimal transport cycles.
The F/V/A-type ATPases include the F (F1FO) ATPases found in mitochondria,
chloroplasts and bacterial cells, the vacuolar-type ATPases found in Golgi-derived
organelles including lysosomes and secretory vesicles in animal cells and vacuoles and
tonoplasts in plant and fungal cells, and the A-type ATPases found in archaebacteria.

Page 9.1
These ATPases transport H+ by a rotary mechanism and have a two-part structure including
a membrane stalk and a large head. They differ from the plasma membrane ATPases in
being insensitive to vanadate and in lacking a stable phosphorylated intermediate. The
subunit composition of the F, V and A ATPases is also much more complex than that of
the P-type ATPases. They are characteristically sensitive to N,N'-
dicyclohexylcarbodiimide (DCCD), which reacts with carboxyl groups and typically
inhibits proton-translocating enzymes.
The F or F1/FO class of H+-ATPases are found in mitochondria, chloroplasts and
bacterial cells and are, of course, responsible for ATP synthesis rather than ATP
hydrolysis. Nevertheless, they should be considered here because they are functionally
similar to the V and A type ATPases, and they can catalyze active transport of H+ when the
free energy of the reaction favors that direction. These enzymes have a complex subunit
composition including a hydrophilic F1 component which carries the catalytic site and a
hydrophobic FO component which is thought to be the proton channel. The F1 component
includes five kinds of subunits (α, β, γ, δ, ε) in a stoichiometry of 3:3:1:1:1. The
nucleotide binding sites lie between the α and β subunits. At any given time, these sites
are in the tight, loose or open state depending on the position of the γ subunit. The FO
component includes three kinds of subunits (a, b, c) in a stoichiometry of 1:2:~12. DCCD
inhibits the F1FO ATPases by reacting with a glutamate residue in the c subunit. The c
subunits are relatively small (2 transmembrane segments) and the cluster of c subunits
together with the γ subunit form an axle around which the α3β3 head rotates. In the ATP
hydrolysis mode, the F1FO ATPase is thought to function as follows: ATP hydrolysis
drives the rotation of the α3β3 head group relative to the γ subunit and the c cluster attached
at the base. This rotation causes the three nucleotide binding sites to step alternately
through the loose, tight and open conformations. In the process, ATP is bound, hydrolyzed
and the products released. Concomitantly, 3 H+ are pumped through the FO component for
each ATP hydrolyzed.
The V or vacuolar ATPases are similar in structure to the F ATPases, although their
normal function is H+ transport driven by ATP hydrolysis rather than ATP synthesis driven
by H+ movement. The V ATPases differ from the F ATPases in that they are inhibited by
the antibiotic bafilomycin A1 and by N-ethylmaleimide.
ATPases couple a chemical reaction to vectorial movement of an ion. For this to
work, three criteria must be met. 1) The energetics must be such that the chemical reaction
releases more free energy than is required for ion translocation. 2) The chemical reaction
and ion translocation must be obligatorily coupled processes. One cannot be allowed to
occur in the absence of the other. 3) The enzyme must change its affinity for the ions
during the translocation cycle.
First, let us consider the energetics. For ion transport, the change in the
electrochemical potential of the transported ion is:

(9.1) ∆µj = µj,final - µj,initial

= µj° + RT ln Cj,final + zjFψfinal - µj° - RT ln Cj,initial - zjFψinitial

= RT ln (Cj,final/Cj,initial) + zjF(ψfinal- ψinitial)

Page 9.2
The total free energy change required for transport is the sum of the electrochemical
potential differences (∆µj) for each of the transported ions multiplied by the stoichiometry
nj for that ion:

(9.2) ∆Gions = Σj nj ∆µj

If we define nj as positive for ions transported from outside to inside and negative for ions
transported from inside to outside, then

(9.3) ∆µj = RT ln (Cj,in/Cj,out) + zjF(ψin- ψout)

This convention simplifies the bookkeeping of directionality and allows us to use the
membrane potential (ψin- ψout) without worrying about the sign. Now, for the chemical
reaction, the free energy change is

(9.4) ∆GATP = ∆G°ATP + RT ln ([ADP][Pi]/[ATP])

where ∆G°ATP = -7.3 kcal/mol. Of course, ion transport will proceed as long as

(9.5) ∆GATP + ∆Gions ≤ 0

Equation 9.5 emphasizes that the ATPases catalyze coupled reactions. It would not do for
either ATP hydrolysis or ion movement to occur independently. If ATP hydrolysis
occurred independently of ion movement, then the free energy released by ATP hydrolysis
would be wasted. If ion movement were allowed to proceed in the absence of ATP
hydrolysis, then the ions would move in the wrong direction: from higher concentration to
lower concentration.
To catalyze ion movement against a concentration gradient, the enzyme must take
the ion from a solution in which the ion is present at a lower concentration and release the
ion into a solution in which the ion is present at a higher concentration. Thus, in addition
to moving the ion, the enzyme must change its affinity for the ion in order to function
efficiently. The ATPase must have a high affinity for the ion on the low-concentration side
and a low affinity for the ion on the high-concentration side. Part of the energy released by
ATP hydrolysis, therefore, must go into changing the enzyme's affinity for the ion.

Section 9.2. Electron-transfer chains

In mitochondria, chloroplasts and bacterial cells, substrate oxidation is used to
create an H+ gradient. This was first suggested by Peter Mitchell in 1961 as the celebrated
"Chemiosmotic Hypothesis." We will discuss some of the details of this energy coupling
later. Here, let us just review the energetics of redox-driven H+-translocation.
In electron transfer reactions, an electron is passed from an electron donor to an
electron acceptor.

Donorred + Acceptorox → Donorox + Acceptorred

The free energy change associated with this reaction may be expressed in the usual way:

Page 9.3
(9.6) ∆Gredox = ∆G° + RT ln ( [Donor] red [Acceptor ]ox
Rather than tabulate ∆G° values for all combinations of electron carriers, however, electron
affinities are catalogued by considering each half- reaction separately:

Donorred → Donorox + n e-

Acceptorox + n e- → Acceptorred

The reduction potential E of the donor is defined as

(9.7) E = E° + (RT/nF) ln {[Donor]ox/[Donor]red}

where n is the number of electrons transferred in the reaction and E° is the midpoint or
standard reduction potential for the electron donor. A similar equation defines the
reduction potential of the electron acceptor. The reduction potentials are related to the free
energy change as

(9.8) ∆Gredox = nF Edonor - nF Eacceptor

= nF (E°donor -E°acceptor) + RT ln ( [Donor]red [Acceptor ]ox
Since E is measured in volts, ∆G will be in units of joules/mole.
The best characterized redox-driven proton-pump is cytochrome oxidase, the
enzyme catalyzing the transfer of electrons from cytochrome c to oxygen. Mitchell
originally suggested that cytochrome oxidase accepted electrons from cytochrome c
externally and that it carried the electrons across the membrane to O2. Since the reduction
of O2 to H2O takes up two protons, Mitchell hypothesized that this enzyme contributed to
proton translocation simply by removing H+ from the mitochondrial matrix. Since E° for
cytochrome c is +0.24 volts and E° for O2/H2O is +0.816 volts, this electron transfer step
releases a lot of energy and one might expect it to be captured to a greater extent. In 1978,
Krab and Wikstrom showed that this is the case; cytochrome oxidase not only takes up
protons because of the reduction of oxygen but it also physically transports protons across
the membrane.
Cytochrome oxidase contains two hemes (a and a3) and two copper atoms. One Cu
and heme a are on the cytoplasmic side of the inner mitochondrial membrane and accept
electrons from cytochrome c. The other Cu and heme a3 are on the matrix side of the
membrane and react with O2. Cytochrome oxidase from bovine heart consists of eight
subunits. Three (I, II and III) are coded by the mitochondrial DNA and the rest are coded
by nuclear DNA.

Page 9.4
Section 9.4. Bacteriorhodopsin
Bacteriorhodopsin is a light-driven proton pump. It is found in Halobacterium
halobium, a bacterium found growing in hot salt springs. In the presence of oxygen,
Halobacterium uses respiration as a source of metabolic energy. Under anaerobic
conditions, however, Halobacterium uses light energy to produce ATP directly.
Halobacterium is not photosynthetic since it does not split H2O (produce O2) or exhibit
light-driven electron flow. Instead, it produces a "purple membrane" which is essentially
pure bacteriorhodopsin: 25% lipid and 75% bacteriorhodopsin. Bacteriorhodopsin pumps
H+ out of the cell in the presence of light. This sets up an H+ gradient, just as respiration
would, and the proton gradient can be used to make ATP in the usual chemiosmotic way.
Bacteriorhodopsin is a particularly well characterized ion pump, because it can be
obtained in large quantity and essentially pure. Moreover, it is a rather simple protein
having a molecular weight of 26,000. The amino acid sequence has been determined
(Ovchinnikov et al., 1979) and its three-dimensional structure elucidated (Henderson and
Unwin, 1975). Bacteriorhodopsin consists of seven membrane-spanning domains. A
cofactor, retinal, is responsible for light absorption. Retinal is covalently bound to the ε
amino of a lysine residue via a Schiff's base linkage.
The energy of a quantum of light depends on the wavelength (λ):

(9.10) Elight = hν = hc/λ

where h is Planck's constant and c is the speed of light. For an einstein (mole) of quanta,
the energy is

(9.11) ∆Glight = hcN/λ

The value of hcN is 28,592 nm.kcal/einstein.

Section 9.5. Group Translocation

Group translocation is a transport mechanism in which the substrate is metabolized
as it is transported. The energy released by that metabolism is used to drive transport. For
example, glucose is brought into some bacterial cells as glucose 6-phosphate.
Phosphoenolpyruvate serves as the source of both the energy and the phosphate. There are
three proteins involved: Enzyme I, Enzyme II, and HPr. Enzyme I is a soluble protein that
catalyzes the phosphorylation of HPr using phosphoenolpyruvate. P-HPr then donates the
phosphate to glucose bound to enzyme II. Enzyme II is specific for the sugar. This system,
studied by Saul Roseman, is found in E. coli, Salmonella and Staphylococcus.


