J . POURADIER 185 SOME ASPECTS OF TME REACTION OF BASIC CHROMIUM SALTS WITH HIDE PROTEIN BY K. H. GUSTAVSON Garvcrinaringens Forskningsinstitut, Stockholm, Sweden Received 2nd July, 1953 Experimental evidence of the dominating function of the ionized carboxyl groups of collagen for the fixation of cationic chromium complexes is given. Thus, collagen with its carboxylic groups completely discharged loses its affinity for cationic chromium compounds completely. By complete esterification of the carboxyl groups, the binding of these chromium complexes is prevented. Results from interactions of basic chromium chlorides and sulphates with gelatin, in 1 % solution, removing the uncombined chromium by means of cation exchanger, show that the basic salts are preponderantly uni-pointly fixed by gelatin.From data on the average charge of the cationic chromium complexes and the available number of carboxylic groups of collagen for fixation of the electro- positive chromium complexes, an approximation of the proportion of the uni- and multi- point fixation of these complexes by collagen indicates that at least 90 % of the amount of chromium held by collagen is unipointly fixed, 10 % being bipointly attached to the protein. The chromium complexes functioning as the cross-linking agent are to be found among the small amount of bipointly fixed chromium.186 REACTION OF CHROMIUM SALTS WITH HIDE Chrome tannage concerns generally reactions between aqueous solutions of basic sulphates and chlorides of chromium and collagen, the greatest part of the chromium complexes removed by collagen being in irreversible combination with the protein.The fixed chromium salt makes collagen resistant to the action of proteolytic enzymes and to hot water, even to boiling water by rationally performed tanning. This stabilization of the collagen lattice by chromium complexes is generally considered to be due to the cross-linking of the collagen chains through stable chromium bridges. Various types of chromium complexes, differing in composition, molecular size and electrical charge, are present in solutions of the baFic salts. In the present paper, some recent investigations of these systems will be discussed, selecting €or the main experiments as simple systems as possible that contain cationic chromium complexes mainly, to facilitate theoretical discussion and for didactic reasons.Results of recent investigations of the reaction of basic chromium salts with collagen, some modified collagens and polyamides are surveyed in this paper. EXPERIMENTAL THE COMPONENTS.-C~~~~~WH compounds.-A standard solution of basic chromium chloride with composition corresponding to the empirical formula : Cr2(0H)2Cl4 . 2NaC1, prepared by adding one mole of sodium hydroxide to one mole of chromium chloride, cr(oH&c13, represented the type of chromium salts with ions of little or no compkxing tendency. The basic chromium salts are in the following characterized by their acidity percentage, i.e. the number of equivalents of Cl or SO4 groups associated with chromium expressed in per cent. The chromium chloride is thus 67 % acid.The solution of basic chromium sulphate investigated was prepared by reducing a solution of sodium bichromate with sulphur dioxide, boiling the solution after complete reduction. The equilibrated solutions contained 0.6 equiv. Cr/l. (15.2 g/l. Cr203). By ion exchange analysis of the solutions of the chromium salts, employing Dowex 50 in the sodium cycle 1 and Amberlite IRA 400 as hydrochloride,2 only cationic chromium complexes were found in the basic chromium chloride. No complex-bound C1-groups were present in the chloride. From determination of the diffusion coefficient,3 it is known that there are two atoms of chromium present in the molecule of the 67 % acid chlorides and sulphates on the average. Hence, the complexes present in the chromium chloride may be represented by the formula (Cr2(QH)2)4+.The analysis of the solution of the 67 % acid chromium sulphate gave 96 % cationic and 4 % non-ionic chromium complexes. The acidity of the sulphato-chromium cations by means of the cationic and anionic resins, according to methods earlier described, gave a mean value of 33 %. Hence, the cationic complexes are represented by the general type of cations (Cr2(QH)2S04)2+. Whereas the electrochemical composition of the solutjon of the basic chromium chIoride does not change by increasing the chrome concentration, only positively charged chromium complexes being present, the formation of non-ionic and anionic chromium is greatly facilitated by increasing the concentration of the solutions of the type of basic chromium sulphate concerned.4 At a concentration of 6 equiv.Cr/l., for example, the solution contained 45 % cationic, 50 % non-ionic and 5 % anionic chromium complexes. In some experiments, solutions of masked chromium salts of greater complexity were employed : sulphito-chromium sulphates, obtained by adding increasing amounts of sodium sulphite to a solution of a neutral salt-free 67 % acid chromium sulphate, cor- responding to Cr2(OH)2(SO4)2 ; and further phthalato-chromium sulphate, obtained by adding 0.5 mole of sodium phthalate per mole Cr of the aforementioned chemically pure chromium sulphate.5 The complex composition of these solutions is given in connection with the experimental results. Substrates.