The Molecular Basis of Biological Specificity

by Linus Pauling
Institute of Orthomolecular Medicine
2700 Sand Hill Road, Menlo Park, California 94025

       During the decade 1930-1940 I formulated a general theory of the

molecular basis of biological specificity, involving the idea that biological

specificity results from the interaction of complementary molecular structures,

with hydrogen bonds among the most important of the weak intermolecular

forces between the interacting molecules.  The most striking example of

specific biological interactions of this sort is the interaction between the two

complementary strands of the DNA molecule in the double helix discovered by

Watson and Crick twenty-one years ago.

       My early work was on the determination of the structure of crystals

by the x-ray diffraction technique, the determination of the structure of gas

molecules by electron diffraction, and the application of quantum mechanics to

physical and chemical problems, especially the structure of molecules and the

nature of the chemical bond.  In 1929, when Thomas Hunt Morgan came to the

California Institute of Technology, bringing with him a number of very able younger

biologists, I began to become familiar with biological problems, and to think

about possible ways in which biological specificity could be explained in terms

of interactions between molecules.  I worked on several problems of biological

specificity, from the molecular point of view, without success; one of them was

the problem of explaining the self-sterility of the marine organism Ciona (the


2.

sea squirt), which was being studied by Morgan.  In 1934 the problem of the

shape of the oxygen equilibrium curve of hemoglobin attracted my attention.

Consideration of the structure of hemoglobin led to the idea that investigation

of the magnetic properties of this substance and its derivatives would provide

valuable information, and work along these lines , in collaboration with C. D,

Coryell and a number of students, was initiated. Alfred E. Mirsky of the

Rockefeller Institute for Medical Research, who had been studying hemoglobin

for several yearsp came to Pasadena for a year , and he and I formulated a

theory of the structure of native, denatured, and -coagulated proteins, based

upon the concept that a native protein molecule consists of one polypeptide

chain (or of two or more such chains) folded into a uniquely defined configuration,

in which it is held by hydrogen bonds between the peptide nitrogen and oxygen

atoms, as well as by other Wak forces , with denaturation involving a loss of

this well-defined structure. .l

In 1936, while I was on a short visit to the Rockefeller Institute

for Medical Research, Carl Landsteiner asked me how I would explain the

observed properties of antibodies and antigens by means of their molecular

structure.  I thought about this problem during the following years, and consulted

with Landsteiner about the interpretation of sometimes conflicting experimental

results.  By 1940 I had formulated a theory of the structure and process of

formation of antibodies. 2 This theory was based upon the concept that the

specific combining region of an antibody molecule is complementary in structure to

a portion of the surface of the antigen, with the antigen-antibody bond resulting


3.

from the cooperation of weak forces (electronic Van der Waals forces, electrostatic

interaction of charged groups, and hydrogen bonding) between the complementary

structures, over an area sufficiently large that the total binding energy could

resist the disrupting influence of thermal agitation.  Precipitating and, agglutinating

antibodies were assumed to be bivalent, consisting of a central part, with structure

common to all or almost all antibodies produced by the animal, and two end parts,

the combining regions, with structure complementary to that of the antigen.

(The idea of complementary structures for antibody and antigen was suggested

by Breinl and Haurowitz, 3 Alexander4, and Mudd. 5  There is some intimation

of it in the early work of Ehrlich and Bordet. ) The complementary combining

regions were assumed to be formed by the folding of polypeptide chains in the

presence of the antigen, in such a way that the forces of attraction would mold

the folding chain into a structure complementary to that of a portion of the antigen,

with the folded chain then being held in this configuration by hydrogen bonds and

other interactions, even after the antibody has dissociated from the antigen

on which the combining group was molded.  Dan Campbell, David Pressman,

and a number of other workers in our laboratory carried out experimental studies

that verified the valence 2 for precipitating and agglutinating antibodies 67 and

that left no doubt that the combining regions of antibodies are complementary in

structure to the homologous haptenic groups of the antigens. a  The fit of the

combining region of the antigen to the hapten was shown to be close, better than

20 pm in some cases a and the effects of Van der Waals attraction, electrostatic

forces@ and hydrogen-bond formation were separately verified in quantitative


4.

hapten-inhibition studies.  A satisfactory theoretical explanation of quantitative

values of free energy of combination of haptens with antibodies homologous to the

0-, m-, and p-azobenzene arsenic acid groups on the basis of known intermolecular

`interactions was reported in 1945. 9 For several haptens with various groups

substituted in the positions of the azo group in the hapten of the immunizing antigen

the standard free energy of combination, as given by hapten inhibition constants,

was found to be proportional to the calculated Van der Waals interaction with the

surrounding antibody, which includes proportionality to the electric polarizability

of the group.  For groups forming hydrogen bonds the energy of the hydrogen

bond (1. 5 to 3 kJ mole -1
                            , representing the difference in energy of the hydrogen

bond formed by the hapten with antibody and with water) was needed, in addition

to the term corresponding to electronic Van der Waals interaction.  The effect
                                                               .