J.P. Abrahams, A.G.W. Leslie, R. Lutter, and J.E. Walker (1994) Structure at 2.8A
resolution of F1- ATPase from bovine heart mitochondria, Nature 370, 621-628.
R. Addison, (1986) Primary structure of the Neurospora plasma membrane H+-ATPase
deduced from the gene sequence, J. Biol. Chem. 261, 14896-14901.

Page 9.5
K.B. Axelsen and M.G. Palmgren (1998) Evolution of substrate specificities in the P-type
ATPase superfamily, J. Mol. Evol. 46, 84-101.
P.D. Boyer, (1997) The ATP Synthase - a splendid molecular machine, Ann. Rev. Biochem.
66, 717-
T. Elston, H. Wang and G. Oster, G. (1998) Energy transduction in ATP synthase, Nature
391, 510-513.
S.P.A. Fodor, J.B. Ames, R. Gebhard, E.M.M. van den Berg, W. Stoeckenius, J.
Lugtenburg and R.A. Mathies (1988) Chromophore structure in bacteriorhodopsin's
N intermediate: Implications for the proton-pumping mechanism, Biochemistry 27,
M. Forgac (1999) Structure and properties of the vacuolar (H+)-ATPases, J. Biol. Chem.
274, 12951-12954.
K.M. Hager, S.M. Mandala, J.W. Davenport, D.W. Speicher, E.J. Benz, Jr. and C.W.
Slayman (1986) Amino acid sequence of the plasma membrane ATPase of
Neurospora crassa: Deduction from genomic and cDNA sequences, Proc. Natl.
Acad. Sci. USA 83, 7693-7697.
P. Henderson and P.N.T. Unwin (1975) Three-dimensional model for purple membrane
obtained by electron microscopy, Nature 257, 28-32.
E. Hilario and J.P. Gogarten (1998) The prokaryote-to-eukaryote transition reflected in the
evolution of the V/F/A-ATPase catalytic and proteolipid suunits, J. Mol. Evol. 46,
P.L. Jorgensen, J.M. Nielsen, J.H. Rasmussen, and P.A. Pedersen (1998) Structure-
function relationships of E1-E2 transitions and cation binding in Na, K-pump
protein, Biochim, Biophys. Acta 1365, 65-70.
P.M. Kane (1999) Vacuolar ATPases: structure, function, assembly and biosynthesis, J.
Bioenerg. Biomembr. 31, 1-83.
K. Kinosita, Jr., R. Yasuda, H. Noji, S. Ishiwata and M. Yoshida (1998) F1-ATPase: a
rotary motor made of a single molecule, Cell 93, 21-24.
K. Krab and M. Wikström (1987) Principles of coupling between electron transfer and
proton translocation with special reference to proton-translocation mechanisms in
cytochrome oxidase, Biochim. Biophys. Acta 895, 25-39.
S. Lutsenko and J.H. Kaplan (1995) Organization of P-type ATPases: Significance of
structural diversity, Biochemistry 34, 15607-15613.
D.H. MacLennan, C.J. Brandl, B. Korczak and N.M. Green (1985) Amino-acid sequence
of a Ca2++Mg2+-dependent ATPase from rabbit muscle sarcoplasmic reticulum,
deduced from its complementary DNA sequence, Nature 316, 696-700.
D.H. MacLennan, W.J. Rice and N.M. Green (1997) The mechanism of Ca2+ transport by
sarco(endo)plasmic reticulum Ca2+-ATPases, J. Biol. Chem. 272, 28815-28818.
M. Mohraz, M.V. Simpson, and P.R. Smith (1987) The three-dimensional structure of the
Na, K-ATPase from electron microscopy, J. Cell Biol. 105, 1-8.
J.V. Moller, B. Juul and M. LeMaire (1996) Structural organization, ion transport and
energy transduction of P-type ATPases, Biochim. Biophys. Acta 1286, 1-51.
V. Muller, C. Ruppert and T. Lemker (1999) Structure and function of the A1A0-ATPases
from methanogenic archaea, J. Bioenerg. 31, 15-27.
N. Nelson and W.R. Harvey (1999) Vacuolar and plasma membrane proton-
adenosinetriphosphatases, Physiol. Rev. 79, 361-385.

Page 9.6
D. Njus, P.M. Kelley, and G.J. Harnadek (1986) Bioenergetics of secretory vesicles,
Biochim. Biophys. Acta 853, 237-265.
H. Noji, R. Yasuda, M. Yoshida and K. Kinosita, Jr. (1997) Direct observation of the
rotation of F1-ATPase, Nature 386, 299-302.
Y.A. Ovchinnikov, N.G. Abdulaev, M.Y. Feigina, A.V. Kiselev, and N.A. Lobanov (1979)
The structural basis of the functioning of bacteriorhodopsin: An overview, FEBS
Lett. 100, 219-224.
M.G. Palmgren and K.B. Axelsen (1998) Evolution of P-type ATPases, Biochim. Biophys.
Acta 1365, 37-45.
G. Rudnick (1986) ATP-driven H+ pumping into intracellular organelles, Ann. Rev.
Physiol. 48, 403-413.
Y. Sambongi, Y. Iko, M. Tanabe, H. Omote, A. Iwamota-Kihara, I. Ueda, T. Yanagida, Y.
Wada and M. Futai, M. (1999) Mechanical rotation of the c subunit oligomer in
ATP synthase (F0F1): direct observation, Science 286, 1722-1724.
B. Schulenberg, R. Aggeler, J. Murray and R.A. Capaldi (1999) The γε-c subunit interface
in the ATP synthase of Escherichia coli: cross linking of the ε subunits to the c
subunit ring does not impair enzyme function, that of γ to c subunits leads to
uncoupling, J. Biol. Chem. 274, 34233-24237.
R. Serrano, M.C. Kielland-Brandt and G.R. Fink (1986) Yeast plasma membrane ATPase
is essential for growth and has homology with (Na++K+), K+- and Ca2+-ATPases,
Nature 319, 689-693.
G.E. Shull and J.B. Lingrell (1986) Molecular cloning of the rat stomach (H++K+)-ATPase,
J. Biol. Chem. 261, 16788-16791.
G.E. Shull, L.K. Lane and J.B. Lingrel (1986) Amino-acid sequence of the beta-subunit of
the (Na++K+)ATPase deduced from a cDNA, Nature 321, 429-431.
G.E. Shull, A. Schwartz and J.B. Lingrel (1985) Amino-acid sequence of the catalytic
subunit of the (Na++K+)ATPase deduced from a complementary DNA, Nature 316,
D. Stock, A.G.W. Leslie, and J.E. Walker (1999) Molecular architecture of the rotary
motor in ATP synthase, Science 286, 1700-1705.
C. Toyoshima, M. Nakasako, H. Nomura and N. Ogawa (2000) Crystal structure of the
calcium pump of sarcoplasmic reticulum at 2.6 Å resolution, Nature 405, 647-655.
J. Weber and A.E. Senior (1997) Catalytic mechanism of F1-ATPase, Biochim. Biophys.
Acta 1319, 19-58.

Page 9.7
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

As we have discussed, some molecules and ions cross biological membranes by

simple diffusion through the hydrophobic milieu and some pass through specific channels.
The passage of other molecules and ions across biological membranes is facilitated by
means of transporters. Except in those few cases in which the transporters draw upon
chemical or electromagnetic energy to transport ions against concentration gradients, this
facilitated transport is known as facilitated diffusion.

Section 10.1. Ionophores

Of the agents which mediate facilitated diffusion, the simplest are antibiotics
known as ionophores, hydrophobic compounds which can complex an ion and carry it
across a lipid bilayer. Ionophores are traditionally classified in two categories: neutral
ionophores such as valinomycin and carboxylic ionophores such as nigericin. To these two
categories, we should add a third: those agents that carry hydrogen ions across biological
membranes. H+ carriers are commonly called uncouplers of oxidative phosphorylation (or
simply uncouplers) because of their effect on mitochondrial respiration. Because their
effect is more precisely to transport protons across membranes, we will refer to them here
as protonophores.
The protonophores are weak acids capable of permeating the lipid bilayer in their

Carbonylcyanide p-trifluoromethoxyphenyl
2,4-Dinitrophenol hydrazone (FCCP)

CH3 CH 3
5-Chloro-3-tert-butyl-2'-chloro-4'-nitrosalicylanilide Carbonylcyanide m-chlorophenyl
(S-13) hydrazone (CCCP)
Figure 10.1. Structures of some protonophores

anionic form. Of course, they can also pass across the membrane in their uncharged
(protonated) form, so they carry H+ across the membrane and bring H+ concentrations on

Page 10.1
the two sides of the membrane to electrochemical equilibrium. In terms of structure, the
protonophores are generally aromatic compounds that are both hydrophobic and capable of
distributing the negative charge over a number of atoms in the molecule. Structures of
some protonophores are shown in Figure 10.1.
Neutral ionophores generally have a cyclic structure so they are effectively
hydrophobic doughnuts capable of complexing a cation in the hole. Valinomycin, for
example, is a ring composed of three units each of D-hydroxyisovalerate, L-valine, L-
lactate and D-valine linked by alternating peptide and ester bonds (Figure 10.2).
Valinomycin complexes K+, Rb+ or Cs+, but Na+ and Li+ are too small to be effectively
bound. The selectivity for K+ is 10,000 times greater than the selectivity for Na+.
Valinomycin will cross the membrane either with or without a bound ion so it carries
charge across the membrane. Ion transport mediated by valinomycin, therefore, depends
upon the membrane potential. Moreover, valinomycin will create a membrane potential by
transporting capacitative charge.