-Hide powder &yon) and iso-electric calf skin pelt (split) served as intact collagen substrates.Further, the following modified specimens of collagen were used : hide powder with its carboxyl groups practically completely inactivated by esterifi- cation and hide powder with its ionic groups bIocked by irreversibly fixed condensed naphthalene-disulphonic acid. The methylated collagen prepared by the method of Fraenkel-Conrat and Olcott,6 as described in an earlier paper 7 contained 17.4 % N, 2 6 % OCH3 groups and 2-7 % Cl present as HCl attached to the basic protein groups.K. H. GUSTAVSON 187 Its acid binding capacity was 0.1 niequiv./g collagen (0.1 N HCl). The hide powder which had its ionic groups blocked by condensed naphthalene disulphonic acid (three consecutive treatments of hide powder in 5 % solution of the acid at a final pH of 1.2) contained 27.5 % fixed sulpho acid based on the weight of collagen. Its acid binding capacity was 0.07 mequiv.H ion per g collagen (as 0.1 N HCl). Finally, a hydrated polyamide, earlier described,g represented a substrate with free -CO . NH- groups as the main reacting group. Its N content was 12.5 % of the weight of polyamide. Its acid binding capacity 0.05 mequiv. H+ and base binding 0 0 2 mequiv. OH- per g polyamide. In some experiments, a cation exchanger with carboxyl ions as functional groups was used. It was equilibrated against Na ions at pH 7-0 and contained 4.8 mequiv. free carboxyl ions per g dry resin.9 GENERAL 0RIENTATION.The reactive sites of collagen are : (i) ionic groups, such as the carboxyl and ammonium groups (the e-amino group of the lysine residue and the guanidyl group of arginine); (ii) non-ionic groups, such as the peptide link and the hydroxy groups of the residues of serine, threonine and hydroxyproline.In reactions of metal ions with proteins, any of these groups may be involved. The Zn ion appears to have preferential affinity for the imidazole group of the histidine residue of globular proteins.10 Collagen also binds Ag and Pb ions in the same manner as a recent investiga- tion of Grassmann and Kusch 11 shows. Cu2+ ion reacts primarily with carboxyl ions of amino acids, forming stable internal complex salts.12 With collagen and gelatin the copper ion Cu2+ is attached to the carboxyl groups,l3 whereas in some globular proteins the imidazole group forms the binding site.14 Hence, apparently various types of binding groups are involved in the interaction with cations of heavy metals, according to the nature of the protein, the pH of the reacting system and the specific environment.The mechanism of chrome tanning was conceived by Wilson 15 as a reaction between the carboxyl group of collagen and chromium ion. However, it was difficult to conceive such a reaction from the general reactivity of proteins, postulated by the classical concept of protein reactions. The introduction and application of the zwitter-ion concept to protein reactions removed the theoretical apprehensions and, indeed, the attraction and fixation of cationic chromium complexes by the carboxyl ions should be the logical out- come.16 Hence, since chrome tanning takes place in an acidic medium, a competition between protons and cationic chromium complexes should be set up in the chrome tanning process.This deduction agrees with the established facts of the process. Before discussing the principal evidence for the rnechani sm of the chrome fixation, it is appropriate to mention that certain findings have long been known which indicate the carboxyl group to be the main reacting group of collagen in the fixation of ordinary chromium salts. For instance, basic chromium complexes of green colour diffusing into gelatin gel colour the gel violet,l' the colour of the chromium carboxylate. The light absorption curve of solutions of slightly degraded gelatin treated with chromium salts is similar to that of chromium-amino acid complexes,l8 in which compounds the binding of the carboxyl group directly on the chromium atom has been proved.RESULTS AND DISCUSSION EXPERIMENTAL EVIDENCE FOR THE BINDING OF CATIONIC CHROMIUM COMPLEXES BY THE CARBOXYL IONS OF coLLAGEN.-The following experimental evidence for the function of the carboxyl ion as the main binding group of cationic chromium complexes may be cited. The eflect of inactivating the ionic protein groups by sulpho-acids.