The Reaction of Hydrogen Chloride With Dry Proteins
[Read before the Otago Branch, April 12, 1949; received by the Editor, March 7, 1949.]
The ability of the proteins to combine with acids and bases is one of their most general chemical properties. The great volume of literature concerned with this subject has been reviewed by Schmidt (1945), and a comprehensive discussion of the acid- and base-binding capacity of the fibrous proteins has been given by McLaughlin and Theis (1945). Of the many experimental methods employed, the majority have depended upon titration with acid or alkali in water or in aqueous salt solutions of various ionic strengths. As Schmidt has pointed out, titration methods are subject to several uncertainties, some of which become particularly acute when the maximum combined acid or base is to be measured. Nevertheless, it is generally considered that acid or base bound by proteins bears a stoichiometric relationship to their amino-acid content.
The problem would be simplified if the reaction between protein and acid could be carried out in the absence of water, but we are then restricted to acids which, in the anhydrous state, have no oxidizing or other destructive action on the protein molecule. The most suitable is hydrogen chloride, which has been used by several investigators.
Bancroft and Barnet (1930) applied the phase rule to the interaction of gaseous acids with proteins and amino acids, and the same principle was used by Czarnetzky and Schmidt (1934) in their method of gaseous titration. By plotting partial pressure of hydrogen chloride against hydrogen chloride taken up by the protein, they obtained evidence of compound formation, but did not isolate the compounds free from excess acid. Beek (1932) attempted to prepare compounds of collagen and hydrogen chloride, but his compounds varied widely according to the method of preparation. It will be shown here that it is possible to prepare a continuous series of stable compounds of hydrogen chloride and collagen whose composition can be related to the amino-acid content of the protein.
It was thought desirable to study the reaction with various modifications of collagen and also with another protein of widely different amino-acid composition. For the latter purpose silk fibroin was chosen on account of its very low content of the amino acids possessing basic groups in the side-chain, such as arginine, histidine and lysine.
Collagen in the form of standard hide powder was dried to constant weight in vacuo at 100° over phosphorus pentoxide.
Deaminised collagen. Collagen was treated with sodium nitrite solution and acetic acid at low temperature, according to the method of Highberger and Retzsch (1939). It was then washed, dehydrated with acetone, and dried as above.
Hypochlorite-treated collagen. Collagen was treated with ice-cold alkaline sodium hypochlorite solution in the manner recommended by Highberger and Salcedo (1940) for oxidation of guanidine groups. It was found very difficult to wash the product free from chloride ions, so it was washed and dried as before and residual chloride was estimated as described below. The chloride content, calculated as HCl in the dry product, was 0·73%. This figure was used to correct the analyses made later.
Hypochlorite-treated collagen subsequently deaminised. A portion of the previous product was deaminised, washed, and dried. Its chloride content was 0·48%, calculated as before.
Silk. Unweighted and undyed degummed silk fabric was dried as above. Its ash content was 0·08%. The acid-binding capacity of the ash, calculated on the weight of the original dry silk, was 0·03%. As the history of the sample was unknown, its hydroxyamino-acid content was measured by the periodate-ammonia method of Rees (1946) and found to be 0·136 gram-equiv. per cent., in good agreement with published values.
Hydrogen chloride was prepared from pure dry sodium chloride and sulphuric acid, and dried by passing through three sulphuric acid bubblers.
Ether was distilled from solid potassium hydroxide, allowed to stand over anhydrous calcium chloride, and then refluxed over and distilled from metallic sodium. It was stored in the dark over sodium wire.
Two methods of investigating the reaction were adopted
1. Vacuum-dry samples of the various protein materials were weighed in stoppered glass weighing bottles which were then opened in a desiccator and evacuated over concentrated sulphuric acid. After seven days, dry air was admitted and then displaced by a stream of dry hydrogen chloride passed through the desiccator by means of two stopcocks. In early experiments on rate of reaction, the desiccator was opened at intervals and one sample removed for analysis. The apparatus was then quickly evacuated to remove moist air, dry air was again admitted, and more dry hydrogen chloride was passed through. In later experiments, however, in order to exclude all traces of moisture, the desiccator was kept closed from the beginning to the end of the chosen reaction period. To measure the uptake of hydrogen chloride over different intervals of time, experiments were carried out in a series of desiccators, none of which was disturbed by opening during the experiment. A beaker of concentrated sulphuric acid was kept in the desiccator throughout. When a sample was removed from the hydrogen chloride, it was weighed and transferred to another
evacuated desiccator over calcium oxide to absorb excess hydrogen chloride. After a time in the desiccator, the sample was heated to constant weight in vacuo at 100° over calcium oxide.