Enniatin B
Figure 10.2. Some neutral ionophores

Unlike the neutral ionophores, the carboxylic ionophores have a linear structure
with a carboxyl group on one end and one or two hydroxyls on the other (Figure 10.3).
They cyclize by head-to-tail hydrogen bonding and will cross the membrane with the
carboxyl group either protonated or complexed to an ion. Nigericin, for example, will
cross the membrane carrying either H+ or K+. It functions, therefore, as a K+/H+ exchanger.
Because nigericin does not carry a net charge across the membrane, transport is not
affected by the membrane potential nor does it contribute to the creation of a membrane

Page 10.2
Section 10.2. Transporters
Integral membrane proteins which catalyze facilitated diffusion differ from
ionophores in some important ways. First, the proteins span the membrane creating a
pathway for facilitated diffusion. They do not diffuse across the membrane as do
ionophores. Although transporters are sometimes called carriers, the latter term suggests a
protein that binds to a substrate and moves with it through a medium, not a protein through
which the substrate moves. Therefore, the term carrier may be applied to ionophores but
should not be used for transporter proteins.
A second important characteristic of transporters is that they are inserted into the
membrane with a fixed orientation. Therefore, whereas the binding process for ionophores
is the same on both sides of the membrane, transporters may be asymmetric. In particular,
the binding constants observed on the two sides of the membrane may be quite different.
In fact, for transporters to achieve maximum efficiency, the transporter should have a
binding constant higher than the expected substrate concentration on one side of the
membrane and lower than the expected concentration on the other side. This will allow the
transporter to bind the substrate on one side and release it on the other.
Finally, transporters customarily facilitate coupled transport of two or more
substrates. That is, like nigericin, transporters facilitate either exchange diffusion of two or
more substrates or codiffusion of two or more substrates. For example, the sodium/hexose
transporter, commonly found in the plasma membrane of animal cells, mediates
codiffusion of Na+ and a hexose molecule (glucose) into the cell (Figure 10.4). As we shall
see, this coupled transport is perhaps the most important mechanism for transporting
substances across biological membranes.

Page 10.3
Outside Inside

Na +

Figure 10.4. Codiffusion mediated by the sodium/hexose transporter

Section 10.3. Equilibria of Facilitated Diffusion

The ionophores and transporters mediating facilitated diffusion will reach
equilibrium when the free energy change for the processes they mediate is zero. For
transport processes

(10.1) ∆G = Σi ni ∆µi = Σ ni (µiin - µiout)

= Σi ni [RT ln (Cin/Cout) + ziF ψm]

ni is the stoichiometry for transport of substrate i and is positive if i is transported inward

and negative if i is transported outward.
For protonophores, which carry only H+,

(10.2) ∆G = RT ln ([H+]in/[H+]out) + F ψm

We obtain the equilibrium reached by protonophores by setting ∆G = 0. Not surprisingly,

this yields the Nernst equation:

(10.3) ψm = (RT/F) ln ([H+]out/[H+]in)

For nigericin, which exchanges K+ and H+,

(10.4) ∆G = RT ln ([H+]in/[H+]out) + Fψm - RT ln ([K+]in/[K+]out) - Fψm

By setting ∆G = 0, we obtain

Page 10.4
(10.5) [H+]in/[H+]out = [K+]in/[K+]out

Nigericin will reach equilibrium when the [H+] and [K+] gradients are proportional.
For the sodium/hexose transporter,

(10.6) ∆G = RT ln ([H]in/[H]out) + RT ln ([Na+]in/[Na+]out) + F ψm

Upon setting ∆G = 0, we obtain the equilibrium condition:

(10.7) [H]in/[H]out = [Na+]out/[Na+]in exp (-F ψm /RT)

Section 10.4. Kinetics of Facilitated Diffusion

The kinetics of facilitated diffusion may be analyzed by considering the simple
system illustrated in Figure 10.5. Let us imagine that this facilitated diffusion is mediated
by an ionophore so that we may make the following assumptions:
1. The rate constants for transmembrane movement of the liganded ionophore (k1)
are the same in both directions.
2. The rate constants for transmembrane movement of the empty ionophore (k2) are
the same in both directions.
3. The binding constants (K) on both sides of the membrane are the same.
Finally, we will make the important assumption of rapid equilibrium. That is, we
will assume that the rate limiting steps are the ones involving transmembrane movement
and that the binding processes on the two surfaces of the membrane occur relatively
quickly. This allows us to assume that the binding reactions are always at equilibrium.

(10.8) K = To Co/TCo = Ti Ci/TCi

The rate at which the total internal transporter concentration changes is

(10.9) d(TCi+Ti)/dt = k1TCo + k2To - k1TCi - k2Ti

If we assume that the transporter distribution reaches a steady-state, then d(TCi+Ti)/dt = 0,


(10.10) 0 = TCo (k1 + k2K/Co) - TCi (k1 + k2K/Ci)

Multiplying both sides of equation 10.10 by KCi + 2CiCo + KCo gives

(10.11) 0 = TCo[(k1Co + k2K)(Ci+K) + (K+Co)(k1Ci + k2KCi/Co)]

- TCi[(k1Ci + k2K)(Co+K) + (K+Ci)(k1Co + k2KCo/Ci)]

If we let T be the total amount of transporter in the membrane, then

(10.12) T = TCo + To + TCi + Ti

= TCo (1 + K/Co) + TCi (1 + K/Ci)

Page 10.5
Multiplying both sides of equation 10.12 by k2K(Co - Ci) gives

(10.13) k2KT(Co-Ci) = TCo(k2K-k2KCi/Co)(K+Co)

- TCi(k2K-k2KCo/Ci)(Ci+K)

Upon adding equations 10.11 and 10.13, we obtain

(10.14) k2KT(Co-Ci) = (TCo-TCi)[(k1Co + k2K)(Ci+K)+(K+Co)(k1Ci+k2K)]

Note that the flow of substrate J = k1 (TCo-TCi) so

(10.15) J =
(k1Co + k2K)(Ci+K) + (k1Ci + k2K)(Co+K)

This expression gives us the steady-state flux in terms of three parameters: k1T, K, and the
ratio k1/k2. If we restrict our consideration to the initial velocity of transport (Ci = 0), then
equation 10.15 reduces to

(10.16) J = (k1k2CoT)/(k1Co + 2k2K + k2Co)

Equation 10.16 may be rewritten as

(10.17) 1/J = (k1+k2)/k1k2T + 2K/k1CoT

If we define

(10.18) Jmax = k1k2T/(k1 + k2)

(10.19) Km = 2Kk2/(k1 + k2)


(10.20) 1/J = 1/Jmax + Km/JmaxCo

Initial velocity of transport, therefore, can be described by only two parameters, Jmax and
Km. As in enzyme kinetics, Jmax is the velocity at infinite substrate concentration; Km is the
substrate concentration giving one-half the maximum velocity. Note that the Km measured
for initial velocity of transport is equal to K, the binding constant for the substrate, only if
the rate constants (k1 and k2) are about equal.

Page 10.6
Outside Membrane Inside

TC o TC i

Co Ci

To Ti

Figure 10.5. A simple model of facilitated diffusion

Section 10.5. Facilitated Diffusion Mediated by Ionophores

McLaughlin and his colleagues have analyzed the kinetics of H+ transport mediated
by the ionophores FCCP and CCCP (Benz and McLaughlin, 1983; Kasianowicz et al.,
1984). In the absence of a membrane potential, the model illustrated in Figure 10.5 applies
at least in the short term. Values for the rate constants k1 and k2 and the binding constant
(expressed as pK) are given below:

k1 10,000 sec-1 12,000 sec-1
k2 700 sec-1 175 sec-1
pK 6.0 5.9

First, note that protonation reactions occur on a time scale of 1011 sec-1 so the
transmembrane rates (k1 and k2) are clearly limiting. Second, note that the rate constants
for FCCP are larger than those for CCCP consistent with the more potent protonophoric
activity of FCCP. The simple model shown in Figure 10.5 applies to protonophores with
two restrictions. First, it applies only in the absence of a membrane potential since a
membrane potential will change the rate constants for the movement of the anionic species.
In fact, because the membrane potential will facilitate the movement of the anionic species
in one direction and slow its movement in the other, k2 will not be the same in both
directions. This introduces an additional parameter. Second, the model applies only for
short time periods. Because the ionophore slowly partitions between the aqueous and
membrane phases, the membrane-bound concentration may not remain constant over
longer time periods.

R. Benz and S. McLaughlin (1983) The molecular mechanism of action of the proton
ionophore FCCP (carbonylcyanide p-trifluoromethoxy phenylhydrazone), Biophys.
J. 41, 381-398.

Page 10.7
J. Kasianowicz, R. Benz and S. McLaughlin (1984) The kinetic mechanism by which
CCCP (carbonyl cyanide m-chlorophenylhydrazone) transports protons across
membranes, J. Membrane Biol. 82, 179-190.
R.I. Macey (1978) Mathematical models of membrane transport processes, in Membrane
Physiology (T.E. Andreoli, J.F. Hoffman, and D.D. Fanestil, eds.), Plenum, New
York, pp. 125-146.
S.G.A. McLaughlin and J.P. Dilger (1980) Transport of protons across membranes by
weak acids, Physiol. Rev. 60, 825-863.
B.C. Pressman (1973) Properties of ionophores with broad range cation selectivity, Fed.
Proc. 32, 1698-1703.
B.C. Pressman (1976) Biological applications of ionophores, Ann. Rev. Biochem. 45, 501-
B.C. Pressman and M. Fahim (1982) Pharmacology and toxicology of the monovalent
carboxylic ionophores, Ann. Rev. Pharmacol. Toxicol. 22, 465-490.

