The Polypeptide-Chain Configuration in Hemoglobin and Other Globular Proteins
The citation for the published article is: Pauling, Linus, and Robert B. Corey. "The Polypeptide-Chain Configuration in
Hemoglobin and Other Globular Proteins." Proceedings of the National Academy of Sciences 37, no. 5 (May 15, 1951): 282-285.
Item is handwritten.
Number of Image Pages:
13 (835,346 Bytes)
1951-03 (March 1951)
Corey, Robert B.
Original Repository: Oregon State University. Library. Ava Helen and Linus Pauling Papers
Reproduced with permission of the Ava Helen and Linus Pauling Papers. Oregon State University Library.
Medical Subject Headings (MeSH):
The Search for the Molecular Helix
Two Nobel Prizes
[ . . . ] would like to have this MS as soon as possible.
The Polypeptide-chain Configuration in Hemoglobin and other Globular Proteins
By Linus Pauling and Robert B. Corey
Gates and Crellin Laboratories of Chemistry, California Institute of Technology*
Communicated March --, 1954
In the immediately preceding papers we have described several hydrogen-bonded planar-amide configurations of polypeptide chains,
and have discussed the evidence bearing on the question of their presence in fibrous proteins. It seems worth while to consider
the possibility that these configurations -- the pleated sheet, the 3.7-residue alpha helix, the 5.1-residue gamma helix,
and the three-chain collagen helix -- are represented in
[END PAGE ONE]
[BEGIN PAGE TWO]
molecules of the globular proteins.
It may first be noted that many globular proteins, such as ovalbumin, can on denaturation be converted into a form showing
the beta-keratin x-ray pattern^1. The fiber-axis residue distance that is observed, about 3.3 A, is the same as for beta
keratin, for which we have suggested the pleated-sheet configuration^2, and it seems reasonable that the same structure should
be represented by these denatured proteins. It is, of course, to be expected that a layer structure, such as the pleated
sheet, would be assumed by a protein when pressed flat, and the extension of the chains in the
[END PAGE TWO]
[BEGIN PAGE THREE]
pleated-sheet structure makes it reasonable that such a structure should also be assumed by a protein when drawn into a fiber.
The most significant published data bearing on the structure of globular proteins are those on horse carbonmonoxyhemoglobin
that have been obtained through the well-planned and diligent efforts of Perutz and his co-workers^(3,4). These data have
been published mainly as a set of sections of a three-dimensional Patterson diagram. We have observed that the data provide
some support for the idea that the 3.7-residue helix is a principle feature of the structure of this protein.
Perutz has pointed out that his data indicate that the hemoglobin
[END PAGE THREE]
[BEGIN PAGE FOUR]
molecule is about 57A long, and between 34A and 57A in other dimensions, and that there are present rods extending in the
57A direction, and packed in a pseudohexagonal array, with the centers of the rods about 10.5A apart. He concluded that the
rods probably have the same structure as the molecules in a keratin, for which we have recently suggested the 3.7-residue
There are several facts that favor the view that the 3.7-residue helix is represented in hemoglobin. First, there is the
similarity to keratin, pointed out by Perutz, and the evidence supporting the 3.7-residue helical configuration
[END PAGE FOUR]
[BEGIN PAGE FIVE]
for the fibrous proteins with the alpha-keratin structure^5. Closely related is the fact that from the density and the average
residue weight for hemoglobin one would predict that molecules with this helical configuration would be spaced about 11A apart
(from center to center) in agreement with Perutz's conclusion that the rods in hemoglobin are about 10.5A apart. (A calculation
of this sort at once eliminates the 5.1-residue helix, for which the predicted average spacing of the rods is 14A.)
Another bit of supporting evidence is provided by the integrated vector density in a strip of the xz Patterson section through
the origin of the 3-dimensional diagram and in the direction of the axes of the rods. Bragg,
[END PAGE FIVE]
[BEGIN PAGE SIX]
Kendrew, and Perutz^6 have reproduced this quantity, plotted as a function of the distance from the origin, in connection
with their painstaking analysis of the data for hemoglobin and also for myoglobin^7 and discussion of the correlation of the
data with alternative polypeptide configurations. The function has peaks at almost 5A, 11.5A, 16.5A, 21.5A, 27A, 32A, etc.