-In com- parative experiments using regular hide powder and the hide powder devoid of ionic groups (combined with the maximum amount of the condensed naphthalene disulphonic acid) the fixation of basic chromium sulphate was determined by running simultaneously series with intact hide powder. Upon treating portions of substrate equal to 2.0 g collagen with 100 ml of the solution of basic chromium sulphate for 6h, intact hide powder fixed 5.0 % Cr2O3 and the inactivated hide powder 1.4 % Cr203, on the basis of collagen.At a concentration of 3.0 equiv. Cr/l. the corresponding figures were 6-9 and 2.4 %. The condensed naphthalene disulphonic acid was selected as the inactivating agent of the ionic protein groups because it does not form insoluble compounds with the basic chromium sulphate and moreover, since it would be expected not188 REACTION OF CHROMIUM SALTS WITH HIDE to interfere with the reactivity of the non-ionic protein groups, primarily the peptide links. The data show that the uptake of chromium by the hide powder devoid of ionic groups is considerably greater than the amount which can be bound by the free carboxyl groups (0.07 mequiv.per g collagen) which is estimated to be about 0.4 % Cr2O3. The additional fixation is probably due to secondary interaction of the chromium complexes with free sulphonic acid groups in the sulpho acid fixed by collagen, i.e. an example of cation exchange. This explanation is supported by the finding that basic chromium chlorides are still more heavily fixed, these complexes evidently being capable of a still more extensive co-ordination with these sulpho groups. However, this experiment indicates the ionic protein groups to be the main binding sites of cationic chromium complexes. The efect of the discharge of the carboxyl ions of collagen.-The first conclusive evidence for the dominant role of the carboxyl ions of collagen in the binding of cationic chromium complexes was obtained by investigation of the behaviour of basic chromium chloride towards collagen which had its carboxyls in the non-ionized state, the ionic groups being discharged by equilibrating the collagen in solution of hydrochlorjc acid The calf skin split was pickled with hydrochloric acid in 5 % solution of sodium chloride to an equilibrium pH of 1.1.Under these conditions the ionized carboxyl groups of collagen are completely discharged. The solution of the 67 % acid chromium of PH 1.0.19 FIG. 1.-The fixation of 67 % acid chromium chloride by collagen (I) and HC1-saturated collagen 01) as a function of the concentration of chromium. chloride at a concentration of 6 equiv. Cr/l.was boiled for 30 min to increase its resistanca to acids. Solutions of various chrome content containing 5 vol-% NaCl and a sufficient amount of hydrochloric acid to give pH 1.0 were made up immediately before use. Portions of lightly pressed pickled pelt of pH 1.0 (= 4 g collagen) were treated in 200 ml of these solutions for 6 h. This rather short time was selected to minimize the interaction of the added hydrochloric acid with the basic chromic chloride. Blank series were run in the same manner with untreated calf skin split (isoelectric), except that no acid was added to the solutions of the boiled chromium chloride. The results are given graphically in fig. 1, showing the chrome fixation as a function of the chrome concentration of the solutions.No chromium is fixed by collagen with completely discharged carboxyl groups from solutions of basic chlorides adjusted to pH 1.0 in concentrations up to 60 g/l. Cr2O3. In the more highly concentrated solutions, only smalI amounts of chromium are fixed. Hence, the curves prove the chrome fixation by collagen from solutions of purely cationic chromium chlorides to take place on the ionized carboxyl groups of the protein. The results of the corresponding experiment with the basic chromium sulphate em- ploying pelt equilibrated with sulphuric acid at pH 1.1 in 5 volume % sodium sulphate, are graphically represented by fig. 2. It is to be noted that by increased concentration of the basic chromium sulphate the formation of non-cationic chromium complexes will be prominent whereas with the basic chloride of fig.1 such formation does not occur. The curves of fig. 2 show that no chromium is fixed by the acid-saturated collagen from solutions of chromium concentrations less than about 20 g/l. Cr2O3. However, withK . H . GUSTAVSON 189 increasing chrome concentration, collagen devoid of ionized carboxyl groups gradually commences to bind chromium. This fixation consists probably mainly of non-ionic complexes which are attached to non-ionic protein groups. Hence, in tanning with basic chrome sulphate liquors of considerable chromi um content, at least two types of reaction are to be recognized : (i) the ionic fixation of cationic chromium complexes and (ii) the fixation of