Page 10.8
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

In 1961, Peter Mitchell (1961) proposed that the H+ concentration gradient drives
the synthesis of ATP. In the same year, Crane (1965) showed that the sodium
concentration gradient drives sugar and amino acid transport across epithelial cell plasma
membranes. In the years since, it has become evident that ion concentration gradients are
an important energy store in all cells. Ion pumps are therefore crucial transducers that
transform chemical-bond energy into electrochemical potential energy. Coupled transport,
in turn, uses the electrochemical potential of one substrate to drive the transport of a

Section 11.1. The Master Pump Concept

It is obviously possible to imagine all sorts of combinations of coupled transport.
Fortunately, this picture may be greatly simplified by applying the concept of a master
pump. According to this concept, each cellular membrane possesses one master pump
which uses chemical energy to transport one or two inorganic ions. This pump creates
transmembrane gradients in the electrochemical potential of the transported ions across the
membrane in which it is located. Other ions and molecules are transported across that
membrane by coupling their movement to the movement of the transported ion(s).
It is evident that a given pump can store electrochemical potential energy only
across the membrane in which it is located. Thus, the master pump must be the
predominant active transport enzyme in the membrane. For example, the Na+/K+ ATPase,
which is located in the plasma membrane of animal cells, can store energy only across the
plasmalemma. Accordingly, transport of other ions and molecules across animal cell
plasma membranes is likely to be coupled to Na+ or K+. Similarly, transport across
bacterial cell, chloroplast thylakoid and mitochondrial inner membranes is likely to be
coupled to the H+ gradients generated by the redox chains in those membranes. In Golgi-
derived organelles in both plant and animal cells, the master pump is the V-type H+-
translocating ATPase, while, in plant and fungal cell membranes, the master pump appears
to be the P-type H+-translocating ATPase.
A master pump must have three attributes: 1) high capacity, 2) high efficiency, and
3) low dissipation. Low dissipation is probably the reason that pumps almost exclusively
transport the relatively impermeant inorganic cations. The exception is H+ which, because
of its small size, is relatively permeant. The saving feature of H+ is that the rate of
dissipation (leakage current) is proportional to the concentration and H+ is present at
exceedingly low concentrations. Thus, because of low concentration, H+ gradients exhibit
a low dissipation.
High capacity is the requirement that the ion gradient involve concentrations that
are relatively large compared to the concentrations of the compounds that are to be
transported. Thus, if the master pump is to drive uptake of glucose and other metabolites
present at millimolar concentrations, the ion gradient established by the pump should be an
order of magnitude or two greater than this so the ion gradient is not greatly perturbed by
the secondary transport systems. For example, the Na+ concentration outside animal cells
is usually on the order of 100 mM while that inside is on the order of 10 mM.
Consequently, the Na+ gradient has a high capacity and is suitable for the transport of
millimolar concentrations of metabolites. By contrast, the Ca2+ concentration is on the
order of 5 mM outside animal cells and 10-7 M inside. While the concentration gradient is

Page 11.1
large, the capacity of the system is low. Ca2+ concentration gradients would be greatly
perturbed by the transport of millimolar concentrations of metabolites. The H+ ion is
anomalous because it is normally present intracellularly at very low concentrations (10-7
M). Although this is necessary to reduce dissipation of the gradient, it might be expected
to adversely affect capacity. H+, however, is unusual in that its concentration is buffered
and the capacity of the system is determined not by the absolute H+ concentration but by
the buffering capacity. Because the cytosol is well buffered, the capacity of the H+ gradient
is in the millimolar range eventhough the H+ concentration is micromolar or less.
The concept of a master pump is significant because it suggests a likely mechanism
for all transport systems in a given membrane once the master pump has been identified.
The master pump concept raises some interesting questions. First, why is transport
coupled to a single master pump? A number of possible reasons can be imagined.
Coupled transport may be more efficient than independent pumps. Ion gradients generally
store smaller packets of energy than ATP. By choosing the proper coupling stoichiometry,
the energy required for transport can be matched more precisely, thereby using energy more
efficiently. Coupled transporters may be inherently more efficient than pumps. An
indication that this may be so is provided by bacteria. Under anaerobic conditions, some
cells produce ATPases for transporting amino acids. However, under aerobic conditions,
the cells dispense with the amino acid pumps and transport amino acids by coupling
transport to the H+ gradient. Coupling transport to a single master pump may also serve a
control function. This is particularly evident in the case of sequential coupled transport
(tertiary and quaternary transport). For example, the mitochondrion does not need to
import citrate as a substrate for the TCA cycle if no phosphate is available for
phosphorylation. Because citrate transport is coupled to the phosphate gradient, citrate
transport automatically stops when phosphate is depleted.
There are two reasons why a transport system might not be coupled to the master
pump. First, if the transport system has a high capacity itself, it may adversely affect the
ion gradients established by the master pump. Second, if the transported substrate serves a
regulatory function, then it may be desirable to control its concentration separately.
Exceptions to the master pump concept may exist for both of these reasons. The K+/H+
ATPase is a possible example of an exception for the reason of high capacity. The ATPase
must transport large quantities of H+ into the lumen of the gut and kidney. If H+ movement
were coupled to the Na+/K+ ATPase, it might radically affect the magnitude of the Na+
gradient. The Ca2+ ATPase may be an exception for the reason of independent regulation.
The Ca2+ concentration serves a regulatory function in animal cells, so its concentration
should not be dependent on the size of the Na+ gradient.
Ion pumps may catalyze a net transport of charge. Because this increases charge
separation across the membrane, it is termed electrogenic. Coupled transport may also
produce a net transfer of charge (non-neutral exchanges). These transport systems are
normally driven by the membrane potential, however. Therefore, these transporters tend to
dissipate rather than create a membrane potential. For that reason, they should be termed
electromotive or electrically dissipative rather than electrogenic.

Section 11.2. Animal Cell Plasma Membrane

Animal cells, which find themselves in a medium that is high in sodium, can use
the sodium gradient for transport. Since no other membrane can rely on a high sodium
capacity, it is not surprising that the Na+/K+ ATPase is limited to the plasma membrane of

Page 11.2
animal cells. The Na+/K+ ATPase transports both Na+ and K+ and, in principle, the K+
gradient could also be used for transport. In this regard, it is significant that the membrane
potential in animal cells is usually negative inside. This electrical potential adds to the
chemical potential gradient for Na+ but nearly cancels the potential gradient for K+. Thus
the Na+/K+ ATPase puts the chemical energy into the Na+ gradient rather than the K+
gradient. Since K+ ions can contribute little energy for transport, it is not surprising that K+
is usually not involved in coupled transport across the plasma membrane. For example, in
the squid axon, the membrane potential is -90 mV, the K+ equilibrium potential is -102
mV, and the Na+ equilibrium potential is +50 mV. A K+ ion can contribute only 12 mV of
energy to transport whereas a Na+ ion can contribute +140 mV. This illustrates how the
energy stored by the Na+/K+ ATPase is put mainly into the Na+ ions and not into K+.

3 Na+ Na+

+ Glucose
ADP + Pi

Figure 11.1. Coupled transport across the animal cell plasma membrane.

In animal cell plasma membranes, a glucose transporter uses the Na+ gradient to
bring glucose into the cell. It does this by mediating the cotransport of 1 glucose molecule
and 1 Na+ ion. The energetics of this coupled transport may be analyzed by applying the
same logic used earlier. The total free energy change is the sum of the electrochemical
potential changes of all transported species (equation 10.1):

(11.1) ∆G = Σi ni ∆µi = Σ ni (µiin - µiout)

= Σi ni [RT ln (Cin/Cout) + ziF ψm]

Both glucose (G) and Na+ are transported from outside to inside, so they may both be
assigned a stoichiometry of +1. Then, at equilibrium,

(11.2) ∆G = RT ln ([G]in/[G]out) + RT ln ([Na+]in/[Na+]out) + F ψm = 0

Consequently, the glucose concentration gradient will depend on the sodium gradient and
on the membrane potential as

(11.3) [G]in/[G]out = {[Na+]out/[Na+]in } exp (- F ψm/RT)

Page 11.3
Section 11.3. Inner Mitochondrial Membrane
The electron transfer chain in the mitochondrial inner membrane uses energy
derived from substrate oxidation to pump H+ out across the membrane. This gradient, of
course, is used to drive the synthesis of ATP. It is also used to transport many metabolites
across the inner mitochondrial membrane. ATP and ADP cross the inner mitochondrial
membrane via a transporter catalyzing ATP/ADP exchange. ATP carries one more
negative charge than ADP, so the exchange is electrically dissipative. The membrane
potential (inside negative) generated by H+ transport, will drive ATP/ADP exchange in the
direction of ATP efflux and ADP influx.

(11.4) ∆ψ = RT/F ln { [ATP]out[ADP]in/[ATP]in[ADP]out }

PO4- enters mitochondria via PO4-/OH- exchange. This is an electroneutral exchange but it
will respond to the H+ concentration gradient by virtue of the inverse relationship between
H+ and OH- concentrations. Consequently, PO4- will accumulate in mitochondria because
of the higher internal OH- concentration.

(11.5) [PO4-]in/[PO4-]out = [OH-]in/[OH-]out = [H+]out/[H+]in

The electrochemical potential of one proton (∆µH+) is

(11.6) ∆µ H+ = ∆ψ + RT/F ln {[H+]in/[H+]out}

From equations 11.4 - 11.6, it should be apparent that the electrochemical potential of one
proton drives the uptake of ADP and phosphate into the mitochrondrial matrix and the
export of ATP from the matrix. The electrical part of ∆µ H+ drives ADP uptake and ATP
export, while PO4- uptake is driven by the chemical part of ∆µH+.


H ++ 1/2 O2
H+ ATP 4-
+ H 2O


Figure 11.2. Coupled transport across the inner mitochondrial membrane.

Page 11.4
Section 11.4. Secretory Vesicle Membranes
Secretory vesicle membranes have a vacuolar ATPase that pumps H+ into the
intravesicular space. Consequently, we would expect transport across secretory vesicle
membranes to be coupled to the H+ gradient. Catecholamines (epinephrine, norepinephrine
and dopamine) are indeed transported across the secretory vesicle membrane via a
catecholamine/proton exchange. The stoichiometry of this exchange is 2 protons per
protonated amine so the equilibrium catecholamine gradient is

(11.7) [RNH3+]in/[RNH3+]out = {[H+]in/H+]out}2 exp (F∆ψ/RT)

The equilibrium gradient depends on the square of the H+ gradient because two protons are
exchanged for each amine. Because the charge on the amine cancels the charge on one of
the two protons, there is a net movement of only one charge in the exchange. For this
reason, the catecholamine gradient depends on the membrane potential only to the first

RNH 3+


2 H+
+ Pi

Figure 11.3. Coupled transport across secretory vesicle membranes.

