Hemoglobin and myoglobin

Interrelationship
Between Structure and
Function in Hemoglobin
and Myoglobin
ERALDO

ANTONINI

Institute of Biological Chemistry, University of Rome, Rome, Italv4
Introduction
............................................................
-Interaction
Between
Heme and
Basic Properties
of Hemoglobin
and Myoglobin
............................................................
Protein.
Reversible
oxygen binding.
..............................................
Sltability
of heme-protein
complex.
.......................................
...................
Changes in conformation
of protein
due to heme binding.
Structure

and Function
of Myoglobin.
......................................
Structure
of myoglobin.
................................................
Functional
properties
of myoglobin.......................................
.....................................
Structure
and Function
of Hemoglobin.
Over-all
structure
and amino acid sequence of mammalian
hemoglobin.
.........
.............................................
Physicochemical
properties.
Reactions
of hemoglobin
with ligands.....................................
Oxygen
dissociation
curve of hemoglobin
and heme-heme
interaction.
..........

Phenomenological
Aspects of Heme-Heme
Inter action. ........................
.....................................................
Species variations.
Type of ligand. ........................................................
Temperature
..........................................................
...........................................
Concentration
of hemoglobin.
Effect of pH. ..........................................................
Salt concentration
of medium.
...........................................
..................................
Kinetic
effects of heme-heme
interaction.
............................................
Modification
of hemoglobin.
Mechanism
of Heme-Heme
Interaction.
....................................
..
Relation
of shape of oxygen dissociation
curve to number

of hemes per molecule.
Conformational
changes and heme-heme
interaction.
.......................
Bohr Effect. .............................................................

123

‘25
‘25
127
I29
1.30

130
132
137
138
I39
141
141
I44
I44
I44
I45
I45
146
148
148

I49
I53
I54
156
‘59

INTRODUCTION

The physiological
function of respiratory
heme proteins lies in the ability of
the ferrous iron of the prosthetic group to bind molecular
oxygen reversibly.
This
is a highly specialized
function
at a molecular
level; in hemoglobin
especially the
characteristics
of the oxygen equilibrium,
as shown by the isolated protein
in
solution, are such as to meet the various physiological
needs of the whole organism
(3r). Accordingly,
in the different
animal
species the structural
and functional

properties
of hemoglobins
change widely, in each case adapting
to different physiological situations.
In trying to analyze the interrelationship
between structure and function
in
myoglobin
and hemoglobin,
we may start by emphasizing
that the physiologically
I*3


124

ERALDO

ANTONINI

Volume $5

important
reaction occurs at the iron atom of the prosthetic group, this being the
point of attachment
of the oxygen molecule.
We can therefore look at the whole
myoglobin
and hemoglobin
molecules as very complex coordination

compounds
of the iron.
In the various heme proteins the reactivity
of the metal atom is addressed
in a very specific direction
and is precisely controlled
due to interactions
with
increasingly
complex systems which represent successive steps of chemical organization (I I 3, I 22). The development
of the characteristic
reactivity
of hemoglobin
may be traced as follows.
r) Heme. The prosthetic group of hemoglobin
and myoglobin
is a very stable
coordination
compound
of iron with protoporphyrin
(61, I I 3, 123). In heme, the
ferrous atom at the center of the porphyrin
ring has four of its six coordination
sites occupied by bonds with the porphyrin.
The fifth and sixth free coordination
positions are available for combination
with ligands, one on each side of the plane
of the porphyrin.
The isolated ferrous heme already shows some of the characteristic reactivity
of hemoglobin

and myoglobin
in its ability to bind reversibly
a
number of ligands such as CO, R-CN,
etc., in a one-to-one ratio, giving specific
compounds
with well-defined
spectroscopic
and magnetochemical
properties
( IOI, I 02, I I 3, 202).
These ligands are bound to one of the free coordination
sites,
the other one being occupied
by a water molecule
(87, I I 3). The isolated ferrous heme is not able, however,
to bind molecular
oxygen reversibly
since it is
rapidly oxidized to ferric heme in the presence of this gas.
2) Heme and a simple nitrogen ligand
in the J;fth position (hemochromogen). The
addition to heme of a simple nitrogen compound
such as pyridine, which is bound
in the fifth coordination
position of the iron, has a marked quantitative
effect on
the reactivity
at the sixth position which becomes more similar to that of heme
proteins. Thus, for example,

the combination
velocity constant for the reaction
with CO passes from I .5 X IC? M-~ set-l for heme to 5 X 10~ M-~ sec.-r for pyridine hemochromogen
(I 80) ; the oxidation-reduction
potential
becomes much
more positive (51). The interactions
between the ligands at the various coordination sites of the iron are now being explored
theoretically
by the “ligand
field
theory”
(75, 128, 203).
3) Myoglobin.
This protein contains one heme group and one polypeptide
chain. The interaction
of the heme with the specific protein occurs at the fifth
coordination
position of the iron and through
additional
bonds of the porphyrin
with the polypeptide
chain.
The complex becomes able to bind molecular
oxygen reversibly
in addition
to the other ligands mentioned
above. The functional
behavior
of myoglobin

largely corresponds
to the simplest case for a reversible
reaction and appears to
be essentially controlled
by the “local environment”
of the heme group in the
protein.
4) Hemoglobin. Every hemoglobin
molecule
consists of several polypeptide
chains, each containing
a heme group. Here, in contrast to myoglobin,
the reactivity of the heme iron depends not only on its “local environment”
in the
polypeptide
chain, but is also dominated
by distant, indirect
interactions
with
other groups in the molecule which may belong to a different polypeptide
chain.


January

HEMOGLOBIN

1965

AND

MYOGLOBIN

I*5

These interactions
determine
the characteristic
features of the functional
behavior
of hemoglobins
which
are physiologically
important:
the value of the oxygen
affinity, the shape of the dissociation
curve, the Bohr effect.

BASIC

PROPERTIES

INTERACTION

OF
BETWEEN

HEMOGLOBIN
HEME

AND
AND

MYOGLOBIN-

PROTEIN

It was mentioned
before that ferrous heme and hemochromogens
may be
considered
primitive
models for the function of respiratory
heme proteins in view
of their ability to combine reversibly with a number of ligands such as CO, R-CN,
etc., giving addition
compounds
in which the basic linkage between the iron and
the molecule
coordinated
at the sixth position
is essentially
the same as in the
corresponding
ones with hemoglobin
and myoglobin.
However,
it is only on
combination
with the protein of hemoglobin

and myoglobin
that the ferrous iron
in the heme group acquires the specific and physiologically
important
property
of binding
reversibly
molecular
oxygen, giving a stable oxygenated
derivative.
This property
of hemoglobin
and myoglobin
is lost on denaturation
of the protein, the heme-protein
complex then behaving like a simple hemochromogen
in
its reaction with ligands (IO). The oxygen capacity is therefore closely dependent
on the specific interaction
of the heme with the native protein.
The simplest form of a hemoprotein
capable
of reversible
oxygenation
is
represented
by myoglobin,
a protein which contains only one polypeptide
chain
and one heme group per molecule. Vertebrate

and invertebrate
hemoglobins
are
usually made up of several subunits, each consisting of one polypeptide
chain with
an associated heme. Mammalian
hemoglobins,
which have a molecular
weight
near 65,000, are made up of four chains, identical
in pairs (QI and f3 chains) (54,
133, 138). It is evident, however,
that the hemoglobin
subunits and myoglobin
have a similar over-all structure,
and that, more specifically,
the basic linkages
between the heme and the protein are essentially
the same (54, 108, 132). The
subject can be therefore discussed from a uniform point of view for both hemoglobin
and myoglobin.
The specific combination
of the heme with the native protein, in rnyoglobin
and hemoglobin,
is responsible
for the following
major effects: a) the ferrous iron
in the heme binds oxygen reversibly, without
being oxidized;
b) the heme protein

complex
has a great stability,
its association
constant being extremely
high at
neutral pH; G) the protein undergoes, on binding
the heme group, a drastic change
in conformation,
which is especially evident in hemoglobin.
It is quite clear that
though all these effects may be correlated,
they are to some extent independent
and structurally
due to different bonds between the heme group and the protein.
Reversible 0 xygen Binding
The basis of the oxygenation
process in hemoglobin
and myoglobin,
as contrasted with the oxidation
which takes place when ferrous heme or simple heme
compounds
are exposed to oxygen, is not completely
clear although much prog-