non-ionic chromium complexes by means of groups other than the carboxyl.The former reaction has been proved to involve that type of chromium fixation which is effecting the high degree of hydpothermal stabiliza- tion of the collagen lattice in the chrome tannage. The egect of esterifying the carboxyl groups u f colhgen. -The above experiments of demonstrating the complete blocking of the fixation of cationic chromium complexes by discharge of the carboxyl ions of collagen, is limited to the low pH mentioned. By per- manent inactivation of the carboxyl groups, for instance by their esterification, investiga- tion of the mechanism of the chrome tannage in a more desirable pH range, or that of the ordinary chrome tanning bath, such as pH 2.5-4 is feasible. Such experiments have been performed by Bowes and Kenten,zo who have methylated collagen by means of methyl sulphate, and methyl bromide.In view of the non- specificity of these agents for the carboxyl group and the marked hydrolytic breakdown of the protein incurred by the large number of treatments necessary for complete esterifica- tion, care must be taken in the interpretation of the data. Bowes and Kenten20 found that methylation of collagen decreased the chromium fixation from basic sulphates very markedly. Thus, the amounts of chromium fixed at final pH values below 3 were less than FIG. 2.-The fixation of 67 % acid chromium sulphate by collagen (I) and HzSO4-saturated collagen (11) as a function of the concentration of chromium. 1 % Cr2O3. At pH values about 4, amounts of 2-3 % CrzO3 were fixed by collagen.The authors were convinced that it is not unreasonable to assume that in the absence of carboxyl groups no chromium would be bound by collagen. By means of the method of esterifkation devised by Fraenkel-Conrat and Olcott,6 employing methanol as the esterifying agent in the presence of small amounts of hydrochloric acid, it is possible to inactivate the carboxyl groups of collagen completely without markedly interfering with other protein groups and with a minimum risk of degradation.7 No chromium was fixed by the methylated collagen from solutions of basic chromium chlorides and perchlorates of the cationic type? The same was found to apply to the reaction of the dilute solutions of basic sulphates. Highly con- centrated solutions of the chromium sulphate which contain large amounts of non-cationic chromium possess some affinity for the collagen devoid of carboxylic groups, exactly as found in the experiments of fig.2, although the amounts of fixed chromium were considerably less. Some reactions of chromium complexes of various electrochemical type.-This non- ionic fixation of chromium reaches very large proportions and is dominating in systems of esterified collagen and certain complexed (“ masked ”) chromium salts, as the following experiments will demonstrate. To a stock solution of the 67 % acid chemically pure chromium sulphate increasing amounts of sodium sulphite were added, as noted in table 1. The solutions were adjusted to contain 0.8 equiv. Crll. and aged 3 weeks before use. The complex composition of these solutions given in table 1 was determined by the ion exchange method (Dowex 50, operated in the sodium cycle and Ainberlite IRA 400, in the form of hydrochIoride’).7190 REACTION OF CHROMIUM SALTS WITH HIDE Untreated hide powder and methylated hide powder (= 0 5 g protein) were shaken with 20.0 ml portions of these solutions for such a short period as 2 h, in order to eliminate hydrolysis of the methylated carboxyls.Since it should be of interest to ascertain the behaviour of a substrate with carboxylic ions as the only functional group towards these chromium complexes, series with a carboxyl type of cation exchanger (Amberlite IRC 50) were included. Portions of 0.5 g of dry resin were shaken for 2 h in 100 ml of solutions. The results of these experiments are graphically shown in fig.3. The data of fig. 3 show that the cationic chromium complexes of the blank solution (no. 1) possess a slight affinity for the protein devoid of carboxyl groups TABLE 1 .-ELECTROCHEMICAL COMPOSITION OF CHROMIUM COMPOUNDS % chromium present as cationic non-ionic anionic no. composition PH - 1. cr2(oH)2(s04)2 2 9 96 4 0 2. WOH)2(S04)2 + Na2SO3 3.4 63 37 0 3. @r2(QH)2(SO4)2 + 2 Na2SQ3 40 12 79 9 5. WOH)2(SO4)2 + 3 Na2SO3 5.7 86 12 (2) 0 85 15 4. Cr2(QH)2(SO& + 2 5 Na2S03 4-8 only. On the other hand, the untreated hide powder binds considerable amounts of chromium. By the addition of sodium sulphite to the solution of the chromium sulphate, the pH is increased which connotes increase of ionized carboxyl groups of collagen, and hence of fixation of cationic chromium.Further, non-cationic complexes are gradually formed by the penetration of sulphite groups into the chromium complexes. These complexes are fixed by both types of hide powder. Solutions no. 4 and 5, mainly containing non-ionic chromium, react more ex- tensively with the esterified collagen than with the intact protein. This is probably due to increase of the number of co-ordination active sites of the collagen chains, FIG. 3.