Section 11.5. Some Coupled Transporters

Transporter Substrates Inhibitors

Sodium/glucose transporter Hexose/Na+ cotransport
Vesicular monoamine transporter monoamine:H+ exchange Reserpine
(Major facilitator superfamily)
Serotonin transporter Serotonin/Na+ cotransport Fluoxetine
Anion exchange (Band 3) protein Cl-/HCO3- exchange Stilbene
Sodium/proton antiporter Na+/H+ exchange Amiloride
Ca2+/Na+ antiporter 3 Na+/Ca2+ exchange
Mitochondrial nucleotide transporter ATP4-/ADP3- exchange Atractyloside
Cation/chloride symporter Na+/K+/2Cl- cotransport Furosemide

For a much longer listing, go to

Page 11.5

Bell, G.I., C.F. Burant, J. Takeda, and G.W. Gould (1993) Structure and function of
mammalian facilitative sugar transporters, J. Biol. Chem. 268, 19161-19164.
Crane, R.K., Miller, D., and Bihler, I. (1961) in Symposium on Membrane Transport and
Metabolism (A. Kleinzeller and A. Kotyk, ed.), Academic Press, London, p. 439.
Crane, R.K. (1965) Na+-dependent transport in the intestine and other animal tissues, Fed.
Proc. 24, 1000-1006.
Mitchell, P. (1961) Coupling of phosphorylation to electron and hydrogen transfer by a
chemi-osmotic type of mechanism, Nature 191, 144-148.
Schuldiner, S. (1994) A molecular glimpse of vesicular monoamine transporters, J.
Neurochem. 62, 2067-2078.
West, I.C. (1980) Energy coupling in secondary active transport, Biochim. Biophys. Acta
604, 91-126.

Page 11.6
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000

Section 12.1. Thermodynamic Efficiency

Biological systems, like other machinery, often use the energy released by one
process to drive another. Maximizing the efficiency of this energy coupling is obviously of
great importance. The thermodynamic definition of efficiency η is

(12.1) η = [output power]/[input power] = - [JoutputXoutput]/]JinputXinput]

where J is the flow and X is the force driving that flow. Conservation of energy requires
that output power not exceed input power:

(12.2) JinXin ≥ - JoutXout

Consequently, 1 ≥ η ≥ 0.
The flows are related by the stoichiometry of coupling n:

(12.3) n = Jout/Jin

Combining equations 12.1 and 12.3 gives

(12.4) η = - n[Xout/Xin]

Maximum efficiency is attained when η = 1, but this is the equilibrium condition (JinXin = -
JoutXout). Therefore, no net flow will be produced. Consequently, productive energy
coupling with maximum efficiency requires that η be close to but not equal to one, and that
the coupled flows be close to but not at equilibrium.

Section 12.2. Characteristics of Biological Redox Reactions

Among the most important phenomena mediated by biological membranes are
transport processes that are linked to redox reactions. These participate in numerous
cellular phenomena including the vital bioenergetic functions of mitochondria and
chloroplasts. To understand these processes, it is necessary to consider both the equilibria
and the kinetics of redox reactions.
Redox reactions involve the transfer of n reducing equivalents (either e- or H) from
a donor to an acceptor:

Donorred + Acceptorox → Donorox + Acceptorred

As discussed in Chapter 9 (equation 9.8), the free energy change for this reaction is

(12.5) ∆Gredox = nF Edonor - nF Eacceptor

= nF (E°donor -E°acceptor) + RT ln ( [Donor]red [Acceptor ]ox

Page 12.1
The free energy change will be less than zero when Edonor < Eacceptor. Therefore, reducing
equivalents will be transferred from carriers with lower reduction potentials to carriers with
higher reduction potentials and this will generally be in the direction of increasing E°.
At equilibrium, ∆Gredox = 0 and the ratio of products to reactants is equal to the
equilibrium constant K12. Therefore,

(12.6) K12 = exp{(E°acceptor - E°donor)(nF/RT)}

The mechanisms of biological redox reactions have significant implications,

because biology uses both hydrogen atom carriers and electron carriers and exploits the
difference between them. Although it is common to speak of electron transfer chains and
electron transfer reactions in biology, typical “outer-sphere” electron transfer reactions are
rare. Outer-sphere electron transfer reactions occur spontaneously at rates determined only
by the relative reduction potentials of the donor and acceptor and by the intrinsic reactivity
of each reactant. The Marcus theory for electron transfer reactions in solution defines the
rate constant for electron transfer from compound 1 to compound 2 (k12) in terms of self-
exchange rate constants for each compound (k11 and k22) and the equilibrium constant for
the reaction:

(12.7) k12 = {k11k22K12f12}1/2

f12 is a collision factor which may be defined as

(12.8) log (f12) = (log K12)2/(4 log {k11k22/Z2})

where Z = 1011 M-1s-1. In practice, f12 is usually about equal to 1. A significant implication
of the Marcus theory is that electron transfer is an intrinsically non-specific process and
will occur most rapidly between the lowest-potential electron donor and the highest-
potential electron acceptor. Electron transfer, therefore, is not compatible with the
sequential redox reactions of a redox chain. In fact, electron carriers tend to be avoided in
biological systems because their reactions are uncontrolled.
Instead, biological redox reactions employ organic compounds that donate or accept
hydrogen atoms. These compounds do not react spontaneously so their redox reactions can
be controlled by enzymes. The enzymes that react with these organic hydrogen-atom
carriers, however, are generally metalloproteins. The metal ligand will accept and donate
only electrons, not hydrogen atoms. The active site of the metalloenzyme, therefore, must
provide its substrate (the hydrogen-atom carrier) with a way to exchange a proton as well
as an electron. We will call this phenomenon concerted proton/electron transfer.
Concerted proton/electron transfer implies that biological redox reactions will
naturally release or consume H+. The generation of proton gradients, therefore, is a natural
consequence of the vectorial organization of these reactions. Because electron transfer
reactions are avoided in biology, terminology such as “electron transfer chain” is
misleading. Instead, we will refer to the respiratory chain in mitochondria or to redox

Page 12.2
Section 12.3. The Secretory-Vesicle Ascorbic Acid-Regenerating Chain
A good, simple example of a biological redox chain is the ascorbic acid-
regenerating system in secretory vesicles. The function of this system is to provide
reducing equivalents for redox reactions occurring inside the secretory vesicles. Two
reactions, those catalyzed by dopamine β-monooxygenase (DβM) and peptidylglycine α-
amidating monooxygenase (PAM), are especially significant. DβM converts dopamine to
norepinephrine and mediates catecholamine biosynthesis in adrenal chromaffin cells, in
peripheral sympathetic nerve endings and in noradrenergic neurons in the central nervous
system. PAM and an associated lyase lead to amidation of the carboxyl termini of many
peptide hormones including vasopressin, oxytocin, VIP, neuropeptide Y, α-MSH,
substance P, calcitonin and gastrin.
The monooxygenases both incorporate one atom from O2 into the substrate and
reduce the second oxygen atom to H2O. The reducing equivalents are provided by
intravesicular ascorbic acid (vitamin C) which in turn is oxidized to the radical anion,
semidehydroascorbate. The intravesicular ascorbate is recycled by importing reducing
equivalents across the vesicle membrane through cytochrome b561 (Figure 12.1).
Cytochrome b561, which spans the secretory-vesicle membrane, is reduced in turn by
cytosolic ascorbic acid. This transport of electrons into the vesicles is driven by both the
membrane potential (interior positive) and the pH gradient (interior acidic) generated by a
V-type H+-translocating ATPase in the vesicle membrane. The pH-gradient favors inward
electron flow because the midpoint potential of ascorbate is pH-dependent being higher at
the lower internal pH. Thus, contrasting with the mitochondrial respiratory chain in which
H+ is transported using energy from redox reactions, the secretory-vesicle system drives the
redox reaction using energy from the proton gradient.

Figure 12.1. Mechanism of ascorbic acid regeneration in secretory vesicles. 1)

semidehydroascorbate reductase; 2) cytochrome b561, 3) dopamine β-monooxygenase, 4)
H+-translocating ATPase.

Page 12.3
O -H+ O -H+ O
H H +H+ H

AH2 AH- A=
E°1 +e- -e- -H +e- -e- E°2

O O -H+ O O
. .
H +H+ H
AH . A-.

+e- -e- E°3


-H2O O
Figure 12.2. Interconversion of ascorbate species. Protonation reactions are shown
horizontally and electron-transfer reactions are shown vertically.

Page 12.4
To understand how this system works, it is important to understand that ascorbic
acid functions as a donor of single hydrogen atoms. At physiological pH, ascorbic acid
exists predominantly as a monoanion (AH-). Similarly, the oxidized form occurs as a
radical anion (A-) because this form is stabilized by its capacity to distribute the unpaired
electron over a number of atoms. The fully oxidized form, dehydroascorbate, is not
normally formed because that requires formation of a very unfavorable reaction
intermediate (A). Consequently, cytochrome b561 must react with ascorbic acid and its
radical anion by exchanging the equivalent of a single hydrogen atom. Because the metal
ligand (Fe) will only accept an electron, there must be a mechanism to facilitate ascorbate
deprotonation. This implies that cytochrome b561 reacts with
ascorbate/semidehydroascorbate by concerted proton/electron transfer. Because the
reaction involves only one reducing equivalent and causes no other chemical changes, it is
an especially simple example of this kind of redox reaction.
The distinction between concerted H+/e- transfer and e- transfer is illustrated by
comparing cytochrome b561 with cytochrome c. The rate at which ascorbate reduces
cytochrome c is relatively slow at neutral pH, strongly pH-dependent, and does not saturate
at high ascorbate concentrations. By contrast, the rate of cytochrome b561 reduction is
faster, only weakly pH dependent, and saturates at ascorbate concentrations over about 1
mM. This implies that cytochrome c is reduced by the ascorbate dianion (A=) by outer
sphere electron transfer whereas cytochrome b561 is reduced by the ascorbate monoanion by
concerted H+/e- transfer.
A probable but unproven implication of concerted H+/e- transfer is that the substrate
binding site is sufficient to make the substrate reactive. By creating a mechanism for
proton transfer, the binding site:substrate complex becomes capable of engaging in electron
transfer reactions. This means that metalloproteins can be thought of as two (or more)
separate electron transfer centers: the metal and the bound substrate(s). The bound
substrate will normally exchange electrons with the metal because that pathway is
available, but the bound substrate may also react with other redox centers that may happen
to be in the vicinity. Consequently electrons may occasionally go astray, for example,
reducing O2 to superoxide (O2-).
Cytochrome b561 reacts with ascorbate on both sides of the membrane and therefore
must possess two ascorbate binding sites. The site on the cytoplasmic side probably
contains a histidine residue, because histidine modification inhibits reduction of
cytochrome b561 by external ascorbate. Cytochrome b561 has recently been cloned and
sequenced and has a molecular weight of 30,061. Hydropathy analysis suggests that the
protein has six transmembrane domains with both amino and carboxyl termini being on the
cytoplasmic side of the membrane. The protein contains seven histidine residues, two of
which act as ligands to the heme. The other five are candidates for the ascorbate binding
sites which mediate concerted H+/e- transfer.