We have evaluated a corresponding function for the 3.7-residue helix, by including interatomic vectors deviating by not more
than 2A from the direction of the helical axis, and weighting the vectors proportionately to the product of the atomic number
of the two atoms. The function obtained in this way for an 18-residue 5-turn helix with fiber-axis residue length 1.53A has
maxima at 5.1A, 10.6A, 16.7A, 21.4A, 27.5A, 32.6A, etc., in excellent agreement with the experimental points.
[END PAGE SIX]
[BEGIN PAGE SEVEN]
Another test of the proposed configuration can be made by comparison of the calculated and observed radial distribution functions.
Perutz pointed out that the Patterson diagram shows a strong shell at about 5A from the origin. We have obtained a radial
distribution function corresponding to his data for hemoglobin by numerical integration over the controlled Patterson sections
published in his paper; this function is shown in Figures 1 and 2. It is seen that it has a maximum at about 4.8A. The calculated
radial distribution functions for the 5.1-residue helix are also shown in Figure 1. The three curves represent respectively
the function for the four main -- chain atoms C, C^1, O, and N only,
[END PAGE SEVEN]
[BEGIN PAGE EIGHT]
the function for the four main-chain atoms and a beta-carbon atom in one of the two alternative positions, and the function
for the four main-chain atoms and a beta-carbon atom in the other positions. It is seen that there is no agreement with the
hemoglobin curve. The same three calculated radial distribution functions for the 3.7-residue helix are given in Figure 2.
We think that the rough agreement with the hemoglobin curve is to be considered as significant; it is to be remembered that
even with inclusion of the beta-carbon atom only about 60 percent of the heavy atoms in the molecule have been taken into
consideration in the calculation. The neglected side-chain atoms are, of course, far more randomly arranged than the main-chain
atoms of the
[END PAGE EIGHT]
[BEGIN PAGE NINE]
helix, and would for this reason tend to distribute their vectors rather uniformly, and thus not to mask the characteristic
features of the function due to the main-chain and beta-carbon atoms.
The comparison of radial distribution functions may thus be centered as giving additional evidence in favor of the suggestion
that the rods that Perutz has reported to be present in the hemoglobin molecule have the 3.7-residue helical configuration.
We think that it is not unlikely that this polypeptide configuration is represented in other globular proteins also. In particular,
its presence in myoglobin, which is closely related to hemoglobin, would not be surprising; however, it must be pointed out
that the Patterson projection for myoglobin on
[END PAGE NINE]
[BEGIN PAGE TEN]
a plane perpendicular to the axis of the rods, given by Bragg, Kendrew, and Perutz^6, seems hardly to be compatible with this
structure. It is possible, of course, that side-chain atoms happen to cooperate effectively in changing the aspect of this
projection, or that the axes of the rods do not lie exactly along the direction of projection. The evidence favoring the
3.7-residue helix for myoglobin is contained in Kendrew's description of the myoglobin molecule, as deduced from his data,
as consisting of a layer of four rods about 9.5A apart and with vector maxima spaced 5A apart in the direction of the axes
of the rods. The layers themselves are about 15A apart, which suggests that if the
[END PAGE TEN]
[BEGIN PAGE ELEVEN]
structure does involve the 3.7-residue helix the side chains are distributed as in crystalline muscle^5, in which the molecules
have an effectively elliptical cross-section, with major and minor diameters 13.1A and 9.8A, respectively.
This investigation was aided by grants from The Rockefeller Foundation, The National Foundation for Infantile Paralysis, and
the United States Public Health Service. We acknowledge with gratitude the assistance and encouragement of our colleagues
in the Gates and Crellin Laboratories of Chemistry throughout the period during which the studies reported in this series
of papers and also the investigations on which this work is based were made. We are
[END PAGE ELEVEN]
[BEGIN PAGE TWELVE]
especially grateful to Professor Verner Schomaker, who has helped by giving us the benefit of both his deep understanding
of structural chemistry and his profound critical insight.
[END PAGE TWELVE]
[BEGIN PAGE THIRTEEN]
1. Astbury, W.T., and Lomax, R., J. Chem. Soc., 1935, 846; Astbury, W.T., Dickinson, S., and Bailey, K., Biochem. J., 29,
2. Pauling, L., and Corey, R.B. These Proceedings, 37, (1951).
3. Bayes-Watson, J., Davidson, E., and Perutz, M.F., Proc. Roy. Sec., A191, 83 (1947).
4. Perutz, M.F., ibid., A195, 474 (1949).
5. Pauling, L., and Corey, R.B.
6. Bragg, W.L., Kendrew, J.C., and Perutz, M.F., Proc. Roy. Soc., A203, 321 (1950)