126

ERALDO

ANTONINI

Volume 45

ress in this direction
has been made. The subject has recently
been critically
discussed by Wang (201).
It must be pointed out that the difference between heme
proteins and simple heme compounds,
as far as the oxidation
of the iron in the
presence of oxygen is concerned,
is dramatic
but only quantitative.
Qualitatively,
hemoglobin
and especially myoglobin
have a tendency to autoxidation
in the
presence of oxygen (45a, I I 3).
For the reversible combination
with oxygen, the presence of the lig’and to the
iron in the fifth coordination
position and the environment
of the heme group are
important.
The bond of the iron atom with the polypeptide
chain has now been unequivocally
identified
by X-ray studies (54, 107, 108) : it is represented

in hemoglobin and myoglobin
by a linkage with the imidazole
nitrogen
of a histidine
residue,, This is residue number 87 in the hemoglobin
a chain, number 92 in the
p chain and number
91 of the myoglobin
chain (54). The distance between the
iron atom and the histidine nitrogen
is 1.9 A. This bond, which is normal to the
plane of the heme group, satisfies one of the two available coordination
sites of the
iron. The other coordination
position of the iron on the other side of the heme
plane is occupied by a water molecule
(83, 84) in the deoxygenated
derivative or
by the ligand. Th.e ligand may be oxygen, CO, NO, R-CN,
or a nitroso aromatic
compound
(I I 3). It is clear that there must be strong interactions
between the
two groups which are coordinated
to the iron. Lumry (I 14) has discussed in detail
the mechanism
by which the protein can control the reactivity
of the metal acting
through this iron-imidazole
bond.

The iron-protein
bond does not appear, by itself, enough to give the metal
the capacity of binding
oxygen; other interactions
with the protein must play a
major role.
Whatever
th.e mechanism,
in hemoglobin
and myoglobin,
by which binding
of oxygen is not accompanied
by a rapid oxidation
of the iron, it is likely that it
is an indirect
one and does not involve the primary
ligand-iron
bonds. This is
suggested by the fact that in the case of carbon monoxide,
the derivatives
with
hemoglobin
or myoglobin
have a spectral and magnetochemical
character identical
to those of the corresponding
simple heme compounds
and similar free energy of
formation
( I 80, 20 I).

Also there is no correlation
between
the oxidation-reduction
potentials
of
different heme compounds
and the velocity of autoxidation.
Thus the Em [potential when (ox.) = (red.)] for ferro-ferri
pyridine hemochromogen
is about the
same as for ferro-ferrihemoglobin
(51, I rg), while some modified
hemoglobins
show a large drop in Em which is not accompanied
by a significant
increase in
the speed of autoxidation
(I 6, 46).
Great importance
has been ascribed for the reversible oxygen binding
to the
so-called “distal histidine,”
that is to a histidine
residue (58cu, 63p, 66,) which is
seen in the X-ray models to lie in front of the heme, in the opposite direction
to
the iron binding
imidazole
(54). It has been suggested that the iron-oxygen
complex is stabilized

by interactions
(hydrogen
bond) of the oxygen mollecule with
this residue. In agreement
with this view is the fact that in some abnormal
hemoglobins, such as the so-called “hemoglobin
M,” which are characterized
by a


much greater speed of autoxidation,
this histidine is replaced by other amino acid
residues (64, 92). However,
this imidazole
may not be generally essential in the
oxygenation
process because some invertebrate
myoglobins
contain only one histidine per heme group and this appears to be the one directly
linked to the iron
(I57)*
The environment
of the heme group in the protein molecule is certainly very
important
for oxygenation.
It must be noted that although
the heme is near the
surface of the molecule,
it is largely surrounded
by the polypeptide

chain which
folds around it (54, 103, 106). It is a con-n-non observation
that whenever
hemoglobin is treated with reagents that presumably
tend to unfold the protein, e.g.
concentrated
urea solutions,
the autoxidation
is much faster, apart from any
irreversible
denaturation
of the protein.
The X-ray studies (54, 104, I 05) and the chemical studies on the amino acid
sequence (42, 58) have shown that in hemoglobin
and myoglobin,
the region of the
polvpeptide
chains,
around
the
heme
group,
is
particularly
rich
in aromatic
side
d
chains. These provide a hydrophobic
environment

to the heme which might prevent electron transfer from the oxygen to the ferrous iron. As a matter of fact,
Wang and co-workers
(20 I, 202)
have been able to make a synthetic model which
undergoes
reversible
oxygenation
by embedding
heme in a polystyrene
matrix
containing
imidazole.
On oxygenation,
this compound,
which is studied in the
form of a thin, solid film, also shows spectral changes similar to those occurring
in
hemoglobin
and myoglobin.
Stability

ofc Heme-Protein

Comjlex

The high stability of the heme-protein
complex also depends on the specific
interaction
between the heme and the native protein.
On denaturation

of hemoglobin and myoglobin,
not only does the heme lose the capacity of binding
reversibly molecular
oxygen, but also the linkage of the protein with the heme becomes looser and unspecific
(I o, I I 3).
Hemoglobin
and myoglobin
can be reversibly split into heme and the protein
part, globin, by such procedures
as the treatment
of the proteins with acid acetone
(9, 89). The reconstituted
hemoglobin
prepared
in the past by these globins often
showed differences
in some important
properties
from natural
hemoglobins
(73,
I I 3, 204). Recently, however,
by improvement
of the methods, globin preparations have been obtained from hemoglobin
(I 55, I 56) and myoglobin
(I 97) which,
on coupling
with heme, give reconstitution
products
indistinguishable

from the
original
pigments
in all the major physicochemical
and functional
properties
studies on the kinetics and
(II, 147, 151, I 58, I 97). These findings stimulated
equilibria
of the interaction
between heme and globin.
Rossi Fanelli and Antonini
(154) have shown that the heme in hemoglobin
and myoglobin
is not a fixed or immobile
prosthetic
group, but that there is a
definite dissociation
of the pigments even at neutral pH. The dissociation
of the
heme group from the native protein can be followed by “heme transfer”
reactions
which occur when a given hemoprotein
is mixed with another hemoprotein:
hemoprotein

x + globin

y -+ hemoprotein

y + globin

x


ERALDO

128

ANTONINI

Volume 45

Thus, the interaction
between heme and globin may be treated as a labile reversible
equilibrium.’
The stability constant of the heme-protein
complex at neutral pH is very high,
largely at acid and alkaline pH (29, 70, 194).
of the order of 1013 M- I, decreasing
Myoglobin
globin has a greater affinity for heme than hemoglobin
globin (30,
I 54). In the case of hemoglobin
globin, the equilibrium
between heme and globin
is dominated
by strong stabilizing
interactions
between