.

of electric charge was determined by comparison of haptens closely similar in

shape, but with a difference in electric charge: In one caselo comparison of

haptens with either trimethylammonium ion or tertiary butyl group, and in the

other case with either carboxylate ion or nitro group.  In each comparison

there was indication of a complementary electric charge in the antibody, close to

the charge in the immunizing antigen.  The magnitude of the effect showed the

charge in the antibody to be within 320 pm (first case) or 260 pm (second case)

of the minimum distance permitted by the Van der Waals radii of the groups.

I think that this work, which was based on earlier work by Landsteiner and his

collaborators, 12 leaves no doubt that the specificity of antibodies is the result of
                                                            .

the complementariness in structure of the combining group and a portion of the

surface of the homologous antigen.


5.

       It became evident that non-biological specificity could also be

explained in terms of complementariness.  I gave an example in a lecture on

analogies between antibodies and simpler chemical substances: 13  "The

reaction shown by simple chemical substances that is analogous to that of

specific combination of antigen and antibody is the formation of a crystal of a

substance from solution. A crystal of a molecular substance is stable because

all of the molecules pile themselves into such a configuration that each molecule

is surrounded as closely as possible by other molecules - that is, if a molecule

were to be removed from the interior of a crystal, the cavity that it would leave

would have very nearly the shape of the molecule itself.  We can say that the

part of a crystal other than a given molecule is very*closely complementary to

that molecule.  Other molecules, with different shape and structure, would not

fit into this cavity nearly so well, and in consequence other molecules in general

would not be incorporated in a growing crystal.  This is the explanation of the

astounding chemical process of purification by crystallization - from a very

complicated system, such as, for example, grape jelly, containing hundreds of

different kinds of molecules, crystals which are nearly chemically pure may be

formed, such as crystals of cream of tartar, potassium hydrogen tartrate. "

      In the same paper it is stated that "although crystallization is the

only simple chemical reaction that shows striking similarity to serological

reactions with respect to specificity, there are many physiological phenomena

that are similarly specific, and for which the specificity can be given a similar

explanation,  The specificity of the catalytic activity of enzymes is due to a

surface configuration of the enzyme such as to make the enzyme complementary

to the substrate molecule,, ora rather, to the substrate molecule in the strained


6.

state that occurs during the catalyzed reaction.  The specific action of drugs

and bactericidal substances have a similar explanation.  Even the senses of

taste and odor are based upon molecular configuration rather than upon ordinary

chemical properties - a molecule which has the same shape as a camphor

molecule will smell like camphor even though it may be quite unrelated to

camphor chemically,  I am convinced that it will be found in the futures as our

understanding of physiological phenomena becomes deeper, that the shapes and

sizes of molecules are of just as great significance in determining their

physiological behavior as are their internal structure and ordinary chemical

properties. "

In 1940 Max Delbriick and I 14 published a discussion of the

intermolecular forces operative in biological processes.  P. Jordan had advanced

the idea that there exists a quantum-mechanical stabilizing interaction that

operates preferentially between identical or nearly identical molecules or parts

of molecules, and is of great importance for biological processes, including the

production of new genes identical with the old ones.  Delbriick and I pointed out

that the specific quantum-mechanical forces between identical molecules could not be
large enough to cause a specific attraction between like molecules under the
conditions of excitation and perturbation prevailing in living organisms, and

therefore could not be effective in bringing about autocatalytic reactions.  We

wrote that "It is our opinion that the processes of synthesis and folding of highly

complex molecules in the living cell involves in addition to covalent-bond

formation, only the intermolecular interactions of Van der Waals attraction and

repulsion, electrostatic interactions, hydrogen-bond formation, etc. , which are

now rather well understood,  These interactions are such as to give stability to

a system of two molecules with complementary structures in juxtaposition, rather


7.

than of two molecules with necessarily identical structures; we accordingly feel

that complementariness should be given primary consideration in the discussion

of specific attraction between molecules and the enzymatic synthesis of moleeules. "

