Reprinted from Ciba Foundation Symposium on Significant Trends in
Medical Research, 1959, pp. 3-10

 

MOLECULAR STRUCTURE IN RELATION TO
BIOLOGY AND MEDICINE

L. PAULING

California Institute of Technology, Pasadena

Tue molecules that compose the body of a human being may
be conveniently divided into two classes: small molecules and
large molecules. Small molecules are molecules containing 10
or 20 or perhaps 100 atoms; examples are glucose, acetylcholine,
glycine and other amino acids, thiamine and other vitamins.
Large molecules are molecules containing hundreds or thousands
or tens of thousands of atoms; examples are the proteins and the
nucleic acids. To understand the human body in health and in
disease we need to know the structure of the small molecules and
the large molecules. The large molecules are especially important
because it is they that carry biological specificity: their structural
differences determine the differences between species of living
organisms and also the differences between individuals of the
same species.

During the past 40 years great progress has been made in the
determination of the precise structures of small molecules. The
work has been done by the use of several methods, of which the
study of gas molecules by electron diffraction and of crystals by
X-ray diffraction are the most important; much information is
also now being provided by microwave spectroscopy and
nuclear-spin and electron-spin magnetic resonance. Outstanding
among the X-ray studies are the astounding achievements of
Mrs. Dorothy Hodgkin and her co-workers, in determining the
structures of penicillin and vitamin By.

During the recent decades there has also been developed a

3
4 L. PAULING

powerful theory of atomic and molecular structure. This theory
and the detailed experimental information about the structure of
simple molecules and crystals have permitted chemical structure
theory to be greatly extended and refined. I think that it now
encompasses all of the significant structural principles necessary
for the understanding of large molecules as well as small ones, and
that no important new structural features will be discovered in the
course of the structure determinations of proteins and nucleic
acids that we expect to be made during the next decade or two.

Moreover, I think that the nature of the forces responsible for
intermolecular interactions is now well understood, and that
these forces—London dispersion force of van der Waals attrac-
tion, the force of van der Waals repulsion, the formation of
hydrogen bonds, the interaction of elcctrically charged groups,
the formation of fractional covalent bonds—can be confidently
analysed as the basis of the characteristic intermolecular inter-
actions leading to biological specificity, such as the interactions of
antibodies and antigens.

I believe that biological specificity in general results from a
detailed complementariness in structure of interacting molecules.
There exists an overwhelming mass of evidence that the specific
combining power of an antibody molecule for its homologous
antigen molecule results from a complementariness in structure
that permits the co-operation of several weak interactions that
separately would not produce a significant bond between the
molecules. This evidence has been provided by the work of
Landsteiner with antigens containing haptenic groups with
known chemical structure and by later studies along the same lines
(Campbell, Pressman, Haurowitz). The combining powers of
antibody molecules with haptens related in structure to the
haptenic group of the immunizing antigen are found to change in
the ways predicted for the interaction energy of the haptens with
an antibody-combining region closely complementary to the
original haptenic group. The replacement of an atom or radical
MOLECULAR STRUCTURE 5

by a larger atom or radical causes a decrease in combining power,
attributed to van der Waals repulsion (steric hindrance); replace-
ment of a radical by a smaller one or by one of equal size with
decreased polarizability (decreased power of van der Waals
attraction), decreased electric charge, or decreased power of
hydrogen-bond formation causes a decrease in combining power;
replacement by one of approximately equal size and polariza-
bility (such as a methyl group by a chlorine atom), even with much
different chemical properties, results in no change in combining
power.

The idea that the antigen or a fragment of it serves as the
template against or about which a plastic material, the precursor
of the antibody, is moulded through the operation of the forces
of intermolecular attraction is an attractive one. We can under-
stand the process of hardening of the antibody molecule in its
complementary configuration through the formation of hydro-
gen bonds and the operation of other forces between the different
parts of the folded polypeptide chain of the molecule. This
postulated mechanism of formation of antibodies provides an
explanation of many observations, such as the astounding versa-
tility of the antibody-producing mechanism—the ability of an
animal to manufacture specific antibodies against haptenic groups,
such as the p-azobenzenearsonate ion, that probably have never
constituted a part of the environment of the forebears of the
animal. However, much remains to be discovered about the
mechanisms of manufacture of proteins, and it may be found that
these mechanisms are complex ones, involving a succession of
steps. From the analysis of possible modes of operation of inter-
atomic and intermolecular forces I have reached the conclusion
that every step involving specificity will be found to depend for
its specificity on a detailed complementariness in structure of the
interacting molecules.