the heme binding
sites
(1% ‘3, 30)The kinetics of the combination
of heme with globin has been investigated
in detail by Gibson and Antonini
(69, 70). The rate of reaction
of monomeric
heme derivatives,
such as CO-heme,
with globin is extremely
high, the apparent
second-order
rate constant being about 108 M-~ set-l. The reaction, which is very
similar
for hemoglobin
and myoglobin
globin,
is complex
and spectrophotometrically
appears to consist of two kinetically
distinct steps: the rapid, almost
diffusion-controlled
formation
of a labile complex
and the breakdown
of the
complex to form the heme protein according
to the scheme:
heme

derivative

+ globin

t

kl

complex

k3 ) hemoglobin

derivative

( 2>

k2

In comparison
with denatured
globin, the greater affinity of native globin for
heme must be mainly due to the low rate of dissociation;
the over-all rate for the
combination
of heme derivatives
with denatured
globin is similar or higher than
that for the combination
of heme with native globin (69). The loose association
of heme with simple ligands such as imidazole

or pyridine is reflected, on the other
hand, in a low rate of combination
(69).
An analysis of the structural
basis of the stability of the heme-protein
complex
in native myoglobin
and hemoglobin
may now be attempted.
First of all it must
be pointed out that the stabilization
is obtained
mainly through
linkages of the
protein with the porphyrin
part of the heme group rather than through the ironprotein bond. This is shown by the fact that protoporphyrin
combines with globin
with a specificity and affinity similar to that of protoheme
(158).
The kinetics of
combination
is also very similar for heme and porphyrin
(68). In agreement with
this, thie X-ray data show extensive contacts of several segments of the protein
chain with the porphyrin
part of the heme (105,
I 32). Some of the linkages
of the
porphyrin
with the protein must involve the vinyl side chains in positions 2 and

4 of the porphyrin
ring. These side chains, which are deeply buried in the interior
of the molecule,
probably
make hydrophobic
bonds with aromatic
side chains.
This is suggested by the fact that the affinity and rates of reaction with globin of
hemes modified
in positions 2 and 4 of the porphyrin
side chains are much lower
than those for protoheme
(70, I 54). These changes in affinity and rates of reaction
parallel
the decrease in the hydrophobic
character
of the side chain; thus the
affinity and rates of reaction with globin decrease in the order proto-(a-q vinyl),
1 F’rom a historical
point of view it is interesting
to recall that the dissociation
into herne and globin was a central feature in Bohr’s theory
(2. P!zysioZ. I 7: 685,
in Hill’s theory
(Bioche~n.
J. I 5: 579, 192 I) on the oxygen
equilibrium.

of hemoglobin
and

later

~gtq)


January

1965

TABLE

I.

HEMOGLOBIN

AND

129

Rates of reaction of CO-heme with globin”
IO‘-’

CO-Heme

Proto
Meso
Deutero
Hemato
Dimethyl

MYOGLOBIN

x

Kl

(34 -1 seP)

5ot
5t
5
I
deuteroheme

disulfonate

IO-= X kz
(set”)

10’~ x kg
(set-‘)

I8t
I .6t
I .6

3.7

0.5

2

1.3
094

0

* Values
for the velocity
constants
describing
the reaction
of CO-heme
with globin
according
to the scheme of eq. 2.
t These
values
are minimum
only;
larger
values
in
the same ratio are also possible.
Refs. 69, 70.

meso-(2-4
ethyl),
deutero-(2-4
hydrogen),

hemato-heme-(2-4
hydroxy-ethyl)
(Table I ).
A pre-eminent
role in the stabilization
of the heme-protein
complex has also
been ascribed to the propionic
acid groups of the porphyrin
(84, I I 3, 126, I 27).
In the X-ray models these groups appear to be directed toward the outside of the
molecule
and in myoglobin
interact with arginine,
histidine,
and serine (105).
Reconstitution
products from native globins and heme-lacking
carboxyl
groups
have been obtained
(85, I 26, I 27) ; these compounds
do have the over-all properties of hemoglobin
or myoglobin
but are very easily denatured,
suggesting a very
low affinity of these modified hemes for globin.
Changes in Conformation

of Protein Due to Heme Binding

Globin
undergoes
a drastic change in conformation
on binding
the heme
group. This is evident on comparison
of the properties of myoglobin
and especially
hemoglobin
globin with the corresponding
heme proteins.
Both globins have a much lower stability to denaturing
agents than the hemoproteins, are more susceptible to digestion by proteolytic
enzymes, and cannot be
easily crystallized
( I[55, I 56, 197). They also differ from the full hemoproteins
in
the reaction with dyes (23), in immunochemical
behavior ( I 37), and in the helical
content (I gg). Hemoglobin
globin shows a behavior different from hemoglobin
in
associating-dissociating
into subunits; at neutral pH and moderate
ionic strength,
its molecular
weight is about 41,000, a little higher than half that of hemoglobin
(15, 156, IW
Upon the addition

of the heme to the globin all the original
properties are at
once restored (I 5 I, I 58). These observations
indicate that the heme group is an
essential factor in determining
the conformation
of native hemoglobin
and myoglobin.
The heme group acts on the conformation
of the protein through
the various
linkages that it makes with the polypeptide
chain. In this respect the porphyrin
much more so than the iron-protein
part of the heme is especially important,
bonds. The properties
of the compound
obtained
by coupling
protoporphyrin
to
hemoglobin
globin (protoporphyrin-globin)
are very similar to those of hemoglobin
and correspondingly
different
from those of globin
(158). On the other hand,

ERALDO

I30

ANTONINI

Volume 45

although
reconstituted
hemoglobins
containing
hemes different
from protoheme
have the basic characteristics
of hemoglobin,
they show quantitative
differences
from protohemoglobin
in many properties
(I I, 22, I 52, I 59). This suggests that
the effect of the nonnatural
hemes on the conformation
of the protein is less pronounced,
this being correlated
with the lower affinity that these heLmes have for
globin. Thus, for instance, the resistance to heat denaturation
decreases in the
order proto-, meso-, deutero-, hematohemoglobin
(2 2).

STRUCTURE

-4ND

FUNCTION

OF

MYOGLOBIN

As we have seen there are two aspects of the functional
properties
of hemoglobin and myoglobin
that, to some extent, can be analyzed
separately:
r> the
possibility
of reversible
oxygenation,
and 2) the character
of the reaction with
oxygen, carbon monoxide,
and other ligands. In the previous section we dealt
with the basic function of the molecule 9 we shall now try to analyze the equilibria
and kinetics of the reactions of myoglobin
and hemoglobin
with the physiological
ligand, oxygen, and with other substances and their relation
to the structure

of
the proteins.
l

Structure of Myoglobin
h4yoglobin
represents the simplest case of a respiratory
heme protein,
both
structurally
and functionally
(I 35, I 96). The molecule consists of one polypeptide
chain containing
one heme group. From the brilliant
findings of Kendrew
and
co-workers
(I 05-108),
we now know the details of the structure of this protein in
the crystal and the arrangement
in space of most of the atoms which constitute

it.
The over-all shape of the molecule appears spherical in the X-ray model, in
agreement
with the hydrodynamic
behavior in solution (Table 2). As mentioned
before, the heme group is near the surface of the molecule in a pocket formed by
the folding of the chain; the plane of the heme group is approximately
normal to

the surface of the molecule.
The polypeptide
chain of myoglobin
is made up of about 153 residues; most
of the chain (~80 %), both in the crystal and in solution (39, I 98, I gg), is arranged
to form a right-handed
a-helix.
The molecule
contains little water; the hydrophobic side chains are mainly
directed
toward
the interior
and the polar ones
toward the surface.
The structure of the native protein appears to be held together especially by
nonpolar
interactions
between side chains and between the latter and the heme
group. As a whole the nonpolar
interactions
are far greater than the polar ones.
The polar side chains which lie predominantly
on the surface of the molecule
interact
especially with the solvent, seldom with other polar side chains (105).
However, in the native protein, not all the polar side chains are available for interaction with the solvent; this would be suggested by the fact that some of the ioniz-