We mentioned that "The case might occur in which the two complementary

structures happened to be identical; however , in this case also the stability of

the complex of two molecules would be due to their complementariness rather

than their identity. " Some time later 15 I discussed the matter of gene replication

in more detail: "I believe that the genes serve as the templates on which are

molded the enzymes that are responsible for the chemical characters of the

organisms, and that they also serve as templates for the production of replicas

of themselves.  The detailed mechanism by means of which a gene or a virus

molecule produces replicas of itself .is not yet known.  In general the use of a

gene or virus as a template would lead to the formation of a molecule not with

identical structure but with complementary structure.  It might happen, of course,

that a molecule could be at the same time identical with and complementary to the

template on which it is molded.  However, this case seems to me to be too unlikely

to be valid in general, except in the following way.  If the structure that serves as

a template (the gene or virus molecule)consists of, say, two parts, which are

themselves complementary in structure, then each of these parts can serve as the

mold for the production of a replica of the other part, and the complex of two

complementary parts thus can serve as the mold for the production of duplicates

of itself. " The same statements were made in the spring.of 1948 in lectures

in Oxford, Cambridge, London, and elsewhere.

       The hydrogen bond was recognized by Latimer and Rodebush as an


8.

important structural feature over fifty years ago, 16 In their 1920 paper

they mentioned that "Mr. Huggins of this laboratory in some work as yet

unpublished has used the idea of a hydrogen kernel

a theory in regard to certain organic compounds. "

pointed out the importance of the hydrogen bond in

             1

held between two atoms as

Mirsky and Pauling in 1936

the determining of the structure

of proteins, A In the same year Huggins also discussed protein structures in

a more detailed way, with hydrogen bonds between the NH and CO groups of the

main chains. 17 A few years later Huggins described several helical structures

for polypeptide chains, with intrachain hydrogen bonds. 18  These structures were

needlessly restricted to having an integral number of amino-acid residues per turn

of the chain, and, moreover, Huggins did not require the amide groups to be

planar, although the planarity of these groups had been recognized since 1932, 19

and had already been verified by several determinations of the structure of

simple peptide crystals in our laboratory. It is unfortunate that Huggins was

handicapped by these two erroneous assumptions in his imaginative and otherwise

sound attack on the problem of the secondary structure of proteins.  The same

two erroneous assumptions provided a similar insuperable barrier to the vigorous

attack made by Bragg, Kendrew, and Perutz on the same problem. 20 In the

meantime,, Corey and other investigators in Pasadena had determined the crystal

structures of a number of amino acids and simple peptides, and Pauling and Corey

had discovered the alpha helix and the parallel-chain and antiparallel-chain

pleated sheets. 21 The discovery of the alpha helix left no doubt about the importance

of helical structures and of hydrogen bonds in determining the secondary structures

of proteins.


9.

         I had been interested in the nucleic acids since 1933, when

Sherman and I calculated the resonance energy of guanine and other purines. 22


My colleague       Robert B. Corey had made some x-ray

diffraction photographs of fibers of nucleic acid, which were, however; of

somewhat poorer quality than those published by Astbury and Bell. 23 I began

work on the problem of interpreting the x-ray photographs on 26 November 1952;

on the preceding day I had attended a seminar in biology in the California Institute

of Technology, at which Professor Robley Williams of University of California,

Berkeley, showed a slide of an electron microscope photograph of molecules of


sodium ribonucleate.  He said that the small fibrils had a diameter of about

1. 5 nm, and that they were apparently cylindrical, in that only one diameter was

shown.  The x-ray photographs indicated an identity distance along the axis of

the molecule of 340 pm, and, with the measured density of ribonucleic acid,

about 1.62 g cm -3 s it was indicated that the fibers contain two or three molecules,

probably helices twisted about one another.  The value of the spacing of the

principal equatorial x-ray reflection had been shown to decrease with decreasing

amount of hydration of the fibers, with minimum value 1.62 nm.  I assumed this

value to correspond to essentially anhydrous nucleic acid, and, with use of the

density, I calculated the number of polynucleotide chains per unit to be exactly

three.  This result surprised me, because I had expected the value 2 if the nucleic

acid fibers really represented genes.  I decided, however, that probably the

fibers were artifacts, produced by the process of extraction from cells and the

subsequent stretching.  During the next month I strove to find a way of arranging

the polynucleotide chains in a triple helix, and was successful, although the


10.

structure was described as "an extraordinarily tight one, with little opportunity

for change in positions of the atoms".  The paper in which this structure was

described was communicated to the Proceedings of the National Academy of Sciences

on 31 December 1952, and a copy of the manuscript was sent to Watson and Crick. 24


       In hindsight, it is evident that I made a mistake on 26 November 1952

in having decided to study the triple helix rather than the double helix.  It is likely
that the fibers giving the'equatorial spacing 1.62 nm contained some water, and

also had density smaller than 1.62 g cm -3
                                                * The diameter 1.5 nm observed by
Williams for nucleic acid molecules correspondss with density assumed 1. 6 g cm -3

and unit translation 340 pm along the molecular axis, to two molecules in a

helical structure (calculated diameter 1.6 nm) rather than to three (1. 9 nm).