-The fixation of chromium from solutions of basic chromium sulphate complexed by means of sodium sulphite by : -x-x-x- intact collagen, -e--@--@-- collagen with inactivated carboxyls, -0-0-0- cation exchange resin of carboxylic type. i.e. non-ionic groups, resulting from rupture of cross-links, probably H-bonds, which has been indicated to occur in the methylation process as shown by the findings of Bowes and Kenten 20 and by the author.21 The reactivity of the car- boxyl ions of the resin is evidently restricted to the positively charged chromium complexes.Comparing the curves of resin and collagens, the predominance of the fixation of non-cationic chromium by collagen in systems containing 2-3 moles sulphite per mole of Cr203 is evident. A number of earlier findings, particularly on the influence of pretreatment of collagen in concentrated solution of lyotropic agents and the effect of hydrothermal shrinkage (denaturation) ofK . H. GUSTAVSON 191 collagen, point to the binding of non-ionic complexes of chromium to non-ionic protein groups, such as the peptide link by co-ordination (W-bondings).The hydroxy group of serine and hydroxy-proline may also be involved.21 The find- ings discussed prove the exclusive function of the ionized carboxyl groups of collagen for the initial attraction of electropositive chromium complexes and for the final irreversible fixation of these complexes by collagen, the high complexing power of the carboxyl group directing the ultimate attachment. The great avidity of collagen lacking in carboxyl groups for non-ionic complexes also provides evidence for the participation of groups other than the carboxylic group in the binding of these chromium complexes by hide protein. THE REACTION OF POLYAMIDES WITH CHROMIUM sALTs.--In this connection, the behaviour of basic chromium salts to an artificial substrate with the -CO.NH- link as its main reacting groups and practically devoid of ionic reactivity is of interest. The modified polyamide, in the hydrated state has been found to be exceedingly reactive towards agents which possess hydrogen-bonding ability, for instance, polyphenols of which vegetable tannins are an outstanding example.8 The cationic chromium complexes present in the standard solutions of the basic chloride and sulphate were not taken up by the polyamide. A mirior fixation of chromium from concentrated solutions of the sulphate was found, however. From dilute solutions of extremely basic chromium perchlorate (33 % acidity), the polyamide fixed large amounts of chromium, amounting to 7.1 % Cr2O3 of its weight. This solution contained 70 % of its chromium in the form of non-ionic complexes.Also the solution of the phthalato-chromium sulphate which consisted of 65 % non-ionic and 35 % cationic chromium complexes, showed affinity for the polyamide, which fixed irreversibly 4.5 % Cr203. The experiments 4 with the polyamide constitute an additional proof of the non- participation of the -CO . NH- group in the fixation of cationic chromium and indicate also that certain non-ionic chromium complexes may be fixed by this link, the most frequently occurring group in proteins. UNI- AND MULTI-POINT FIXATION OF CHROMIUM COMPOUNDS.-with practically conclusive evidence that the ionized carboxyl groups of collagen are governing its fixation of ordinary chromium salts, it is possible to proceed with th.e second part of the problem of the nature of chrome tanning, i.e.the extent of ionic dis- charge of the polyvalent cationic complexes by the carboxyl ion of the protein, on the proportion of unipointly and multipointly fixed complexes. As pointed out earlier, the 67 % acid chromium chloride contains mainly complexes of the type (Crz(OH)#+. It is reasonable to expect from steric considerations that there is a remote possibility for complete reaction of such a complex with four carboxyl ions of collagen. The length of a Cr-O-Cr--chain is estimated to be about 6A and by protolysis small amounts of complexes containing four atoms of chromium may be formed which would have an extension of ca. 18 A. The formation of complexes in situ is more likely to occur with the basic sulphates than the chlorides.EXPERIMENTS WITH GELATIN soLmoN.