2. Dry protein was suspended in anhydrous ether and dry hydrogen chloride was passed in until the ether was saturated. The mixture was agitated occasionally and more hydrogen chloride passed in. After a suitable interval, the greater part of the ether was drawn off through a tube plugged with glass wool and extending to the bottom of the reaction vessel. The last of the ether was evaporated in vacuo and another portion of dry ether was added. After 24 hours the ether was again removed in the same way, and the whole washing operation was twice repeated. The sample was now fairly free from excess hydrogen chloride. Finally it was heated to constant weight at 100° in vacuo over calcium oxide.
Analysis of samples. A weighed portion of the product was digested for three hours on the steam bath with 10% potassium hydroxide solution, when all but a few small particles dissolved. The liquid was acidified with acetic acid and filtered to remove any precipitate of protein degradation products. The total chloride was then precipitated as silver chloride in the presence of nitric acid, filtered, washed, and weighed in a sintered porcelain crucible.
1. Reaction with Gascous Hydrogen Chloride. An experiment was made to determine the rate of absorption of hydrogen chloride by collagen and the ease with which it could be subsequently removed. The collagen used as starting material contained 0·95% water, which is said by Czarnetzky and Schmidt (1934) to accelerate the reaction, but the rate was still found to be very slow. To determine how much of the hydrogen chloride taken up was bound tenaciously, the samples were desorbed at room temperature over calcium oxide in vacuo. It was found that, even with frequent renewal of the calcium oxide and cleaning of the desiccator, there was still a perceptible smell of hydrogen chloride over the samples and continued loss of weight after 80 days. Consequently the practice was adopted of heating samples at 100° in vacuo over calcium oxide after their hydrogen chloride content had been reduced to about 10% by desorption at room temperature. The second column of Table I, which gives the amount of hydrogen chloride taken up after different times of exposure to the collagen
|Time in HCl (hours)||Total HCl absorbed g./100 g. collagen||Final gain in weight g./100 g. collagen||Combined HCl g./100g. collagen||g./100 g. final product|
[Footnote] * Starting material contained 0.95% water. Results are calculated for the perfectly dry material.
samples, shows that the uptake had not reached a maximum even after five weeks. Moreover, the removal of more loosely bound hydrogen chloride is a slow process at room temperature. Thus the plot in Fig. 1 of the amount of hydrogen chloride left after different times of desorption at room temperature over calcium oxide shows that even after 75 days desorption is still proceeding. Desorption at 100° is, however, a more rapid process and the lower curve of Fig. 1 shows that desorption can be completed in about 6 days. In further work, samples were heated to constant weight, 7–10 days usually being required.
In spite of the low rate of uptake of hydrogen chloride and the slow desorption over calcium oxide, the amount of hydrogen chloride firmly bound, in the sense that it could not be removed at 100°, seemed, as
shown in Table I, to have reached a maximum value of about 5·6% after 710 hours.
The series of experiments summarised in Table I was regarded as exploratory. With the information afforded by these trials as a guide, several experiments were made to measure the quantity of hydrogen chloride required to saturate perfectly dry collagen. Great care was taken to exclude water during the reaction and subsequent removal of excess hydrogen chloride. Table II shows the results obtained, together with similar data for the modified collagen preparations and for silk.
|Material||Time in HCl (hours)||Total HCl absorbed g./100 g. protein||Final gain in weight g./100 g. protein||Combined HCl g./100 g. protein||g./100 g. final product|
|Hypochlorite-treated and deaminised||1,008||3.70||4.17||4.02|
One of the most striking features of this reaction with hydrogen chloride is the discrepancy, shown in the third and fourth columns of Tables I and II, between the weight of hydrogen chloride taken up, as estimated by gravimetric analysis for chloride, and the gain in weight of the collagen sample. These two quantities might be expected to be identical, but experimental measurements show decisively that the former is always in excess of the latter. A discrepancy was observed in every experiment and, although not very reproducible, was quite definite for each sample of collagen, deaminised collagen, and silk. It seems possible that these difference figures represent loss of water which was not removed when the original material was dried at 100°. With the products formed from the materials pre-treated with alkaline hypochlorite, however, there was a continual slow loss at 100°, the discrepancy increasing with the time of heating. This effect is demonstrated in Table III for three portions of the same sample. The last two columns of the table emphasize the stability of union between protein and hydrogen chloride, which persisted even through the twelve-day period when traces of water or other substances were being expelled.