The importance of even the smallest structural details of the
large molecules in the human body can be illustrated by the
6 L. PAULING

discussion of the abnormal haemoglobins in relation to the here-
ditary haemoglobinaemias. It is now ten years since sickle cell
anaemia was recognized as a molecular disease and the abnormal
molecule responsible for it, haemoglobin S, was discovered
(Pauling, Itano, Singer and Wells, 1949). During this decade
many other abnormal forms of human haemoglobin have been dis-
covered and many new diseases for which they are responsible have
been described (see Pauling, 1955; Itano, 1956). The mechanism of
the process of change in shape (sickling) of the erythrocytes of
sickle cell anaemia patients has been recognized to be the formation
of spindle-shaped tactoids (liquid crystals of the nematic type) of
unoxygenated haemoglobin S. The formation of these crystals
can be attributed to a self-complementariness in structure of the
molecules of this protein. The self-complementariness is destroyed
when the crystals combine with oxygen, and it is not shown by
normal adult human haemoglobin (haemoglobin A).

A significant start has now been made on the determination of
the difference in structure of haemoglobin S and haemoglobin A.
Shortly after the discovery of haemoglobin S, Schroeder, Kay
and Wells (1950) found that its amino acid composition is nearly
the same as that of haemoglobin A. It was then found by
Ingram (1958), by application of his powerful method of
two-dimensional paper electrophoresis-chromatography to the
enzyme-catalysed haemoglobin hydrolysates, that the difference
in amino acid composition and sequence consists only in the
replacement in each half-molecule of a glutamyl residue (in
haemoglobin A) by a valyl residue (in haemoglobin S$). (The
haemoglobin molecule is shown to have a twofold symmetry
axis, and hence to consist of two identical halves, by the X-ray
diffraction pattern of the crystal.) There are about 600 amino
acid residues in the haemoglobin molecule, and only two of the
600 are different in haemoglobin A and haemoglobin S; yet this
small difference in structure is enough to cause the human beings
who manufacture haemoglobin S to have a serious disease.
MOLECULAR STRUCTURE 7

Something is now known about the location of the glutamyl-
valyl replacement in the polypeptide chains. It was shown by
Rhinesmith, Schroeder and Pauling (1957), by the use of Sanger’s
end-group method, that haemoglobin A contains two polypep-
tide chains of one kind (« chains, with N-terminal sequence val-
leu) and two of a second kind (8 chains, with N-terminal sequence
val-his-leu). It has now been shown by Vinograd, Hutchinson
and Schroeder (1959) that the glutamyl-valyl replacement occurs
in the 8 chains; the « chains of haemoglobin A and haemoglobin
S seem to be identical.

The method used by these investigators is an interesting one.
Haemoglobin A labelled with ##C was made by incubating human
reticulocytes in blood to which t-leucine containing 1C had been
added. A solution containing labelled haemoglobin A and un-
labelled haemoglobin S was brought to pH 5 for some hours—at
this pH the « and B chains separate. The solution was brought
back to pH 7, permitting the chains to recombine, and the two
haemoglobins were separated by column chromatography. The
N-terminal residues of the hybridized haemoglobin S were
labelled with Sanger’s reagent, the protein was partially hydro-
lysed, and the N-terminal peptides were isolated chromato-
graphically and checked for radioactivity. The peptide DNP-
val-leu, characteristic of « chains, was found to be strongly
labelled with “C, and the peptide di-DNP-val-his-leu, character-
istic of 8 chains, only weakly labelled. Hence, it is the 8 chains
that are different in haemoglobin A and haemoglobin S.

Ingram (1959) has found the peptide val-his-leu-thr-pro-glu-
glu-lys from haemoglobin A and val-his-leu-thr-pro-val-glu-lys
from haemoglobin S, and has surmised (because the first three
residues are the N-terminal set for the 8 chains) that the glutamyl-
valyl replacement occurs in the sixth position from the N-
terminus of the 8 chains. In haemoglobin C the same position is
occupied by lysyl. He has also reported that for haemoglobins
Dg and E the abnormalities are in the @ chains and for D, and 1
8 L. PAULING

they are in the « chains. Schwartz and co-workers (1957) have
reported from studies of inheritance of S and G in families carrying
both traits that different genes control the manufacture of haemo-
globin S$ and haemoglobin G, and a similar conclusion about
haemoglobin S and haemoglobin Hopkins-2 has been reached by
Smith and Torbet (1958). It is possible that the abnormalities for
haemoglobins G and Hopkins-2 are in the « chains, and that the
synthesis of « chains and that of @ chains are controlled by dif-
ferent genes.

Jones, Schroeder and Vinograd (1959) and Hunt (1959) have
obtained evidence that human foetal haemoglobin, haemoglobin
F, contains two « chains that are identical with those of haemo-
globin A. Hence, an «-gene abnormality would be expected to
cause the manufacture of two abnormal proteins, an abnormal
foetal haemoglobin and an abnormal adult haemoglobin. This
prediction has not yet been verified by observation.