1965

January

HEMOGLOBIN

AND

MYOGLOBIN

‘3’

TABLE
2. chemical, Physiocochemical, and sbectral !v-obeyties
of mammalian myoglobins
Chemical

and Physicochemical

Properties

Horse

Iron content,
76
Nitrogen,
y0
Minimum
molecular
weight from chemical
analysis
Molecular

weight
Osmotic
pressure
Sedimentation
and diffusion
Sedimentation
constant
S$,,,- (Svedberg
unit)
Diffusion
constant
I 07 cm* se@
tsoionic
point
Number
of amino
acid residues
Partial
specific
volume
Spectral

Properties

Mb

(Horse

Max.
E x

MbO:!

Max.
r x

MbCO
Mb’

Kefs:
globin-

horse
58.

Human

o*297

0.318
16.7

18.800

‘7 l 45o

7.80

I
560

13.0

580
14.4
579
‘3.9
630
3*6
40,

197;

16.7
17.816

‘54
0.743

rnp

10-3

Mb

9.6

10-3

Max.
rnp

e x 10-3
Max.
rnp
E x 10-3

Sperm Whale

‘7 .goo
I .81

I .96

Visible

rnp

Mb

18.482

Myoglobin)

myoglobin-4,

Mb

human

542
‘3.9

540
‘5*4
5oo
9*7
myoglobin-130,

‘53

Soret

Ultraviolet

435
I21

424
207
4o9
‘7’
146;

280
31.2
sperm

whale

myo-

able groups cannot be titrated in the native protein, except after denaturation

(44) .
Table 2 gives the most recent values of various physical and chemical parameters of mammalian myoglobins. This protein, as obtained by the usual methods
of purification, is heterogeneous and contains at least three different electrophoretic components (4, 40, I 30, 150, I 97); the data reported in Table 2 refer,
when possible,to the major homogeneouscomponent.
The amino acid composition of myoglobin from several speciesis now known
(4, 130, 146, 157).
Most of the sequence of human (91) and sperm whale myoglobin (58) has also been determined by chemical methods. In the case of sperm
whale myoglobin, the greatest part of the amino acid sequencehas also been established independently by X-ray analysis at high resolution (107).
The identity of several chemical and physicochemical properties of the
myoglobins extracted from the adult and fetal musclesof the same speciesindicates that there is no fetal form of this protein (163).
The molecular properties and the over-all structure of myoglobin from different animal speciesappear to be very similar (I 78), although there may be differ-


ERALDO

I32

ANTONINI

Volume 45

encesin amino acid compositions. However, the sequencestudies,when available,
show that most of the chains have similar amino acid sequences(I 75).
Functional Propertiesof Myoglobin
The ferrous iron in myoglobin and hemoglobin reversibly binds a number of
gaseousand nongaseousligands: 02, CO, NO, cyanide, alkalisocyanides, and aromatic nitroso compounds. The addition compounds obtained with all these
ligands are qualitatively similar: one molecule of ligand is bound per heme group;
the combination of the ligand is accompanied by spectral changes in the visible,
infrared, and ultraviolet regions of the spectrum and by modifications of the
magnetochemical character (I I 3, I 29). Also the oxidation-reduction equilibrium

(Mb g lMb+ + e) can be considered analogous to the oxygen or carbon monoxide equilibrium.
The reaction of myoglobin with ligands should correspond to a very simple
model since this protein contains only one heme per molecule. It must be kept in
mind, however, that the equations shown below for such a simple model are
tacitly based on certain simplifying assumptions, some of which are discussed
later. From this point of view, the detailed analysis of the reactions of myoglobin
is very important becausethe reaction mechanism can be subjected in this caseto
stringent tests both in the equilibria and kinetics, which is still almost impossible
for hemoglobin.
The equilibrium between myoglobin and a ligand can be represented by
Mb

+

X

G,

k’
k

MbX

K = k’/k

(3)

where X is the ligand; Mb and MbX the unliganded and liganded forms of myoglobin; k’ and k, respectively, are the “on” and “off velocity constants; and K, the
equilibrium constant. It follows from simple physicochemical considerations that
(MbX)/(Mb) = K(X)

(4)

and kinetically
[d(MbX)]/(dt)

= k’(X)

(Mb) - k(MbX)

(5 >

If AH and As represent the enthalpy and entropy changes, AH5 and AS5 the heat
and entropy of activation of the over-all reaction:
AH=AH#o,
--Him
As = As5 on - ASI off

(6)

Usually in these equations it is legitimate to identify concentrations with activity
because the reactions are followed in very dilute solutions. In the case of gaseous
ligands the partial pressurein the gasphase (px) is often used instead of the molar
concentration. The conversion from px to (X) is easily made with the use of solubility coefficients.
When two different ligands, X and Y, are present under conditions where


January

rg65

HEMOGLOBIN

AND MYOGLOBIN

I33

there are no free gas binding groups, the following equation should apply at equilibrium :
MbX+Y$MbY+X

(7)

The partition of myoglobin between the two ligands, it& may be defined as (I

71)

:

M =-Mm
mm
Hence
A4 = Ky/Kz

and

AH,

=

AH, - AH,

Kinetically the rate of displacement of the ligand X by Y-that
reaction
MbX+Y--,MbY+X

(9)

is the rate of the
(4

-is given, under certain simplifying assumptions,by (65) :
- [d(MbX)]/dt
(MbX)

kz k;(Y)
= k:(x) + k;(Y)

where the subscripts indicate the velocity constants for the reaction according to
equationI with the ligands X and Y. Putting
- [d(MbX)]/dt
MbX

>

as r:

;= &+$g

(4

In other words the reciprocal of the observed rate, if (Y) is taken as constant
when plotted versus (X) should give a straight line, the intercept of this line on the,
ordinate being I /k, and the slope k’,lk, .k’, (Y).
The displacement reactions can thus be used for determining the dissociation
velocity constant and for checking the over-all con&ency
of separately
determined rate constants.
The data obtained by several authors in the past 20 years on the equilibria
and kinetics of the reactions of myoglobin with oxygen, carbon monoxide, and
other ligands showed a behavior corresponding to a good approximation to the
simple equations reported above. Thus, the oxygen equilibrium of myoglobin from
different specieshas been found to be a rectangular hyperbola when the fractional
saturation of the ligand is plotted versus the partial pressureof the gas (88, 150,
195)
(Fig. I). The kinetics of the reaction of myoglobin with oxygen, carbon
monoxide, and other ligands also showed the expected behavior, being accurately
first- or second-order kinetics (65, 124).
It must be noted that although, asmentioned before, the myoglobin preparations, with which most of these experiments have been done, are heterogeneousin
electrophoresis or in chromatography, the functional behavior of these fractions
is essentially the same (I 49).
l


ERALDO

FIG.

pH

7.45;

I.
r/2

Vdume

ANTONINI

I

I

I

I

1

I

I

1

2

3

4

5

6

7

Oxygen
dissociation
= 0.05.

curves

of human

myoglobin

at different

temperatures

45

(I 50).