|Before heating||g./100 g. original dry material Heated 6 days in vacuo||Heated 18 days in vacuo|
|Combined HCl by analysis||7.44||4.26||4.28|
2. Reaction with Anhydrous Ethereal Hydrogen Chloride. As in the previous section, a number of products were prepared using different reaction times, the ether being always saturated with hydrogen chloride. The nature of the method made it impossible to measure total hydrogen chloride absorbed and difficult to measure the gain in weight of the protein. Table IV therefore shows the composition of the samples in g. HCl per 100 g. product after drying in vacuo over calcium oxide at 100°, and is directly comparable with the last columns of Tables I and II.
|Time in HCl (hours)||Combined HCl g. HCl/100 g. product|
It has been shown that dry collagen can absorb approximately 30% by weight of dry gaseous hydrogen chloride when the partial pressure of the gas is atmospheric. Most of this is removed by reducing the partial pressure, but a portion is firmly combined, and products have been isolated which lose no hydrogen chloride even at 100° in vacuo. For lack of a better term, these products will be referred to here as collagen hydrochlorides. The reaction is a slow one, and saturation with respect to combined hydrogen chloride is achieved only after several weeks. Removal of loosely held hydrogen chloride is also slow and reminiscent of desorption of water vapour from the same protein.
From a large number of experiments carried out, some of which are reported in the previous section, it became evident that there is a continuous series of collagen hydrochlorides ranging from almost pure collagen to a product containing 5·6–5·7% hydrogen chloride calculated on the weight of the original protein. The same series, with approximately the same maximum, results from both methods of preparation. The single observation on silk shows that this protein also forms a stable hydrochloride with virtually zero vapour pressure at 100°. In view of the long duration of the experiment (10 weeks), it seems probable that the composition found for the silk hydrochloride represents saturation.
The existence of stable hydrochlorides was not indicated by the gas titrations of Czarnetzky and Schmidt (1934). They found evidence for a compound of gelatin with 12·5% hydrogen chloride, having a hydrogen chloride dissociation pressure of about 6·5 mm. at 25°, and a compound of silk fibroin with 1·46% hydrogen chloride, and a dissociation pressure of 14 mm. The compounds lost the whole of their combined hydrogen chloride at lower partial pressures. The virtual identity of composition of collagen and gelatin suggests that their hydrochlorides would also be very similar, and it appears probable that the higher hydrochloride of Czarnetzky and Schmidt was formed and dissociated again during the course of our preparations. However, we are unable to reconcile their data with the very stable
compounds described here. Beek reported several products which he regarded as stoichiometrically related compounds. In the light of our experience of the very slow removal of loosely held hydrogen chloride at room temperature, it seems probable that at least some of Beek's products merely represented intermediate points on a desorption curve.
The composition of collagen is well known, and the question arises whether the amount of firmly held hydrogen chloride can be related to the amino-acid content of the protein. Czarnetzky and Schmidt were able to do this for their products, but it will be noticed that their gelatin hydrochloride contained more and their silk hydrochloride less acid than the compounds described here. The much longer reaction time and the rigorous methods used in the present investigation to remove excess reactant make it desirable to inquire whether a quantitative interpretation can be found for the composition of the latter compounds.