It has been found by Jones and co-workers (1959) that haemo-
globin H represents a new sort of molecular abnormality. Haemo-
globin H was first reported in two children of Chinese descent;
the investigators, Rigas, Koler and Osgood (1956), found no
haemoglobin H in the red cells of the parents, whereas for other
abnormal haemoglobins the trait is generally shown by one or
both of the parents. The haemoglobin-H molecule was found by
Jones and his co-workers to have four polypeptide chains with
the same N-terminal sequence, val-his-leu. This observation
suggested that the molecule consists of four normal @ chains, and
this hypothesis was then verified by a hybridization experiment
with haemoglobin A labelled with *C and unlabelled haemo-
globin H. In the course of this work a new haemoglobin, con-
sisting of four sickle-cell 8 chains, was made; it is likely that at
some future time this haemoglobin too will be found in Nature.
The genetic abnormality that leads to the manufacture of haemo-
globin H is apparently one that inhibits the synthesis of the «
chains. ,
MOLECULAR STRUCTURE 9

Thus, significant progress has been made in the study of the
chemical structure of the abnormal haemoglobins, and yet we
are still far from understanding the properties of the substances in
terms of the structures of their molecules. It is almost certain that
this understanding would not be achieved even though complete
determinations were to be made of the amino acid sequences in
the polypeptide chains of normal adult human haemoglobin and
the various abnormal haemoglobins. Complete sequence studies
made for insulin by Sanger and his collaborators have not led to
an understanding of the physiological properties of this hormone.
What is lacking as yet is knowledge about the detailed method of
folding of the polypeptide chains and the configuration of the
side chains; what is needed is the determination of the complete
molecular structure of these proteins, and also of the proteins and
other substances with which they interact. Despite the vigorous
efforts of many investigators (among them are Perutz, Kendrew,
Mrs. Hodgkin, Corey, Harker, and Bernal, and their collabora-
tors) there has not yet been carried out the complete structure
determination of any protein molecule. I estimate rst March
1967+ 2°5 years as the date when the announcement will be
made that the first complete structure determination for a protein
molecule, the determination by experiment (X-ray diffraction)
of the relative positions in space of all of the atoms in the molecule,
has been accomplished. The determination of the complete
structure of a molecule of deoxyribonucleic acid will probably
occur a few years later. Twenty-five years from now we shall
probably know the complete structures of one hundred protein
molecules and a few nucleic acid molecules. We shall then have
a detailed understanding of the ways in which a few enzymes
carry out their specific activities, the ways in which genes dupli-
cate themselves and accomplish their individual tasks of precisely
controlling the synthesis of protein molecules with well defined
structures, the ways in which abnormal molecules give rise to the
manifestations of the diseases that they cause, the ways in which
10 L. PAULING

drugs and other physiologically active substances achieve their
effects. When this time comes, medicine will have made a
significant start in its transformation from macroscopic and
cellular medicine to molecular medicine.

REFERENCES

Hunt, J. A. (1959). Nature (Lond.), 183, 1373.

IncraM, V. M. (1958). Biochim. biophys. Acta, 28, 539.

Incram, V. M. (1959). Nature (Lond.), 183, 1795.

Irano, H. A. (1956). Ann. Rev. Biochem., 25, 331.

Jones, R. T., ScHRogpER, W. A., Batoc, J. E., and VinoerapD, J. R. (1959),
J. Amer. chem. Soc., 81, 3161.

Jones, R. T., Scuroxper, W. A., and Vinocrap, J. R. (1959). J. Amer. chem.
Soc., 81, in press.

PauLine, L. (1955). Harvey Lect., 41, 216.

Pautine, L., Irano, H. A., Stncer, S.J., and We ts, I. C. (1949). Science, 110,
543

Rannesmitu, H. S., ScaroEpER, W. A., and Pautine, L. (1957). J. Amer. chem.
Soc., 79; 609.

Ricas, D. A., Kotsr, R. D., and Oscoon, E. E. (1956). J. Lab. clin. Med., 47,
SI.

ScHroeper, W. A., Kay, L. M., and Wetts, I. C. (1950). J. biol. Chem., 187,
221.

Scuwartz, H. C., Spaet, T. H., Zuerzer, W. W., Nzet, J. V., Rosrnson,
A. R., and Kaurman, S. F. (1957). Blood, 12, 238.

Smitu, E. W., and Torser, J. V. (1958). Bull. Johns Hopk. Hosp., 102, 38.

VinoeraD, J. R., Hurcuinson, W. D., and Scuroeper, W. A. (1959). J.
Amer. chem. Soc., 81, 3168.

SB