The average number of amino acid residues in the unit cell of collagen was recalculated from the independent work of Bear and Pasternak. An estimate of the reliability of this average number was made.
1) The average number of amino acid residues was calculated by the following formulae:
[Data Equation]
n = average number of residues in a unit cell
f = fraction of total weight due to collagen
N = Avogadro’s number
V = volume of unit cell (hexagonal)
d = density
R = average residue weight
c = meridional spacing
b = equatorial spacing
2) In order to evaluate the fraction f from a nitrogen determination it is necessary to know the per cent of nitrogen in a sample of pure collagen. The per cent of nitrogen was calculated from the amino acid composition of mammalian collagen as given by Bowes and Kenton (Biochem. J., 43, 358 (1948)). In 10^{5} gms. of collagen thre are 1278 gm-atoms of nitrogen or 17,902 gms. The combined amide nitrogen of aspartic acid and glutamic acid was taken as 0.66 per cent from the experimental determination by Bowes and Kenton, although the calculated values from the amino acid composition are 0.66 and 1.08 per cent respectively.
calculated per cent of nitrogen
17.90 (amino acid nitrogen)
0.66 (amide nitrogen)
18.65 total
Bowes and Kenton determined the total nitrogen experimentally and found 18.6 per cent nitrogen. The fourth figure in the calculated value is not really significant since the data of Bowes and Kenton account for only 99.6 gms. of residue for each 100 gms. of collagen. The calculated per cent of nitrogen was taken as 18.6.
3) The following data were obtained from the work of Rougvie and Bear (J. Amer. Leather Chemists Assn., 48, 735 (1953)) on thoroughly dried untreated kangaroo tail tendon.
d = 1.34
c = 2.86 Å
b = 10.6 Å
% N = 18.0
f = 18.0/18.6 = 0.969
R = 92.6 (Bowes and Kenton)
V = 2.86 x 10.6^{2}/cos 30 = 371.1 Å^{3}
n = 0.968 x 6.023 x 10^{23} x 371.1 x 10^{-24} x 1.34/92.6
n = 3.13
In a previous report on July 11, 1955 the fraction f had not been taken into account so that the value of n was erroneously reported as 3.23.
4) The following data were obtained from Dr. R. A. Pasternak of this Institute who worked on partially hydrated untreated kangaroo tail tendon.
d = 1.32
c = 2.88 Å
b = 11.54 Å
% N = 15.72 (wet basis)
f = 15.72/18.6 = 0.845
V = 2.8 x 11.54^{2}/cos 30 = 442.9 Å^{3}
n = 0.845 x 6.023 x 10^{23} x 422.9 x 10^{-24} x 1.32/92.6
n = 3.21
5) A second value of n can be calculated from Dr. Pasternak’s data by using the water and wash determination. The water content (15.9%) and cation content (0.4%) were assumed by to be the only impurities.
f = 1 – (15.9 + 0.4) = 0.837
Using the same data as given in section 4) except for the value of f we obtain
n = 3.18
In the previous report the value of n was calculated by using an average residue weight of 93.3 which Dr. Pasternak calculated from the amino acid composition as compiled by Tristram (Advances in Protein Chemistry, Vol. V, 143 (1949)). The average residue weight of 92.6 (Bowes and Kenton) was taken in order to be consistent with the calculated nitrogen content of collagen.
6) Consideration of errors in n – The presence of impurities in untreated collagen can cause a number of serious but indeterminable errors. a) In the nitrogen determination any nitrogenous impurity will give a high value to the fraction f and cause a positive error. b) In the water and ash determination any impurity that is not water or ash will give a high value to f and cause a positive error. c) The fraction of the weight in the unit cell due to collagen may be greater than the experimental f since some of the impurity may be outside the unit cell and thus cause a negative error. d) The experimental density is the average density of collagen plus the impurities not in the collagen. This error can be plus or minus but it is more likely to be negative in a hydrated sample. e) There may be local variations of density and average residue weight along the helix of collagen but the local values of d/R can be expected to be close to the average value d/Ro. f) It is possible to give some numerical estimates of the errors involved in the deviations of the following quantities are based on two measurements of each quantity.
[Data Talbe]
The two best values of the average residue weight (92.6 and 93.3) have an average deviation of 0.38% but they are based on essentially the same data. No estimate is available for the meridional spacing. The errors due to true density and average residue weight are probably underestimated. Based on these estimates the minimum uncertainty in the value of n is about ± 1.5% or ± 0.05 residues per unit cell. The three values of n give the following average value:
3.13 (Bear)
3.21 (Pasternak)
3.18 ( “
Average = 3.17 ± 0.03
The average deviation falls within the estimated minimum uncertainty of ± 0.05. The present data indicate that n is slightly but significantly larger than 3.