Section 12.4. The Mitochondrial Respiratory Chain

In mitochondria, a pair of reducing equivalents are transferred from NADH or
succinate through the respiratory chain to molecular oxygen. Peter Mitchell championed
the hypothesis that the flow of reducing equivalents down the respiratory chain resulted in
the transfer of protons out across the inner mitochondrial membrane and that the proton
gradient generated thereby provided the energy for ATP synthesis. To analyze the
respiratory chain and the manner in which it transports H+, it is convenient to divide the

Page 12.5
chain into three segments: the NADH oxidase/CoQ reductase, CoQ oxidase/cytochrome c
reductase, and cytochrome c oxidase. Let us consider each of these components in turn.

TABLE 12.1. Electron Carriers in the Mitochondrial Electron Transfer Chain

Redox Pair n E° (mV)

NADPH/NADP+ 2 -324
NAPH/NAD+ 2 -320
FMNH2/FMN 2 -219
FADH2/FAD 2 -219
Fe/S centers 1/center -30 to -300
Succinate/Fumarate 2 -31
Ubiquinone/Ubihydroquinone 2 +9
Ubiquinone/Ubisemihydroquinone 1 -240
Ubisemihydroquinone/Ubihydroquinone 1 +258
Cytochrome b566 1 -30
Cytochrome b562 1 +30
Cytochrome c1 1 +230
Cytochrome c 1 +260
Cytochrome a/CuA 1 +280 to +350
Cytochrome a3/CuB 1 +260 to +900
2 H2O/O2 4 +816

NADH oxidase/CoQ reductase oxidizes NADH and reduces coenzyme Q

(ubiquinone). Thus it takes the reducing equivalents from a reduction potential of -320 mV
through a series of iron-sulfur centers to a potential of about -30 mV. In the process, each
pair of reducing equivalents causes 4 H+ to be transferred across the inner mitochondrial
membrane. The H+/e- stoichiometry was measured in a clever experiment by Rottenberg
and Gutman (1977). A proton gradient can drive a reverse flow of reducing equivalents
from succinate to NAD+ if the flow of reducing equivalents down the respiratory chain is
blocked. Rottenberg and Gutman inhibited the flow of reducing equivalents through
cytochrome oxidase using cyanide, and used ATP to generate a proton gradient via the F1Fo
ATPase. The free energy derived from ATP hydrolysis is

(12.9) ∆GATP = ∆G°ATP + 1.38 log {[ADP][Pi]/[ATP]}

The free energy required to drive reverse redox flow is

(12.10)∆Gox = ∆G°ox + 1.38 log {[NADH][fumarate]/[NAD+][succinate]}

If the two processes are allowed to come to equilibrium, ∆Gox + n ∆GATP = 0, where n is
the stoichiometry of coupling. By measuring the equilibrium concentrations of the various
reactants, Rottenberg and Gutman could calculate ∆Gox and ∆GATP and then determine n.
They found a stoichiometry of 4/3. Since the F1Fo ATPase transports 3H+/ATP, this
segment of the respiratory chain must transport 4 H+ per pair of reducing equivalents.

Page 12.6
The CoQ oxidase/cytochrome c reductase segment of the respiratory chain
transports H+ across the inner mitochondrial membrane through a process that Peter
Mitchell termed the Q Cycle. This mechanism is diagrammed in Figure 12.3. It depends
upon two reactions occurring on opposite sides of the inner mitochondrial membrane. On
the matrix side, ubiquinone is reduced by two reducing equivalents, one taken from
cytochrome b562 and the other taken from an iron sulfur center:

Q + cyt b562 red +FeS red + 2 H+ → QH2 + cyt b562 ox + FeS ox

The reduced quinone is uncharged and hydrophobic and will diffuse to the other side of the
membrane where it reduces cytochrome c1 and cytochrome b566:

QH2 + cyt c1 ox + cyt b566 ox → Q + cyt c1 red + cyt b566 red + 2 H+

Intermembrane Matrix
Space FeS


2 H+ Cyt b566 Cyt b562 2 H+

QH 2 QH 2

Cyt c 1

Cyt c

Figure 12.3. The Q Cycle

Page 12.7
pK1 pK2
OH O- O-
CH3 -H+ CH3O CH3 -H CH3

CH3O R +H+ CH3O R +H+ CH3O R


QH2 QH- Q=

E°1 +e- -e- +e- -e- E°2

O pKr O
. -H+ .
- R

QH . Q-.

+e- -e- E°3



Figure 12.4. Interconversion of quinone species. Protonation reactions are shown
horizontally and electron-transfer reactions are shown vertically.

Page 12.8
This results in the net transfer of 2 H+ across the membrane. Because one of the two
reducing equivalents is recycled by cytochromes b562 and b566, a pair of reducing
equivalents passing completely through the Q cycle will cause 4 H+ to be pumped out
across the mitochondrial inner membrane.
Note that the quinone is a carrier of two hydrogen atoms and would be expected to
react with the metalloproteins on both sides of the membrane by concerted proton/electron
transfer. The requirement for a specific active site for each reaction is indicated by the fact
that the reactions are blocked by different compounds; myxothiazol inhibits quinone
oxidation on the outside and antimycin inhibits quinone reduction on the inside.

Figure 12.5. Dioxygen reactions of cytochrome oxidase.

Cytochrome oxidase is the final segment in the mitochondrial respiratory chain. It

is linked to cytochrome c1 by cytochrome c, a soluble protein found in the space between
the inner and outer mitochrondrial membranes. Because cytochrome c is soluble, it
introduces an experimentally accessible break in the respiratory chain and isolates the
cytochrome oxidase segment. Cytochrome oxidase is a multicenter protein with two heme
groups and two copper atoms. Heme a and CuA are located on the cytosolic side of the
membrane and act as the primary acceptors of electrons from cytochrome c. Heme a3 and
CuB lie on the matrix side of the membrane and function in the four-equivalent reaction by

Page 12.9
which O2 is reduced to 2 H2O. Cytochrome oxidase catalyzes a series of electron transfer
reactions and only the reduction of O2 involves H+. Consequently, the cytochrome oxidase
redox reactions do not naturally lead to formation of a proton gradient. Instead,
cytochrome oxidase actually functions as a proton pump. Redox energy is captured and
used to transport H+ across the inner mitochondrial membrane. H+ transport is believed to
be linked specifically to the flow of electrons through the heme a/CuA region of
cytochrome oxidase.
Having now considered the entire mitochondrial respiratory chain, it is possible to
assess the theoretical energy yield of oxidative phosphorylation. Assuming that the ATP
synthase has a stoichiometry of 3 H+/ATP and that 1 H+/ATP is consumed for transport of
ATP, ADP, and inorganic phosphate, the following stoichiometries hold:

Segment of electron transfer chain H+/2e- P/2e-

NADH/succinate 4 1.0
Succinate/cytochrome c 4 1.0
Cytochrome c/O2 2-4 0.5-1.0

The passage of two reducing equivalents down the entire mitochondrial respiratory chain
will cause 1 atom of O to be reduced to H2O. The theoretical P/O ratio (molecules of ATP
formed per atom of O reduced) is 2.5 - 3.0. According to long-held dogma, the theoretical
P/O ratio is 3 but this was established before it was recognized that the ratio did not have to
be an integer. The observed P/O ratio is typically between 2 and 3.
The observed P/O ratio is expected to be less than the theoretical value because of
H+ leakage or poor coupling. The degree of coupling is indicated by the phenomenon of
respiratory control. In the absence of ADP, mitochondria consume O2 at a slow but
measurable rate. This so-called state 4 rate is the rate at which reducing equivalents may
be transferred down the respiratory chain in the absence of ATP synthesis. Presumably,
this redox flow generates a large ∆µH+ and the energy needed to pump H+ against this
substantial gradient slows the flow of reducing equivalents. Another way to look at this is
to recognize that, in the steady state, electron flow may pump protons only as fast as the
protons come back across the membrane. When ADP is added, the rate of O2 consumption
increases dramatically. This so-called state 3 rate is faster because ∆µH+ is dissipated by
ATP synthesis. Thus, electron flow may pump protons as fast as H+ are consumed by the
ATP synthase (F1Fo ATPase). The respiratory control ratio is the ratio of the rate of O2
consumption in state 3 to the rate of O2 consumption in state 4. Generally this ratio lies
between 3 and 15. A high ratio indicates tight coupling between electron flow and ATP
synthesis. A low ratio signifies poor coupling. Complete uncoupling (for example, by
adding an uncoupler) will accelerate state 4 respiration to the rate of state 3 respiration and
will cause the respiratory control ratio to drop to 1.