In spite of all this a closer examination of the data available until 1960 shows
discrepancies in the absolute values of the various constants and inconsistencies in
the equilibrium and kinetic data. These may be due to experimental errors, to
differences in the methods and conditions of measurements used by the different
authors, or to differences in the myoglobin preparations. In view of this, Antonini,
Gibson, and Wyman (I 4) have recently reinvestigated the reactions of myoglobin

with oxygen and carbon monoxide, using a single crystalline sample of horse myoglobin and making all the equilibrium and kinetic measurements under identical,


January

1965

HEMOGLOBIN

AND MYOGLOBIN

carefully controlled conditions. The
new results, reported in Table 3, are,
with few exceptions, in general agreement with the previous ones and confim again that the oxygen equilibrium
of myoglobin is represented by a rectangular hyperbola and that the kinetics of the reactions conform to equation 5. However, it may be seen from
Table 3 that, especially in the caseof the
oxygen reactions, there is quite a large
difference between the value of the
equilibrium constant directly determined and the ratio of the two velocity
constants. Even more important is the
observation that when the displacement
reaction of oxygen by carbon monoxide
was studied over a very large range of
oxygen concentrations, it was found
that the observed rates departed, at
high oxygen concentrations, from the
behavior predicted by equation 12. The
meaning of theseobservations is not yet
completely clear. In any case they indicate that the model used in describing the myoglobin reactions may be
oversimplified; a possible explanation

of the departure from the simple
model could be the formation of shortlived intermediates in the reaction with
ligands.
Values of the equilibrium and kinetic constants for the reaction of myoglobin with two other types of ligands
are reported in Table 4. Comparison of
the data in Tables 3 and 4 shows that
the change in aflinity on passing from
one ligand to another is reflected by
differences both in the ratesof combination and in the rates of dissociation.
In contrast with the behavior of
hemoglobin, the reactions of myoglobin
with ligands are not affected to any
considerable extent by changes of the
medium. Thus the oxygen affinity and

m
c;
7

CD

“0
c(

“0
P-l

x

x

09

0,

P

cc)

0

b
d

*
0,
“0
d

’


ERALDO

136

ANTONINI

Volume 45

TABLE
4. Equilibrium and kinetic values for reaction of myoglobin
with isocyanides and nitroso aromatic compounds
R, IO-’ M-1

Compound

Methyl
isocyanide
Ethyl isocyanide
Isopropyl
isocyanide
n-Propyl
isocyanide
Tert-butyl
isocyanide
Isobutyl
isocyanide
Nitroso
benzene
o-Nitroso
toluene
m-Nitroso
toluene
p-Nitroso
toluene
20 C; borate

* pH
Refs.

buffer
6.8 phosphate,
3, 67, 112.

k’, IO-'

u-1 MC-'

‘35

65

20*
I

.4*

II”
0.1”

=7
43
I

2

.6

k, see-‘

2.6
0.23
0.8

5-2
28

0.3
0.8

14

o-4

2
0.2

0.5

pH 9.1.
25 C.

the shapeof the oxygen equilibrium curve are essentially independent of pH (I 50,
I g5), salt concentrations (I 50), and the presenceof substancesreversibly bound by
the protein, suchas the dye bromthymol blue (23). Whenever the kinetic constants have been measured under different conditions, they also have been found
to be independent of the composition of the medium (65). These findings in
turn imply that the reactivity of the groups of the protein which bind hydrogen
ions, salts, dyes, etc., remains unchanged in the liganded and unliganded form.
The deduction accords with the fact that, again in contrast with hemoglobin, the

reaction of myoglobin with ligands is not accompanied by any large change in
the conformation of the protein. The best proof of this is the fact that crystals of
deoxygenated myoglobin are essentially isomorphous with those of the liganded
derivatives (I 03).
Another characteristic aspect of the oxygen equilibrium of myoglobin is its
insensitivity to chemical or physicochemical modifications of the protein part of
the molecule, as is shown in Table 5. These data suggestthat in myoglobin the
reactivity of the heme iron is essentially controlled by only the part of the chain
which interacts directly with the heme group. In agreement with this would appear
the fact that the value of Pt is very nearly the same in myoglobins from different
animal species(88, I 50, I 63, 164). However, the myoglobin of Aplysia (I 57) and of
other mollusks (I I 8, I I g) showsa much lower affinity for oxygen than vertebrate
myoglobins; but in this case the amino acid composition is so different from that
of other myoglobins (I 57) that it suggestsdifferences even in the environment of
the heme group (148).
Large changesin oxygen affinity are obtained when the heme group is modified, as shown by the fact that deuteromyoglobin has an oxygen affinity about
three times greater than protomyoglobin (147). The rate of combination with
carbon monoxide is also three times higher for deutero- than for protomyoglobin.
Mesomyoglobin, on the other hand, showsthe samerate of combination with carbon monoxide as protomyoglobin (I 81). In the caseof simple pyridine heme cornpounds, the rates for the combination of carbon monoxide with proto-, deutero-,


January

1965

HEMOGLOBIN

TABLE

5. 02 equilibrium

of mammalian

Treatment

None
Acetic

AND

I37

myoglobin after various treutments

Modification

Produced

Digestion
with
carboxypeptidases
Solution
in 5 M urea
Reconstitution
from
globin
and protoheme
Reconstitution
from
globin

and
deuteroheme

Substitution
of H atoms for vinyl
groups in positions
2 and 4 of porphyrin
conditions:

pH

7; r/2

=

0.1;

20

AH, kcal/
mole 02 gas

hl Pl/¶

None
Acetylation
of amino and other
groups
Removal
of several C terminal

residues
Partial
unfolding
None

anhydride

Experimental
r5o*

MYOGLOBIN

C; Mb

-0.15

-1397

-0.13

I

-0.18
-0.

IO

I

-0.14

-‘3*4

-0.60

concentration

-13.2

I

-10-4

M.

Refs.

16, 147,

and mesohemesare in the ratio I : 3. I : 2.5 (I 80). It is thus clear that the functional
properties of myoglobin are determined by the specific effect that the native protein has on the intrinsic reactivity of the heme.
It should also be mentioned that the oxygen equilibrium of myoglobin is not
affected by large changes in the myoglobin concentration (I 50) ; this indicates the
absenceof functional interactions of any kind between different myoglobin molecules.
In conclusion, it appears that the behavior of myoglobin largely corresponds
to a simple model of an oxygen carrier. The basic function and the affinity for the
ligand depend uniquely on the interaction of the heme group with its binding site
in the native protein. In this respect myoglobin differs substantially from hemoglobin, in which the character of the reactions with the ligand is dominated by
several functional interactions of the heme with other distant groups in the molecule. However, even in myoglobin, deviations from the simplest behavior can be
seenwhich may be associatedwith small changesin conformation in the part of the

chain which surrounds the heme on attachment of the ligand to the iron. It should
be noted in this connection that the state of the heme iron has, even in myoglobin,
a great effect on the denaturation of the protein (47).
STRUCTURE

AND

FUNCTION

OF

HEMOGLOBIN

The most characteristic aspect of the functional behavior of hemoglobin is
that the reactivity of the heme iron is dominated by interactions with other groups
of the molecule which may occupy positions distant from the heme group. As a
reciprocal effect, of course, the reactivity of thesegroups dependson the state of the
iron atom (204, 205). This behavior is linked to the particular structure of the
hemoglobin molecule. In comparison with myoglobin, just as the function of hemoglobin represents a higher degree of physiological specialization, the structure of
the protein also showsa more complex molecular organization.