-Since the collagen lattice is a rigid structure with the ionized carboxyl groups probably interspaced at a considerable distance and the inaccessibility of some reactive groups are to be expected, some preliminary experiments were carried out with gelatin in 1 % solution, in order to avoid the interference of spatial factors. These experiments were followed up by series on tanning with hide powder. The following series of experiments will give an idea of the general approach. To portions of 10.0 ml of 1 % gelatin, quantities of the standard solutions of basic chloride and sulphate equal to 1.0 mequiv. Cr (25-3 mg Cr203) were added. The solutions were shaken after the addition of the chromium salt and left for various lengths of time for reaction to take place.They were then shaken for 20 min with moist Dowex-50, employed as the sodium salt (= 2 g dry resin) in order to remove the uncombined chromium. The chromium combined with gelatin was determined in the filtrate from the cation exchanger. The amount of ion exchanger used corresponded to 10 mequiv. cation and the quantity of cations added to the gelatin solution was less than I mequiv. using the ionic equivalent of the chromium salts. The amount of chromium present in the form of non-ionic192 REACTION OF CHROMIUM SALTS WITH HIDE complexes of the basic sulphate was corrected for. Fig. 4 shows the binding of chromium by gelatin as a function of the time of interaction. It is interesting to note that the chromium chloride + gelatin system reaches the equilibrium state practically instan- taneously (in less than 1 min), whereas a few hours is required for the sulphate systems to reach the maximum chrome fixation, being particularly pronounced for the neutral salt-free chromium sulphate. The gelatin solutions treated with the basic chromium chloride contained generally 14-15 % CrzO3 in combination with gelatin.Since gelatin contains 1.0-1.1 mequiv. carboxyl ions per g gelatin, and assuming about 90 % of these groups to be present as COO- groups at the pH value of 3.5 employed and to have reacted quantitatively with chromium, values of the mean equivalent weight of 140-150, in terms of Cr2O3, are obtained for the cationic chromium complexes fixed by gelatin.Hence, practically all of the chromium complexes combining with gelatin must be attached in the form of complexes of the general type (Cr2(0H)~C13+. With a mean equivalent weight of 38 for the cationic complexes present in the original solution, expressed as Cr2O3, or (Cr2(OH2)/4, the maximum amount of chromium bound by gelatin should only be about 3-6-34 % on the basis of protein. Accordingly, it is shown that in systems with readily accessible reactive protein groups, such as gelatin in dilute solution, the basic chromium chloride behaves towards the protein mainly as a unifunctional agent. FIG. 4.-The fixation of chromium by gelatin (in 1 % solution) as a function of the time of interaction from solutions of I: 67 % acid chromium sulphate, Cr2(0H)2(SO& ; 11: 67 % acid chromium sulphate liquor, Cr2(0H)2(SO& .Na2SO4 ; 111: 67 % acid chromium chloride, Cr2(OH)2Cl4. EXPERIMENTS WITH COLLAGEN.-In tanning of hide powder with the standard solution of the chromium chloride at a concentration of 25 g/l. Cr203, the following figures of the composition of the chromed hide powder were recorded. The amount of chromium fixed was 9.6 % and the amount of combined C1 found was 7.4 % Cl, both values calculated on the weight of collagen. Since the binding capacity of collagen for hydro- chloric acidis 0.9 mequiv. HCl per g collagen (= 3.2 % Cl), at least 4.2 % C1, based on the weight of the protein, should be associated with the chromium complexes fixed by collagen. Since according to the ion exchange analysis, the chloride quantitatively is in the ionic form, no chloro groups being present, the chloride groups in excess of the amount of the maximum binding capacity of collagen for HC1 must be associated with the hydroxo-chromium complexes fixed by collagen as compensating ions to the posi- tively charged chromium complexes which are not completely discharged by the carboxyl ions of collagen.The mean equivalent weight of the hydroxo-chromium complexes fixed by collagen should be of the order of 120 assuming 80 % of the available carboxyl groups (about 0.8 mequiv. per g collagen) to be partaking in the binding at the pH of the system, i.e. a pH 3. Thus, the main part of the chromium combined with collagen must have been fixed as unifunctional complexes. The data indicate that only a small percentage of the fixed chromium is multipointly attached, which is the potential cross-linking agent.The relatively low degree of hydrothermal stability imparted to collagen in tanning with basic chromium chlorides, compared to the corresponding tanning with basic sulphates, may possibly be due to this preference of the cationic complexes of chromium chlorides for unipoint combination with collagen.K. H. GUSTAVSON 193 The drastic improvement of the hydrothermal stability of chromium chloride tanned leather by its subsequent treatment in solutions of salts of bifunctional carboxylic acids, such as adipate, was explained by Holland 22 as due to an increase in the number of cross- linkages between