Section 12.5. The Redox Chain of Photosynthetic Bacteria

The redox chain in purple photosynthetic bacteria is interesting for a number of
reasons. First, it combines features of the mitochondrial respiratory chain and of
photosystem II of the photosynthetic redox chain. As such, it provides some insight into
the evolutionary origins of biological redox chains. Second, the reaction center of
Rhodopseudomonas has been crystallized and its three-dimensional structure has been

Page 12.10
solved. It was one of the first membrane structures for which this was accomplished, and
for it Diesenhofer and Michel shared the 1988 Nobel Prize.
The redox chain incorporates a reaction center which takes reducing equivalents
from cytochrome c2 (a soluble protein), invests energy from light, and then reduces
ubiquinone. Reducing equivalents from ubiquinone return to cytochrome c2 by passing
through b and c cytochromes. In the process, protons are pumped across the membrane.
The proton gradient can be used to make ATP as in mitochondria or chloroplasts.
Alternatively, it can be used to drive reverse electron flow from succinate to NADH.
Cytochrome c2 can pass reducing equivalents to the reaction center or to a cytochrome
oxidase which reduces O2. Thus, the redox chain can create a proton gradient by a cyclic
photosynthetic pathway in the light or it can create a proton gradient by oxidizing succinate
or NADH aerobically in the dark.
The photosynthetic reaction center incorporates four molecules of
bacteriochlorophyll, two of pheophytin and two of ubiquinone. The structure has two fold
symmetry. A bacteriochlorophyll dimer near the outer surface of the membrane donates an
electron upon absorbing light. The electron passes crosses the membrane via a
bacteriochlorophyll and pheophytin molecule on one side of the reaction center. The
pigment molecules on the other side apparently do not function in the photosynthetic
process. The electron finally passes to a quinone on the inside of the membrane. This
quinone (QA) is reduced to the semiquinone but apparently does not become fully reduced.
QA passes the electron to QB, accepts a second electron from the reaction center, and passes
it to QB to form the fully reduced dihydroquinone. The dihydroquinone (QH2B) then
dissociates from the reaction center and is replaced by an oxidized quinone from the
quinone pool.
The quinone reduction results in proton uptake and the nature of the proton uptake
mechanism has received some attention. Note that the oxidized quinone Q may be reduced
to the semiquinone by electron transfer, because the radical species have a pK near
neutrality so both Q- and QH. are formed readily. The second reduction step must form
QH2 because the pK for ionization of the species is greater than 12. Consequently, we
might imagine that QH2 will form by reduction of QH. by concerted proton/electron

Page 12.11
Rhodopseudomonas Redox Chain NAD

Q = Ubiquinone
BChl = bacteriochlorophyll Y Succinate

Cyt b

Cyt c 1

Cyt c 2
Cyt a/a 3

1/2 O 2
Mitochondrial Respiratory Chain


FeS Succinate

Cyt b

Cyt c 1

Cyt c

Cyt a/a 3

1/2 O 2

Photosynthetic Redox Chain

PQ = plastoquinone
PC = plastocyanin
Fd = ferredoxin X
Cyt b559

Cyt f


P 680 P700

Photosystem II Photosystem I
Page 12.12
Section 12.6 The Photosynthetic Redox Chain
The thylakoid membranes in chloroplasts accomplish the following so-called light
reactions of photosynthesis:

H2O + NADP+ + hν → 1/2 O2 + NADPH + H+

2 ADP + 2 Pi → 2 ATP

The free energy of NADP+ reduction is about 40 kcal/mole and that of ATP synthesis is
about 10 kcal/mole. The energy obtainable from a single photon of red light is about 40
kcal/mole. The efficiency of energy conversion in photosynthesis is about 40%, but this
means that the energy from four photons of light are needed to drive one pair of reducing
equivalents through the redox chain. Each electron must receive the energy from two
photons and this is accomplished by placing two photosystems in the redox chain.
The light energy is collected by antenna pigments in the chloroplast and transferred
to the reaction centers. To make photosynthesis proceed efficiently, it is important that
light be distributed evenly between the two photosystems so that they run at the same
speed. Plants accomplish this by using the redox state of the quinone pool to regulate the
distribution of light energy between the two photosystems. If the quinone pool is too
reduced, then PS I is operating too slowly and more light needs to be diverted to it.
Conversely, if the quinone pool is too oxidized, then more light must be shunted to PS II.
This is accomplished by phosphorylation control of light harvesting protein (Allen et al.,
1981). If the quinone pool is reduced, a kinase is activated increasing phosphorylation of
the LHCP. This directs more of the light energy to PS I. If the quinone pool becomes too
oxidized, the kinase is inactivated and the LHCP is dephosphorylated by a phosphatase. A
greater fraction of the absorbed quanta are then shunted to PS II.
Photosystem II is quite similar to the reaction center of bacterial photosynthesis in
Rhodopseudomonas. Consequently, the solution of the structure of the bacterial reaction
center has also illuminated the mechanism of photosystem II.


J.F. Allen, J. Bennett, K.E. Steinback and C.J. Arntzen (1981) Chloroplast protein
phosphorylation couples plastoquinone redox state to distribution of excitation
energy between photosystems, Nature 291, 25-29.
P. Brzezinski (1996) Internal electron-transfer reactions in cytochrome c oxidase,
Biochemistry 35, 5611-5615.
S.I. Chan and P.M. Li (1990) Cytochrome c oxidase: Understanding Nature's design of a
proton pump, Biochemistry 29, 1-12.
K. Krab and M. Wikstrom (1987) Principles of coupling between electron transfer and
proton translocation with special reference to proton translocation mechanisms in
cytochrome oxidase, Biochim. Biophys. Acta 895, 25-39.
R.A. Marcus and N. Sutin (1985) Electron transfers in chemistry and biology, Biochim.
Biophys. Acta 811, 265-322.
H. Michel and J. Deisenhofer (1988) Relevance of the photosynthetic reaction center from
purple bacteria to the structure of photosystem II, Biochemistry 27, 1-7.

Page 12.13
P. Mitchell (1961) Coupling of phosphorylation to electron and hydrogen transfer by a
chemi-osmotic type of mechanism, Nature 191, 144-148.
P. Mitchell (1975) Protonmotive redox mechanism of the cytochrome b-c1 complex in the
respiratory chain: Protonmotive ubiquinone cycle, FEBS Lett. 56, 1-6.
P. Mitchell (1976) Possible molecular mechanisms of the protonmotive function of
cytochrome systems, J. Theoret. Biol. 62, 327-367.
P. Mitchell (1979) Keilin's respiratory chain concept and its chemiosmotic consequences,
Science 206, 1148-1159.
D. Njus and P.M. Kelley (1993) The secretory-vesicle ascorbate-regenerating system: A
chain of concerted H+/e- transfer reactions, Biochim. Biophys. Acta 1144, 235-248.
M.Y. Okamura and G. Feher (1992) Proton transfer in reaction centers from photosynthetic
bacteria, Ann. Rev. Biochem. 61, 861-896.
H. Rottenberg and M. Gutman (1977) Control of the rate of reverse electron transport in
submitochondrial particles by the free energy, Biochemistry 16, 3220-3227.
B.L. Trumpower and R.B. Gennis (1994) Energy transduction by cytochrome complexes in
mitochondrial and bacterial respiration: The enzymology of coupling electron
transfer reactions to transmembrane proton translocation, Ann. Rev. Biochem. 63,
I.C. West, P. Mitchell and P.R. Rich (1988) Electron conduction between b cytochromes of
the mitochondrial respiratory chain in the presence of antimycin plus myxothiazol,
Biochim. Biophys. Acta 933, 35-41.
J.R. Winkler, B.G. Malmstrom, and H.B. Gray (1995) Rapid electron injection into
multisite metalloproteins: intramolecular electron transfer in cytochrome oxidase,
Biophys. Chem. 54, 199-209.

Page 12.14
Fundamental Principles
of Membrane Biophysics


David Njus

Department of Biological Sciences

Wayne State University

© D. Njus, 2000


Section 13.1. Topology and General Principles of Epithelia

Epithelia, such as the lining of the gastrointestinal tract and the nephrons of the kidney,
are layers of cells across which ions and metabolites must be transported. Transport across
epithelial layers follows a common pattern with different functions achieved using different
combinations of transporters. It is possible, therefore, to consider a general mechanism and to
establish conventions for defining the various parts and parameters of the system.
The epithelium separates two compartments: the serosal side (blood) and the mucosal
side (lumen). The epithelial cells are polarized, so membranes with different characteristics face
each side. The basolateral membrane faces the serosal side and the apical (or brush border)
membrane faces the mucosal side. The Na+/K+ ATPase is found in the basolateral membrane but
not in the apical membrane. Usually, the basolateral membrane is permeable to K+ (it has a
barium-sensitive K+ channel) but not to Na+. The apical membrane, by contrast, is permeable to
Na+ but not to K+. The K+ current creates a membrane potential (negative inside) across the
basolateral membrane. The inward Na+ current dissipates the membrane potential across the
apical membrane, so the potential on the serosal side is generally positive relative to the mucosal
side. This transepithelial potential (ψtep)drives current through the paracellular space.
The transepithelial potential is a measure of the tightness of the epithelium. Tight
epithelia have a greater resistance and therefore can have a greater transepithelial potential.
Examples are the frog skin, urinary bladder, colon and distal nephron. Leaky epithelia have a
lesser resistance and a smaller transepithelial potential. Epithelia of the small intestine and gall
bladder are examples.
Many kinds of transporters are found in epithelial cells, but those involved in transport of
the principal ions (Na+, K+, Cl-, H+, Ca2+) are common and relatively few in number. A listing of
these and their diagnostic inhibitors follows.