As a working hypothesis, we may assume that one molecule of the acid is held by each of various basic or potentially acid-binding groups in the amino-acid side chains. In Table V are shown the hydrogen chloride capacities of certain amino-acid residues of collagen and silk fibroin, calculated respectively from the proportions reported by Chibnall (1946) and by Cohn and Edsall (1943).
|Amino acid||Collagen||Silk Fibroin|
|%Amino acid||HCl Equiv. % (1 mol. HCl/residue)||%Amino acid||HCl Equiv. % (1 mol. HCl/residue)|
|Total a + b||5.37||5.37|
|c. Amide nitrogen||0.09||0.23|
|Total a + b + c||5.60|
|Total a + b + c + d||9.68|
An examination of the figures in Tables I and IV shows that the firmly bound hydrogen chloride is fixed rapidly at first and the reaction then becomes much slower, taking many weeks for completion. This suggests that there may actually be two distinct reactions of widely different rates, the first with basic nitrogen retarded only by diffusion of the acid into the protein fibres, and the second a slow attack upon aliphatic hydroxyl groups. This point may be clarified by plotting the data of Tables I and IV in the form log (5·67—x) against time, where x represents g. HCl per 100 g. collagen with a saturation value of 5.67. In each case, all points except the first fall very close to a straight line which can be extrapolated to approximately zero time to give the quantity of hydrogen chloride taken up in the initial rapid reaction. Although the slopes of the two lines are quite different, their intercepts on the HCl axis agree well, yielding values of x equal to
3·38 for the gaseous reaction and 3·30 for the ethereal reaction. These results are almost identical with the figure appearing in Table V for combination with arginine, histidine, lysine, and hydroxylysine.
It seems very probable, then, that collagen on exposure to hydrogen chloride rapidly takes up 3·3% of the gas, which is equivalent to 1 molecule for each basic side chain. At the same time a much slower reaction begins, in which the hydroxyl groups of the straight chain hydroxyamino acids are replaced by chlorine atoms. This proceeds to saturation at 5·6% hydrogen chloride, agreeing fairly well with 5·37%; calculated as total a + b in Table V. Even this small discrepancy is eliminated if we assume that the amide groups react, or it may possibly be explained by reaction of the terminal amino groups of the polypeptide chains. It would appear that the hydroxyl group of hydroxyproline is not affected by hydrogen chloride under the conditions of our experiments.
We can now advance an explanation of the fact mentioned above that the weight gain of collagen is always less than the weight of combined hydrogen chloride. Any reaction of hydrogen chloride with hydroxyl groups will liberate one molecule of water for every hydroxyl group attacked. Therefore, for all the hydrogen chloride taken up in excess of 3·30% there should be loss of combined water to the extent of 18/36·5 of the weight of that hydrogen chloride. Reference to Tables I and II shows that the predicted and observed difference figures are of the same order of magnitude, and in Table I increase, albeit somewhat irregularly, with hydrogen chloride content. It has already been shown (Green, 1948) that loss of water from collagen depends on the drying method chosen. This encourages us to believe that, beneath the irregularities of the experimental data, we have rightly interpreted the course of the reaction.
In Table II we reported some results for the action of hydrogen chloride on certain modifications of collagen. Deaminised collagen had been treated with nitrous acid in an attempt to remove the terminal amino groups of lysine and hydroxylysine and the free acid amide groups. Another sample treated with alkaline hypochlorite should, according to Highberger and Salcedo (1940), have lost its guanidine groups by oxidation, but retained its amino groups. Finally, a sample treated with hypochlorite and then, after washing, with nitrous acid might be expected to have lost all its free basic groups with the exception of histidine residues.
All three samples showed a reduced capacity for hydrogen chloride, but the extent of the reduction could not be simply explained by the theory. The action of nitrous acid, in replacing an amino by a hydroxyl group, should not reduce the acid-binding capacity, but might be expected to increase the time needed to reach equilibrium. It is also possible that side reactions occur, since collagen under the action of nitrous acid acquires a yellowish-brown colour which deepens still further in the presence of alkali. This does not seem entirely compatible with a simple deaminising reaction. Evidence for rupture of the polypeptide chain by alkaline hypochlorite was found by Baker (1947).
The composition of silk fibroin hydrochloride reported in Table II can be shown to agree with our hypothesis in a rather striking manner.
The proportion of basic amino acids is very small, but the hydroxyamino acids, serine and threonine, are relatively abundant. Table V shows that the basic amino acids are quite inadequate to account for the observed uptake of hydrogen chloride, while the hydroxyamino-acid content predicts a value much too high. However, there is strong evidence that a portion of the latter residues are unreactive. Gordon, Martin and Synge (1943) wrote as follows: “Repeated treatment of silk fibroin with dimethyl sulphate in N NaOH at room temperature results in O-methylation of nearly all its tyrosine residues and about one-half of its serine residues. The groups accessible to etherification in this way are also accessible to acetylation when the fibroin is treated with boiling acetic anhydride. The existence is thus established of two categories of serine residue in silk fibroin differing in their accessibility to etherification by dimethyl sulphate.”