the collagen chains, the bifunctional residue of the dicarboxylic acid joining two chromium complexes in collagen tanned with chromium chloride. Thus, for instance, a chloride-tanned leather showing an area loss of 60 % in the boiling test, is made shrinkproof after 2 days’ immersion in a 0.25 M solution of sodium adipate.Since the length of an adipate bridge of the type, Cr-0-Cr-OCO . (CH2)4. OCO . Cr-0-Cr, should be about 23& and the side chain distance between protein chains of hydrated collagen is about 17 A, the steric conditions for the postulated cross-linking are favourable. Corresponding series of experiments were run with the standard solution of the 67 % acid chromium sulphate. The chrome fixation curve of gelatin (no. I1 of fig. 4) shows at equilibrium a value of fixed chromium of 13.2 % Cr203 on the basis of gelatin, which proves beyond any doubt that the largest part of the chrome fixation takes place as unipoint binding of cations of the type (Cr2(OH)2S04)(S04/2)+.Nevertheless, the multi- point interaction and binding, necessary for the cross-linking function of the chromium complexes, is as a rule considerably more pronounced for the basic sulphates than for the corresponding chlorides. In tanning experiments with the basic chromium sulphate in solutions of 25 g/l. Cr203 the hide powder fixed 11.0 % Cr2O3 on the weight of collagen. The total sulphate content was 13.7 %. By removing the ionically-held SO4 groups by shaking the chromed hide (= 2-0 g collagen) in 50 ml of 4 % solution of pyridine €or 1 h,23 the sulphate content was reduced to 6-7 % SO4, on the basis of collagen.Hence, an amount of 7.0 % SO4 was present outside the chromium complexes probably in the form of protein-bound sulphate and as sulphate ions compensating the charged chromium complexes attached to collagen by unipoint binding. Since the maximum binding capacity of collagen for sulphuric acid at equilibrium pH > 1.2 is about 4 5 % SO4, on the basis of collagen, an amount of 2.4 % SO4 at least should be present in the third form, assumed to function as compensating ion to the charged, unipointly fixed sulphato-chromium complex. Basing the calculations on the figures given and assuming 80 % of the ionized carboxyl groups of collagen at a final pH value of about 3.2 (0.8 mequiv. per g collagen) to be in- volved in the fixation of the cationic chromium complexes, the chromium compounds fixed by collagen should consist of about 20 % bipointly attached (Crz(OH)zS04)- complexes (= 1.1 % Cr2O3 on the weight of collagen) and of 80 % unipointly attached complexes of the general type : (Cr2(OH)zSO4)(SO4/2)+ (= 9.9 % Cr203).These figures refer to the amount of available carboxyl groups. Hence, the amount of Cr2O3 incor- porated with collagen is composed of about 90 % of the unifunctional complexes carrying residual electrical charge and 10 % of bifunctional complexes containing both of the valencies compensated by the carboxyls of collagen. CROSS-LINKING BY CHROMIUM coMPLExEs.-The cross-linking complexes are among the 1 % chromium oxide which is bipointly fixed by collagen. Although the uncertain factor in these approximations is the amount of the carboxyl ions reacting with chromium, the result probably gives a fair picture of the actual conditions.Weir’s finding 24 that the optimal stability of collagen is obtained by fixation of such small amounts of chromium as 1 % Cr2O3, estimated from data of the thermodynamic quantities, is no Ionger a puzzle. The decreased stabil- ity of collagen indicated by Weir’s thermodynamic data to result from the binding of large amounts of chromium may then be explained by the predominance of unipoint binding of the chromium complexes by single carboxyl groups which connotes rupture and elimination of a part of the ionic cross-links of the original collagen, which bridges in a small way contribute to the organization of the protein structure .Weir24 believes that only a fraction of the acidic and basic protein groups possess the required spatial orientation for reaction with the chromium complexes forming cross-links. He suggested that the amount of chromium bound to collagen in excess of 1 % Cr203 probably is combined in a similar manner as are other types of tanning agents which decrease the entropy of activation. Weir concluded that this additional chrome will add nothing to the orientation of the G194 REACTION OF CHROMIUM SALTS WITH HIDE protein chains and their stabilization. On the contrary, incorporation of large amounts of chromium will mean some loss of orientation of the protein lattice. These suggestions are in harmony with the findings and the deductions of the present paper and with the concept of the dual nature of chrome fixation based on the present and previous experimental data.It is interesting to note that by interaction of chromium sulphate with gelatin solution, the resulting chromed gel will withstand the action of boiling water even at such low chromium contents as 0-2-0.3 % bound chromic oxide, whereas about 3-4 % Cr203 is the minimum amount of fixed chromium required to render collagen resistant to boiling water. This point has been stressed particularly by ElOd.25 This comparison is a striking illustration of the importance of the steric conditions in reactions of fibrous proteins . CONCLUDING COMMENTS.