Table 13.1. Common Ion Transport Activities in Cells

Transporter Membrane Inhibitors

Na+/K+ ATPase Basolateral Ouabain
K+ channel Either Barium
Na+/Ca2+ exchanger Basolateral
Na+/H+ exchanger Apical Amiloride
Na+/K+/2Cl- cotransporter Either Furosemide, Bumetanide
Cl-/HCO3- exchanger Apical SITS, DIDS
Cl- channel Apical

Section 13.2. Analysis of a Tight Epithelium

The frog skin is the classical example of a tight epithelium. It transports Na+ inward
across the skin. This is achieved by an apical membrane that is permeable only to Na+ and a
basolateral membrane that is permeable only to K+. To analyze the electrical properties of this
system, we define the lumen as the reference potential. The intracellular potential ψapi is then the

Page 13.1

membrane potential across the apical membrane. The potential on the serosal side is the
transepithelial potential (ψtep). The membrane potential across the basolateral membrane is then

(13.1) ψbaso = ψapi - ψtep

Serosal Mucosal
(Blood) (Lumen)
I baso

ATP 2 K +

I api
3 Na +
ADP + Pi Na +

I para

ψ ψ ψ =0
tep api

Basolateral Apical
Membrane Membrane

Figure 13.1. Frog skin

The apical membrane is permeable to Na+ but not K+, so the current across the apical membrane

(13.2) Iapi = (ψapi - ψNa)/RNa

ψNa is the Na+ equilibrium potential across the apical membrane, and RNa is the resistance to the
Na+ current across the apical membrane.
The basolateral membrane is permeable to K+ but not Na+, so the current across the
basolateral membrane is

(13.3) Ibaso = - (ψbaso - ψK)/RK = - (ψapi - ψtep - ψK)/RK

ψK is the basolateral K+ equilibrium potential and RK is the basolateral resistance to the K+

The current through the paracellular space is

Page 13.2

(13.4) Ipara = ψtep/Rpara

At steady state, these currents must all be equal, so

(13.5) Iapi = Ibaso = - Ipara

Upon solving this series of equations, it is apparent that

(13.6) ψapi = ψNa - (ψNa - ψK)RNa/(Rpara+RNa+RK)

(13.7) ψtep = (ψNa - ψK)Rpara/(Rpara+RNa+RK)

This shows that ψapi lies between the apical Na+ equilibrium potential (ψNa) and the basolateral
K+ equilibrium potential (ψK). If RNa = 0, then Na+ will be in equilibrium across the apical
membrane and ψapi = ψNa. At the other extreme, if RK = Rpara = 0, then K+ will be in equilibrium
across the basolateral membrane, and ψbaso = ψK. The transepithelial potential will be zero, so
ψapi = ψbaso = ψK.
The transepithelial potential (ψtep) is a little smaller than the difference between the apical
Na and basolateral K+ equilibrium potentials (ψNa - ψK).

Section 13.3. GI Epithelial Transport

The small intestine functions to transfer nutrients and other important metabolites into the
blood. The intestinal epithelium is leaky, so the transepithelial potential is small. This
maximizes the membrane potential across the apical membrane (the brush border membrane) and
thereby optimizes the energy that Na+ ions can contribute to coupled transport across that
membrane. As an example of transport across the epithelium of the small intestine, hexoses
(such as glucose) are carried across the apical membrane into the epithelial cells by Na+-linked
transporters. Glucose transport across the basolateral membrane occurs by facilitated transport
and is driven only by the sugar's concentration gradient. Note that a Na+-linked transporter
would not function to transport glucose out of the cell into the blood because the Na+ gradient
opposes transport in that direction.

Page 13.3

Serosal Mucosal
(Blood) (Lumen)


ATP 2 K + Hexose

3 Na+ ADP + Pi Na +


+10 mV -40 mV 0 mV

Basolateral Apical
Membrane Membrane

Figure 13.2. Sugar transport in the small intestine

The oxyntic cells of the gastric mucosa are responsible for secreting HCl to acidify gastric
juice in the stomach. Because gastric juice is extremely acidic, H+ and Cl- are moved against
large concentration gradients, and this requires considerable energy. Estimating the H+ and Cl-
concentrations in gastric juice and blood as shown below, the transepithelial ∆µ for H+ and Cl-
may be calculated:

Serosal Concentration Mucosal Concentration ∆µ

H+ 3.4 x 10-8 M (pH 7.4) 150 mM 8.4 kcal/equiv
Cl- 50 mM 150 mM 1.1 kcal/equiv

The energy for H+ and Cl- transport is satisfied by a K+/H+ ATPase located in the apical
membrane. This enzyme actively pumps H+ against the large H+ concentration gradient across
the apical membrane. The concomitant K+ influx neutralizes charge movement. The K+ gradient
is then used to drive Cl- into the lumen via an electroneutral K+/Cl- cotransport. Cl- influx across
the basolateral membrane is driven by the Na+ gradient generated by the Na+/K+ ATPase. Na+
influx is coupled to Cl- influx via parallel transporters: a Cl-/HCO3- exchanger and a Na+/H+

Page 13.4

Serosal Mucosal
(Blood) (Lumen)

K+ +
ATP 2 K +
ADP + Pi H

3 Na+ ADP + Pi Cl -
Na +
Cl -

H2 O +
Cl - K+

CO 2
HCO3 -

Basolateral Apical
Membrane Membrane

Figure 13.3. Gastric mucosa

Section 13.4. Renal Epithelial Transport

In the kidney, the proximal tubule is responsible for reabsorption of nutrients and other
important metabolites. Consequently, like the small intestine, the renal proximal tubule is a
leaky epithelium. This maximizes the energy stored in the Na+ gradient across the apical
membrane and optimizes the efficiency of Na+-linked uptake of amino acids, sugars, phosphate
and sulfate. The proximal tubule also functions in the reabsorption of Na+ and Cl-. This is
accomplished by a Na+/H+ exchanger and an anion exchanger in the apical membrane.
Acidification of the lumen by Na+/H+ exchange facilitates reabsorption of HCO3- as CO2. Na+
and Cl- uptake causes H2O to be reabsorbed by osmosis. Amiloride, which blocks the Na+/H+
exchanger, inhibits reabsorption of Na+ and HCO3-. It also blocks the concomitant uptake of
water and consequently functions as a diuretic.

Page 13.5

Serosal Mucosal
(Blood) (Lumen)

K+ Cl -

ATP 2 K +
HCO 3 -
H 2O
3 Na+ ADP + Pi
CO 2

Basolateral Apical
Membrane Membrane

Figure 13.4. Renal proximal tubule

Serosal Mucosal
(Blood) (Lumen)

Cl - K+

ATP 2 K +
Na +
2 Cl -
3 Na+ ADP + Pi

Basolateral Apical
Membrane Membrane

Figure 13.5. Thick ascending limb - Loop of Henle

In the thick ascending limb of the loop of Henle, NaCl reabsorption must be achieved
without loss of water by osmosis. Na+ and Cl- are taken up from the lumen by the Na+/K+/2Cl-

Page 13.6

cotransporter which is driven by the Na+ and Cl- gradients. The Na+ concentration in the
epithelial cells is kept low by the Na+/K+ ATPase. The Cl- concentration is kept low by a Cl-
channel in the basolateral membrane. Cl- efflux through this channel is driven by the membrane
potential. A K+ channel is located in the apical membrane rather than in the basolateral
membrane. This reduces the K+ gradient across the apical membrane and prevents it from
interfering with the proper functioning of the cotransporter. The K+ channel can also cause the
transepithelial potential to be positive on the lumen side. This arrangement results in the active
transport of NaCl across the epithelium and can make the urine hypotonic relative to the blood.
The loop diuretic furosemide (Lasix) inhibits NaCl reabsorption by blocking the Na+/K+/2Cl-
cotransporter. Because it blocks K+ reuptake from the lumen, furosemide leads to K+ loss.
Amiloride, which blocks water reabsorption by acting on Na+/H+ exchange, does not do this and
is known as a K+-sparing diuretic.

Section 13.5. Airway Epithelia

In airways, mucus production in the lumen depends on H2O secretion by the epithelium.
H2O secretion occurs by osmosis consequent to Cl- secretion. Cl- secretion occurs via a
Na+/K+/2Cl- cotransporter in the basolateral membrane and a Cl- channel in the apical membrane.

Serosal Mucosal
(Blood) (Lumen)


ATP 2 K +

3 Na + ADP + Pi Cl -

Na +
2 Cl -

Basolateral Apical
Membrane Membrane

Figure 13.6. Airway Epithelium

The Na+/K+ ATPase drives Cl- influx across the basolateral membrane, and Cl- flux across the
apical membrane then occurs passively. The symptoms of cystic fibrosis are caused by a defect

Page 13.7

in regulation of the Cl- channel resulting in inadequate H2O secretion making secretions
characteristically salty and too viscous.
Airway epithelia and the epithelia in the nephrons of the kidney are structured in similar
ways, yet airway epithelia function to secrete Cl- while the epithelia in the kidney take up Cl-.
This illustrates how the general structure of an epithelium can be adapted to achieve different
functions simply by incorporating appropriate combinations of transporters in the apical or
basolateral membranes.

W.B. Reeves and T.E. Andreoli (1992) Renal epithelial chloride channels, Ann. Rev. Physiol. 54,
J.R. Riordan (1993) The cystic fibrosis transmembrane conductance regulator, Ann. Rev. Physiol.
55, 609-630.
B. Thorens (1993) Facilitated glucose transporters in epithelial cells, Ann. Rev. Physiol. 55, 591-
E.J. Weinman and S. Shenolikar, (1993) Regulation of the renal brush border membrane Na+-
exchanger, Ann. Rev. Physiol. 55, 289-304.
M.J. Welsh (1987) Electrolyte transport by airway epithelia, Physiol. Rev. 67,1143-1184.
M.J. Welsh, M.P. Anderson, D.P. Rich, H.A. Berger, G.M. Denning, L.S. Ostedgaard, D.N.
Sheppard, S.H. Cheng, R.J. Gregory and A.E. Smith, (1992) Cystic fibrosis transmembrane
conductance regulator: A chloride channel with novel regulation, Neuron 8, 821-829.
C.S. Wingo and B.D. Cain (1993) The renal H-K-ATPase: Physiological significance and role in
potassium homeostasis, Ann. Rev. Physiol. 55,323-347.
E.M. Wright (1993) The intestinal Na+/glucose cotransporter, Ann. Rev. Physiol. 55, 575-589.

Page 13.8