If we follow these workers and assume that of total b in Table V only half is available to hydrogen chloride, we arrive at a calculated figure of 2·80% for silk, in almost exact agreement with experiment.
A more detailed study of the reaction of hydrogen chloride with silk and modified collagen is in progress.
Further support for our hypothesis of a rapid reaction with the basic nitrogen of the side chains comes from the work of Felix and Rauch (1931) on the histone of the thymus gland. These authors prepared the hydrochloride by the action of ethereal hydrogen chloride on an ethereal suspension of the base for only ten minutes, and then heated it to constant weight at 100° in vacuo. They made detailed analyses of this product and of the methyl ester hydrochloride, but did not attempt to relate combined chlorine to the amino-acid content. In Table VI are reproduced some of the data of Felix and Rauch for the composition of their products, to which we have appended combined chlorine calculated on the basis of one hydrogen chloride molecule to each residue of arginine, histidine and lysine. Here again agreement is very good.
|Histone hydrochloride||Histone methyl ester hydrochloride|
|Van Slyke nitrogen||12.38%||12.27% of total N|
|Arginine nitrogen||25.68%||24.18% "|
|Histidine nitrogen||5.52% "||5.92%|
It should be emphasized that throughout the above discussion we have deliberately ignored those reactions which, while they may result in acid binding at low temperatures, do not yield products stable at 100° in vacuo.
Anhydrous hydrogen chloride reacts slowly with dry collagen, and saturation is reached only after several weeks.
Collagen takes up approximately 30% and silk approximately 17% by weight of gaseous hydrogen chloride, but most of this can be slowly removed by lowering the partial pressure.
A definite quantity of hydrogen chloride, depending on the duration of the reaction, is retained by the collagen even at 100° in vacuo. Its amount is lessened by pre-treatment with either nitrous acid or alkaline hypochlorite.
Silk fibroin also forms a stable hydrochloride.
The observed changes can be quantitatively explained by postulating a rapid reaction between hydrogen chloride and the basic amino-acid residues, and a much slower reaction with the aliphatic hydroxyl groups of the side chains.
The author wishes to express his indebtedness to Professor R. A. Robinson, of the University of Malaya, for many valuable discussions. He also wishes to thank Professor F. G. Soper for making available the facilities of his laboratory, and the Royal Society of New Zealand for a grant from their Research Fund.
Baker, R. W. R., 1947. Biochem. J., vol. 41, p. 337.
Bancroft, W. D., and Barnet, C. E., 1930. J. Phys. Chem., vol. 34, pp. 449, 753, 1217, 2433.
Beek, J., 1932. Bur. Standards J. Research, vol. 8, p. 549.
Chibnall, A. C., 1946, J. Intern. Soc. Leather Trades Chem., vol. 30, p. 1.
Cohn, E. J., and Edsall, J. T., 1943. Proteins, Amino Acids and Peptides, p. 358, Reinhold Publishing Corp., New York.
Czarnetzky, E. J., and Schmidt, C. L. A., 1934. J. Biol. Chem., vol. 105, p. 301.
Felix, K., and Rauch, H., 1931. Z. physiol. Chem., vol. 200, p. 27.
Gordon, A. H., Martin, A. J. P., and Synge, R. L. M., 1943. Biochem. J., vol. 37, p. 538.
Green, R. W., 1948. Trans. Roy. Soc. N.Z., vol. 77, p. 24.
Highberger, J. H., and Retzsch, C. E., 1939. J. Amer. Leather Chem. Assoc., vol. 34, p. 131.
Highberger, J. H., and Salcedo, J. S., 1940. Ibid., vol. 35, p. 11.
McLaughlin, G. IX, and Theis, E. R., 1945. The Chemistry of Leather Manufacture, pp. 67 et seq. Reinhold Publishing Corp., New York.
Rees, M. W., 1946. Biochem. J., vol. 40, p. 632.
Schmidt, C. L. A., 1945. The Chemistry of the Amino Acids and Proteins, pp. 720 et seq. Charles C. Thomas, Springfield, Illinois.