I38

ERALDO

ANTONINL

Volume 45

Over-All Structureand Amino Acid Sequence
of IMammalianHemoglobins
Most details of the structure and conformation of hemoglobin have now been
revealed as a result of chemical (42,43, 72, go, Iog, I IO, I 75) and physicochemical
studies (I I 3, 168,204) in progressfor a number of years and of the recent beautiful
X-ray analysis of the crystals made by Perutz and co-workers (54, 131, 132).
Mammalian hemoglobin in dilute solutions, at neutral pH and moderate
ionic strength, has a molecular weight of about 66,000, corresponding to four
times the minimum molecular weight based on the iron content (2, I I 3). The
whole molecule is made up of four polypeptide chains, identical in pairs (92, 132,
138), each one containing one heme group and having a molecular weight of
about 16,000. The different chains have been called the a and @chains on the basis
of their N-terminal sequence(138).
The X-ray studies have revealed that each one of these subunits has a structure similar to that of myoglobin, especially in regard to the folding of the polypeptide chain, the number and distribution of the helical and nonhelical segments,
and the position and orientation of the heme group.
The four hemoglobin chains are arranged in the full molecule to form an irregular tetrahedron, with the heme groups on the surface and near the anglesof it.
The structures of the a and /3 chains are complementary, and these chains are in
close contact, in contrast to the much looser association between the identical
chains.
Mammalian hemoglobins contain several SH groups per molecule (50) : some
of thesecan react with SH reagents in the native protein, the others only when the
protein is denatured or unfolded. Human hemoglobin has two freely reacting SH
groups per molecule, one in each of the @chains (140); their position in the chain
is closeto the heme (54).
In native hemoglobin a number of ionizable groups cannot be titrated. The
unmasking of these groups on denaturation and the significance of this phenomenon have been investigated in detail (I 83, 188).
In the amino acid sequencesof myoglobin and of the hemoglobin a and p
chains a number of amino acid residuesoccupy identical or analogous positions
(42, 54, 108); this is particularly significant in the region of the chain which surrounds the heme group.
In the hemoglobins from different animal species,the amino acid sequenceof

a given chain is largely the same: the number of different residuesincreaseswith
the distance of the animals in the zoological scale (42, 91a). The fl chain shows,in
this respect, greater differences than the QIchain; the a chain, which is also more
similar to the myoglobin, has a more uniform structure in the different species.
These findings have suggestedthat myoglobin and the hemoglobin chains are the
product of the evolution of one primitive molecule, and that the number of differencesin the sequencebetween the two speciesis an expression of mutations correlated to the age at which the two speciesbranched off from a common ancestor
(2 10).

Different hemoglobins can be found in the same animal species(g2-g4). In
adult normal humans, the hemolysate contains over g5 % of a major homogeneous


January

1965

HEMOGLOBIN

AND MYOGLOBIN

I39

component (Al) and 2-4 % of a minor component (AZ), which has cychains identical to those of Al but different P chains (called 6 chains) (92). In fetal life, the major
component of the hemolysate is a different hemoglobin, fetal hemoglobin; here
again the cychains are the same as in hemoglobin Al, but the other chains are
different (y chains) (I 75, I 76). A similar situation is encountered in the other
animal species; however, the number and proportions of the hemoglobin components present in adult blood may vary from speciesto species(74, 94).
In humans, “abnormal hemoglobins” have been found together with the
“physiological” hemoglobins in hereditary pathological conditions (92-94). These
“abnormal hemoglobins” are characterized by a different chemical structure,

often due to a substitution of only one amino acid residue in one of the hemoglobin
chains (92).
The functional properties of the various “physiological” and most of the Ccabnormal” human hemoglobins are remarkably similar; a notable exception, in this
respect, is hemoglobin H. This “abnormal hemoglobin” is thought to be made up
of four identical chains having the structure of the normal /3 chains (92); it shows
an oxygen equilibrium wholly different from that of normal hemoglobin (36), the
equilibrium curve being hyperbolic, like that of myoglobin, rather than sigmoid.
Physicochemical Properties

Some chemical and physicochemical properties of human hemoglobin are reported in Table 6. It must be pointed out, however, that the molecular properties
and the conformation of native hemoglobin are not fixed, but appear to change
widely and in a complex way with a number of variables which involve both the
medium in which the protein is dissolved and the presenceor absenceof a ligand
on the heme iron.
The complexities of the molecular architecture of hemoglobin in comparison
with myoglobin and the structural and functional association between the various
chains are alsoshown by the character of the interaction between the heme and the
protein. For each individual chain the location of the heme group, the type of linkages,and the kinetics of combination of heme are very similar in hemoglobin and
myoglobin; however, in hemoglobin there is a strong stabilizing interaction between the heme binding sites (13, 30). Thus, in the processof reconstitution of
hemoglobin from globin and protohematin no significant amounts of intermediates
can be detected. This means that a stable conformation is only achieved when all
the four chains have bound the heme group-in other words, that the presenceof
the heme on one of the chains influences the binding of heme by the others.
One of the most striking phenomena in hemoglobin is the tendency of the protein to dissociatereversibly into subunits of lower molecular weight under a number of mild conditions which do not involve lossof the oxygen capacity (168). The
average molecular weight gradually drops from 66,000 to about half of this value
in acid (60) and alkaline pH (82), in concentrated salts (38, 76, 79, 165), and in
urea solutions (80, I I I, 182). The product of the splitting is, almost certainly, a
molecule made up by one a and one /3chain (I g, 78, I 68). It was thought for some
time that the dissociation into halves was “asymmetrical,” i.e. produced cy2and

ERALDO

r4o
TABLE

ANTONINI

V&me

45

6. Chemical, physieochemical, and spectral properties of human hemoglobin
Chemical

and Physicochemical

Human

Properties

Iron content,
y0
Nitrogen,
70
Minimum
molecular
weight
from iron analysis
Molecular

weight
from :
Amino
acid sequence
(C&Z)
Globin
(&I*)
CYChain
(without
heme)
p Chain
(without
heme)
Molecular
weight
(neutral
pH; r/2 0.01-0.2)
Osmotic
pressure
Sedimentation
diffusion
Sedimentation
equilibrium
Light-scattering
Number
of amino
acid residues

Hb

Ref.

o*335
16.9
16.700
64,450
61 &I90
15,128
‘5,870

42

79
80
‘35
165

42
a

Chain
0 Chain

42
s--e---

Sedimentation
constant
S”,,, (Svedberg
Diffusion

constant
107 cm2 see-l
Partial
specific
volume
Spectral

Properties

unit)

99, 1659 =77
80, 158, 185
I85

4*35-4.45
6.d.4
0.749

Soret

Visible

Ref.
--

Hb
HbOz

max.

w

e”

loos

x

max.
e x

HbCO

u
IO-’

max.
e x

* Molar

w
10-3

max.
e x

Hb+

430
“9
4x2
‘35
419
IQ1
405
‘55

rnp
IO+

extinction

coefficient

based

555
12.3

on heme

I

54’
13.8

577
14.6

539
13.4
500
9-5

569
‘3.4
630
4.1

I

158

equivalents.