-The chrome tanning process exemplified by the interaction of collagen with basic chromium sulphate, mainly present in dilute aqueous solutions as (Cr~(OH)~S0~)-complexes, may in the light of modern researches be described as follows.The complexes and protons, present in the protolyzed solution of the basic salt, are attracted to and compete for the carboxyl ions of collagen. In order to maintain the electro-neutrality of the system, the corresponding sulphate ions are simultaneously attracted to the cationic protein groups, the main part of these sulphate ions being originally associated with the sulphato-chromium cations which are in great excess over the protons. In view of the considerable distance between adjacent ionized carboxyl groups, which is made probable from considerations of the chemical composition of collagen, and the rather small size of the sulphato- 0x0-chromium complexes of moderately basic chromium sulphates, with an estimated length of 6-12& it seems probable that the main part of the complexes are only able to react with one carboxyl ion, the unipointly fixed complex balancing its residual charge by an equivalent of sulphate ions.The initial ionic reaction with the discharge of electrostatically attracted complexes by the carboxyl ions is probably followed by the penetration of the carboxyl group into the chromium complex, forming a co-ordinate-covalent type of linking of great strength. The direct attachment of the carboxyl group to the chromium atom is indicated by a number of observations. It seems reasonable to believe that only a small part of the complexes, probably not more than one-tenth of the total amount of chromium present in heavily chromed leather, will be able to combine with two carboxyl ions of such close proximity as required for incorporating two carboxyls in the chromium complex.There are two possibilities for this bifunctional fixation (i) intra-molecular binding, which does not contribute to the stabilization of the protein, and (ii) inter-molecular, which combination results in cross-linking of adjacent collagen chains and increased stability of the collagen lattice. In view of the uncertain and unsatisfactory state of our present conception of the structure of collagen, particularly agpavated by the introduction of the helical concept, which appears not to have advanced far enough for the application to intricate chemical reactions of collagen, the discussion of the steric possibilities has been avoided in the present communication although in the final analysis it is likely to be the crucial factor. 1 Gustavson, J. Amer. Leather Chem. Assoc., 1950, 45, 536 ; J. SOC. Leather Trades 2 Gustavson, J. Soc. Leather Trades Chem., 1951, 35, 270. 3 Riess and Barth, Collegium., 1935, 62. Jander and Scheele, 2. anorg. Chem., 1932, 4 Gustavson, Svensk. Kem. Tidskv., 1944,56, 14 ; J. Colloid Sci., 1946, 1, 397 ; J. Soc. 5 Gustavson, J. SOC. Leather Trades Chem., 1951, 35, 160. 6 Fraenkel-Conrat and Olcott, J. Biol. Chem., 1945, 161, 259, 7 Gustavson, J. Amer. Chem. SOC., 1952, 74,4608. Chem., 1950, 34,259. 206, 241. Leather Trades Chem., 1946, 30, 264.K. H. GUSTAVSON 195 8 Gustavson and Holm, J. Amer. Leather Chem. ASSOC., 1952, 47, 700. 9 Gustavson, J. SOC. Leather Trades Chem., 1952, 36, 182. 10 Tanford, J. Amer. Chem. SOC., 1951, 73, 504. Gurd and Goodman, J. Amer. Chem. 11 Grassmann and Kusch, Z. physiol. Chem., 1952, 290, 216. 12 Ley, Z. Elektrochem., 1904, lQ, 954. 13 Thomas and Seymour-Jones, Ind. Eng. Chem., 1924,16,157. Northrop and Kunitz, 14 Tanford, J. Amer. Chem. SOC., 1951, 43, 504. 15 Wilson, J. Amer. Leather Chem. Assoc., 1917, 12, 108. 16 Gustavson and Widen, Ind. Eng. Chem., 1925,17, 577 ; Collegium, 1932, 775. 17 Kuntzel and Riess, Collegium, 1936, 138. 18 Kuntzel, Kolloid-Z., 1940, 91, 152. 19 Gustavson, Svensk Kem. Tidskr., 1940, 52, 75. 20 Bowes and Kenten, Biochem. J., 1949,44, 142. 21 Gustavson, Acta Chem. Scand., 1952, 6, 1443. 22 Holland, J. Int. SOC. Leather Trades Chem., 1940, 24, 221. 23 Gustavson, J. Amer. Leather Chem. Assoc., 1927,22, 60 ; Leder, 1952, 3, 293. 24 Weir, J. Amer Leather Chem. ASSOC., 1949, 44, 142 ; J. Res. Nat. Bur. Stand., 1949, Weir and Carter, J. Res. Nat. Bur. Stand., 1950, 44, 599. 25 Elod and Schachowskoy, Cullegium, 1933, 701. Suc., 1952, 74, 670. See also Edsall, Faraday SOC. Discussions, 1953, 20. J. Gen. Physiol., 1928, 11, 481. 42, 17.