I am now astonished that I began work on the triple-helix structure, rather than

on the double helix. I had not forgotten that Delbriick and I had suggested that the
gene might consist of two complementary molecules, but for some reason, not
clear to me now, the triple-chain structure apparently appealed to me, possibly
because the assumption of a three-fold axis simplified the search for an acceptable

structure.

       I cannot say what would have happened if I had made the other

assumption, that of a double helix, on 26 November 1952, or if I had succeeded

in getting access to the diffraction photographs of DNA that had been made by

Wilkins. There is a chance that I would have thought of the Watson-Crick structure

during the next few weeks. I knew that the purines and pyrimidines were present

in nucleic acid in equal amounts, but I bad not drawn the reasonable conclusion


11.

about purine-pyrimidine pairs.  I knew about hydrogen bonding by purines and

pyrimidines.  Nevertheless, I myself think that the chance is rather small that

I would have thought of the double helix in 1952, before Watson and Crick made

their great discovery. After all, I had spent part of the summer of 1937 in a

search for ways of folding polypeptide chains , with planar amide groups of the

correct dimensions and with hydrogen bonds between the'C0 and NH groups of

residues separated by some distance along the chain, in such a way as to account

for the x-ray diffraction photographs of alpha keratin, but without success.

There was no reason why the alpha helix should not have been discovered then,

rather than eleven years later, when it was discovered after a few hours of work.

There is no doubt that even rather simple ideas sometimes are very elusive.

       It is my opinion that if Watson and Crick had not carried on their

persistent effort, and had not had the benefit of advice about the structures of

the nitrogen bases and hydrogen bonds from Jerry Donohue and information from

the excellent x-ray diffraction photographs of Wilkins, the discovery of the

double helix, which has led to such great developments in molecular biology,

might `well have been delayed for several years.


12.

References

1.


2.


3.


4.


5.


6.

7.

8.

9.


10.





11.





12.





13.


14.


15.

Mirsky, A. E., and Pauling, L. , Proc. U. S. Nat.Acad. Sci. 22, 439 (1936).

Pauling, L. , J.Amer. Chem. Sot. 62, 2643 (1940).

Breinl, F., and Haurowitz, F., Z. physiol. Chem. 192, 45 (1930).

Alexander, J., Protoplasma E, 296 (1931). I

Mudd, S. , J. Immunol. 23# 423 (1932).

Pauling, L., Pressman, D., Campbell, D. H., Ikada, C., and Ikawa, M. B

J. Amer. Chem. Sot. 6& 2994 (1942).

Pauling, L., Pressman, D., and Campbell, D. H., J.Amer. Chem. Sot. 66,

330 (1944).

Pauling, L., Pressman, D., and Grossberg, A. L., J.Amer. Chem. Sot. 66,
784 (1944).

Pauling, L., and Pressman, D., J.Amer. Chem. Sot. 67, 1003 (1945).

Pressman, D., Grossberg, A. L., Pence, L. H., and Paul&g, L.,

J.Amer. Chem. Sot. 62, 250 (1946).
Pressman, D., Swingle, S. S., Grossberg, A. L., and Pauling, L.,

J. Amer. Chem. Sot. 66, 173 1 (1944).

Landsteiner,

Press9 1945.

Pauling, L. s

Pauling, L. s

Pauling, L.,

K o ? The Specificity of Serological Reactions, Harvard University

Chem. Eng. News 22, 1064 (1946).

and Delbrkck, M. o Science 92, 77 (1940),

Molecular Architecture and the Processes of Life

(21st Sir Jesse Boot Foundation lecture# 28 May 1948), published by the

Sir Jesse Boot Foundation, Nottingham, 1948.


13.

16. Latimer, W. M. , and Rodebush, W. H., J.Amer. Chem. Sot. 42, 1419 (1920).

17. Huggins, M. L., J. Org. Chem. A, 407 (1936).

18. Huggins, M. L., Chem. Revs. 32, 195 (1943).

19. Pauling, L., Proc. U. S.Nat.Acad.Sci. 12, 293 (1932).
20. Bragg, W. L., Kendrew# J., and Perutz, M., Proc. Roy. Sot. London

3, A-1074 (1950).

21. Pauling, L., and Corey, R.B., J.Amer. Chem. Sot. 7,2, 5349

22. Pauling, L., and Sherman, J., J. Chem. Phys. i# 606 (1933).
23. Astbury, W. T. p and Bell, F. 0. o Nature l$, 747 (1938).

24.  Pauling, L. p and Corey* R. B. o Proc. U. S. Nat.Acad. Sci. 39,

(1950).

84 (1953).