02 subunits; this was deduced on the basis of the results of “hybridization”
experiments (95, g6). In these experiments mixtures of hemoglobins differing in the cy or
@chains or in both are exposed for some time to conditions where dissociation may
occur. On reversal of the treatment, analysis of the mixture shows the formation of
new components corresponding to “hybrids,”
containing in the molecule chains
from the different original hemoglobins. The asymmetrical
dissociation into cy2
and & subunits was inferred mainly from the fact that no hybrids were formed containing two different types of cy or @ chains in the same molecule, the two a! or p
chains in the new components being identical. It now seems clear that the formation of hybrids is not associated with the dissociation into halves but probably with
the further splitting into quarter molecules, i.e. individual chains (19, 78, 200).
Recent physicochemical
measurements would actually indicate that molecular

Jimuary

rc$~

HEMOGLOBIN

AND

MYOGLOBIN

weights lower than that corresponding
to half are obtained
conditions
of reversible dissociation
(7 I, 77).
Reuctions of Hemoglobin

=4=

for hemoglobin

under

With Ligands

As mentioned
before, the reactions of hemoglobin
with oxygen, carbon monoxide, and other ligands were characterized

by interactions
between the ligand
binding
site and other groups in the molecule (204, 205). These groups may have
the same function,
i.e. may be other heme groups or may involve a different
function such as ionizable
groups.
These interactions
are the basis of the characteristic
sigmoid shape of the
equilibrium
curve, the Bohr effect, and of the effect of salts and carbon dioxide on
the functional
behavior
of hemoglobin.
It should be emphasized
that the functional interactions
do not need to imply proximity
between the interacting
groups.
Actually
it is now clear in some cases that the groups that interact are far apart
and may be located in the same or in another of the polypeptide
chains which
make up the full molecule.
Although
the details of these interactions
are still
largely unknown,

there are indications
that they involve the protein as a whole
and are mediated
through conformational
changes of the protein.
The functional
properties
of hemoglobin
change widely when the protein is
studied under different conditions
so that in order to understand
the behavior
of
this protein the systematic exploration
of the widest range of experimental
conditions is extremely important.
Information
on the molecular
mechanisms
underlying
the reaction of hemoglobin with oxygen and other heme ligands can also be obtained by studies of the
functional
behavior of the protein
after physicochemical
or chemical
modifications which, however, leave unchanged
its oxygen-combining
capacity. The main
difficulty
encountered

in this approach
is the identification
of the real change
produced
in the protein after the various treatments.
In cases such as hemoglobin, in which the protein molecule
participates
as a whole in the function,
it is
often meaningless
to speak of modifications
produced
at the “active sites” because
changes in any part of the molecule may have a dramatic
effect on the functional
behavior, not because the modified
groups are directly
involved in the function,
but because of the changes of conformation
of the protein,
which might
be
aspecific, brought
out by the treatment.
Oxygen Dissociation

Curve of Hemoglobin

and Heme-Heme Interaction

The oxygen dissociation
curve of mammalian
blood, i.e. the curve obtained
by plotting
the fractional
saturation
with oxygen versus the partial
pressure of
the gas, is not a rectangluar
hyperbola,
but has a sigmoid shape in vitro and in
vivo (3 I:j 57). This phenomenon,
which has great physiological
significance
since
it increases the efficiency of the blood as an oxygen carrier between the lungs and
the peripheral
tissues, is due to the inherent
properties
of hemoglobin
(Fig. 2).


I42

FIG.
2. Oxygen
dissociation curve
of sheep hemoglobin
(173).

pH 9.1 in 0.2 M borate.
X Observation
at 19 C; 0 observation
at 0.2 C but
with
oxygen
pressures
all multiplied
by a constant
factor
of 3.737.

ERALDO

ANTONINI

Voku%e 45

60

40

20
x

oxygen

/(I

I1

2

I
4

pressure (mm Hg]
I
I
I
I
6
8

[
66

1

Thus similar behavior is shown by purified hemoglobin
solutions at high dilutions
(8ga), after crystallization
(I~I),
and even after the reconstitution
of the protein
from heme and native globin (I 51).
The sigmoid shape of the oxygen dissociation
curve of hemoglobin
means
that the affinity of the oxygen binding
sites increases with the increase in oxygen

saturation;
in other words the binding
sites are not independent,
the presence of
the ligand on some of these having the effect of increasing
the affinity of the remaining
ones.
Whatever is the mechanism
by which this effect is produced,
it may be called
“heme-heme
interaction”
(I 13). This expression simply states the fact that there
are functional
interactions
between the heme groups and does not really involve
any exact model nor does it imply direct interactions
or even that the interactions
occur between the heme groups in the same molecule.
It must be noted that here, as in the case of myoglobin,
we assume that the
reaction between
the protein
and oxygen or another
ligand is a true reversible
equilibrium..
From the kinetic point of view, it is clear that the rates of the reaction
of
hemoglobin
with ligands cannot be described

by a single combination
and dissociation velocity constant because they should reflect the interaction
between the
binding sites which characterizes
the equilibria
(65).
In the 60 years since the shape of the oxygen dissociation
curve of hemoglobin
was first recognized,
a great number
of different
models have been adopted by
several authors to describe quantitatively
the equilibria
and kinetics of the hemoglobin reactions. It is outside the realm of the present review to describe aPI these
models and their development
parallel
to the increasing
knowledge
of the phenomenological
aspects of the equilibria
and kinetics and especially of the structure
of the protein (7, 49, I I 3, I 36, 168, I 70, 204). Most of the models imply a detailed

1
1:


January

1965

HEMOGLOBIN

AND MYOGLOBIN

I43

and exact mechanism of the heme-heme interactions, which, as we shall discuss
later, is still largely unknown.
At present, the equilibrium data are usually described in terms of the Hill
empirical equation (86) or according to the Adair four-stage model (I). The
Adair model describes the reactions of hemoglobin in terms of individual equilibrium and kinetic constants corresponding to the four hemes contained in the
molecule. The scheme,as it is generally written, is as follows:
Hb4

+

X

k: \

-

=

HbqXl
Hb4X2

+

+

X
X

kg

-

Kl

Hb4X2

x2 = -k::

,

k2
kk

r-

+

X

,
HhX3

K3

=

Hb4X4

K4

=

ki \

-

kl

k2

k3

HbJX3

ki

Hb4X1
kl

k4

( I 3)

kg
k3

ki

k4

where Hbd is the hemoglobin tetramer, X is the activity of the ligand, and EC;k’,
and k the equilibrium constant, the combination constant, and the dissociation
velocity constant, respectively. The determination of these individual constants is
a formidable problem which has been attacked by Roughton and co-workers for
the equilibrium (I 16, I 73) and by Gibson and Roughton for the kinetic constants
[(reviewed by Gibson (65)]. These authors have made remarkable achievements
in determining the values of individual constants through the use of very precise
methods for equilibrium and kinetic measurements. When the model is applied
to the very precise oxygen equilibrium data for sheep hemoglobin, the sigmoid
shape of the curve appears to be due to a very large value of X;, which exceeds
those of ICI, &, and El3 (I 73).
Although the Adair model as formulated above is the more correct in principle among those so far proposed, it may be oversimplified since it appears now
that some aspectsof the equilibria and kinetics of hemoglobin cannot be accomodated in its framework, as is discussedin more detail later.
The Hill empirical equation, which has no direct physical meaning, is the
following :
Y/b

-

y)

=

KXn

( I 4)

where y is the fractional saturation with oxygen, X the activity of the ligand, and
K and n empirical constants. The equation fits quite well the equilibrium data in
the middle range and the parameters K or xi (the ligand activity corresponding
‘to half saturation) and n, measured when y is near x, may be used to define the
affinity for the ligand and the shape of the equilibrium curve under a given set
of conditions. Moreover the value of n may be taken as an over-all measure of the
heme-heme interactions (204). For mammalian hemoglobins under physiological
conditions the value of n in the Hill equation for values of p close to pi is about
3. This indicates that the number of hemes that interact must be at least equal


ERa4LDO

I44

ANTONINI

I~olurr~e $5

or greater than 3. It is easy to show that in the case of infinite interaction
energy
the value of n approaches
the number of interacting
sites (59).
When there are no free gas binding
groups the equilibrium

of hemoglobin
with two different ligands, X and Y, is described by an equation
identical
to that
to Roughton
(I 7 I), the partition.
given for myoglobin.
In this case, according
constant between the two ligands corresponds
to the ratio of the equilibrium
constant for the reaction
with
the same ligands
of the fourth
he-me, i.e.
(&X)/(&Y).
In the case of the gaseous ligands, the affinity of hemoglobin
itI=
for carbon monoxide
is about 300 times greater than that for oxygen (24, 97) and
(65).
that for nitric oxide about 1500 times greater than for carbon monoxide
It is now pertinent to consider the phenomenol .ogical aspect of the heme-heme
curve and the affinity of
interaction
and to describe the shape of the equilibrium
hemoglobin
for oxygen and other ligands under different conditions?

PHENOMENOLOGICAL

A4SPECTS

OF

HEME-HEME

INTERACTIONS

#Species
Vuriations
All the mammalian
hemoglobins
show a great similarity
in the shape of the
oxygen equilibrium
curve, under identical
conditions,
although
the values of p;
may be different. The shape of the oxygen equilibrium
curve may be quite different in ,the case of nonmammalian
(I I 7, 153) and especially invertebrate
hemoglobins (6 3 1 7, 100, 119); however, heme-heme
interactions
are generally present
also in these hemoglobins
Especially in the case of invertebrate
hemoglobins,
the

different shape of the oxygen equilibrium
curve may be associated with the different over-all structure of the protein,
such as the different molecular
weight and
number of subunits (I 8, I I 9, I 43).

Type of Ligand
The shape of the equilibrium
curves of the other ligands that mav be bound
to the ferrous iron is essentially the same as for oxygen. This has been shown in
the case of carbon monoxide,
using whole blood, by Haldane
et al. (56) and,
more recently, by Joels and Pugh (97), although
Roughton
has found significant
dif!I’erences between the shapes of the oxygen and CO equilibrium
cur&s at the
top and bottom regions (I 7 I).
The value of n for the equilibrium
of hemoglobin
and ethyl isocyanide is 2.5
in very dilute hemoglobin
solutions (98), a value identical
to that found for oxygen under the same conditions.
2 To compare
the shape of the equilibrium
curve under different
conditions
in which the

for the ligand
may vary,
it is convenient
to plot the fractional
saturation
y or
of the ligand activity
or of the partial
pressure of the gas.
log [y/(1 - r>l versus the logarithm
The slope of the curve in the plots of log [r/(1
- $1 versus log X, at the point where y = 0.~5,
gives n.
affinity


January

1965

HEMOGLOBIN

,4ND

MYOGLOBIN

I45

In the case of nitroso aromatic compounds, the value of n has been found to
be very low ( I a$, but values of n near 2.3 have also been reported (I 74).

Even in the case of the oxidation-reduction
equilibrium, the shape of the
curve, at alkaline pH’s, is very similar to that found in the oxygen equilibrium
(26, Igo, 191).
It must be noted that for these various ligands, the value of the over-all affinity (or xi) is very different, so that it must be admitted that the heme-heme
interaction is largely independent of the affinity of the hemes for the ligand.
It is striking that in the reactions of ferrihemoglobin with ligands, the situation is wholly different: in this case the value of rz is often exactly I (62, I I 3),
showing the absenceof interactions and the equivalence of the ferrihemes in the
reaction with the ligand.
Temperature
The shape of the oxygen equilibrium curve is also the sameat different temperatures, at least from 5 to 95 % saturation. In very precise measurements on
sheep hemoglobin, Roughton and co-workers have found, extended at the very
bottom and very top of the oxygen equilibrium curve, a small change in shape
at these extreme regions at different temperatures (I 73).
The invariance or variance of the shape of the equilibrium curve with temperature implies equivalence or difference in the values of AH, the heat of reaction, for the various combining sites; the data obtained by Roughton indicate
that the heat for the first combination is about 5 kcal higher than the average
heat.
The temperature generally has a large effect on the affinity of hemoglobin
for the ligand (168, 169, 204). In the caseof oxygen, for human hemoglobin at
alkaline pH (where the contributuon of the heat of ionization of the grou ps involved in the Bohr effect is negligible), this effect accounts for a AH of about
- I 4 kcal for the combination of I mole of gaseousoxygen (27).
There are cases,however, in which the effect of temperature on the oxygen
affinity is practically zero, as in tuna fish hemoglobin (I 53). Even in this casethe
shape of the equilibrium curve doesnot change, at least in the middle range, with
temperature.
It must also be noted that in modified hemoglobins, in which the values of
n and pt may be widely different from those of normal hemoglobin, the effect of
temperature on the oxygen affinity doesnot change after the modification (I 6).
All these facts and the invariance

of the shape of the equilibrium

curve

with

temperature would suggestthat the heme-heme interaction is largely an entropy
effect.
Concentrationof Hemoglobin
The shape of the oxygen equilibrium curve of hemoglobin showssmall but
significant changes with the concentration of the protein. These changes, which


ERALDO

ANTONINI

Volume 45

3
c

2

I

4

5

6

7

8

9

IO

II

PH
F~.G. 3. Bohr efftct and vakes
of n in various
buffers
without
concentration
3-5 mg/ml;
temperature
20 C (20). Q) Results in 0.2
0.4 M acetate; V results in 0.05 M borate;
A results in 0.4 M glycine-NaOH

M

added salt.
phosphate;
mixtures.

Hemoglobin
0 results in

largely account for the difference in shape of the oxygen dissociation
curve between
whole blood and dilute hemoglobin
solution (8ga), are represented
by a decrease
in the sigmoid
character
with the dilution
of the protein.
In the case of human
hemoglobin
the value of n at high salt concentration
(0.5 M phosphate)
passes
from 3.3 to about 2.5 for a change in hemoglobin
concentration
from 6.5 to 0.25
mg/ml
(167). Changes in the same direction,
but differing
in size, have been reported for other hemoglobins
investigated
(89”).

The effect of pH on the shape of the oxygen equilibrium
is, of course, correlated to the effect of pH on the oxygen afEinity, the Bohr effect, which is discussed
later in more detail.

January

1965

HEMOGLOBIN

AND

MYOGLOBIN

=47

1.

0

log c
1
-5

I.

I
-3

-4

1

-2

I
-1

1
0

4

J

1
0

+1

I

3

n

1

log

c
I

-5

1

-4

FIG. 4. Oxygen
affinity
and
solutions
(166). pH 7.0; temperature

shape
20

I

I

-3

-2

of 02 dissociation
C.

In mammalian
hemoglobins
the shape
mains essentially the same at different pH’s

changes; in human hemoglobin,
for instance,
pH 5 to I o (20) (Fig. 3). On the other hand,
small differences in the shape of the extreme
with pH (I 73); however, recent measurements

-t
curve

of Hb

in different

I
chloride

of the oxygen equilibrium
curve re(7, 204) while the value of p 4 largely
the value of n does not change from
in this case, too, Roughton
has shown
parts of the oxygen equilibrium
curve
by the same author, made at con-