The acid-soluble collagen of cod skin

The acid-soluble collagen of cod skin

ARCHIVES OF BIOCHEMISTRY ASD The Acid-Soluble E. GORDON From the Atlantic 88, 373-381 (1960) BIOPHYSICS Regional Collagen YOUNG AND Labor...

3MB Sizes 1 Downloads 198 Views








the Atlantic

88, 373-381 (1960)



Collagen YOUNG


Laboratory, Halifax, Nova Received

of Cod


National Research Scotia, Canada




of Canada,

28, 1960

Cod skin has been found to contain 75% collagen, 10% other protein, 2.5% peptides and free amino acids, 0.6% mucopolysaccharide, 1.0% lipid, and 12cz ash. About 95% of the total solids were dissolved by repeated ext,raction of the skin with mild aqueous solvents of pH 3.4-8.7 at 3-90°C. Collagen has been extracted with buffers of various organic hydroxy acids and their sodium salts at pH 3.6. It was recovered as needle-like fibrils, about 30 X 1 p, with characteristic banding. It contained 18.2% N, 0.02% hexosamine, 8.0% hydroxyproline (== 4.77, of the total N) and a distribution of 18 amino acids quantitatively identical with collagen from cod swim bladder. The specific optical rotation changed from [oL]:’ -349 f 50” to [cx]~~”-107 f 2” and was partially reversible with time and temperature to [cx]:” -223 f 12”. The value for .s&,~ was 3.17 ZIZ 0.10 S, measured at 6-g%., and for [v] 12.8 i 0.3 dl./g. measured at 1.4%. From these values the particle weight was calculated to be 280,000 f 20,OOOfor an elongated rod of 12.1 A. diameter and 2810 A. length on the assumption of B = 0.700. An irreversible change into two components occurred sharply at 13 f 0.5”C., as shown especially clearly in the ultracentrifuge. The value for [T] of the mixture was -1. INTRODUCTION


In a previous investigation the acid-soluble collagen, ichthyocol, extractable from the fibrous tunic of cod swim bladder was examined (1). The collagen of cod skin has also been referred to as ichthyocol, but it,s identity with t,hat of the swim bladder has not been established. It also has been reported to have a relatively low content of hydroxyproline (2%4), a low temperature of denaturation (2), and a low hydrothermal stability (4, 5). The preparation and further characterization of this collagen are herein described. It is of considerable interest in comparison with similar collagens from carp and from mammalian tissues and is of importance as the precursor of some commercial glues. 1 Issued as N.R.C. No. 5658. Paper presented at the IV International Congress of Biochemistry, Vienna, Austria, September 1-6, 1958. Abstract.s, p. 27.

Skins of the cod (Gadus callarias) were collected at the filleting machine in a commercial plant, frozen and kept at -20%. They were freed of adhering muscular and connective tissue by careful scraping and cut into small pieces (about 5 X 10 mm.) with scissors. All manipulations were carried out in the cold room or in refrigerated centrifuges at about 4’C. METHODS The skin was analyzed for moisture by drying in a vacuum oven at 105”C., for ash by ignition in a muffle furnace at 600°C., for total lipid by extraction in a Soxhlet apparatus, for organic nitrogen by micro-Kjeldahl, for hydroxyproline by the method of Neuman and Logan as modified by Bowes, Elliott, and Moss (6). Other methods for the analysis of collagenous preparations were as follows: total protein by quantitative biuret reaction (7), hexosamine by the Elson-Morgan reaction (8)) amino acids by the method of Moore and Stein (9) after hydrolysis for 48 hr. with 6 N HCl in a sealed tube at 105’C.,




pH by the glass electrode, electrophoresis in a Klett Tiselius-Longsworth apparatus, sedimentation in a Spinco model E ultracentrifuge, optical rotation in a Rudolph high-precision polarimeter, viscosity in Ubbelohde capillary viscometers, and electron microscopy with a Metropolitan-Vickers Model EM3A microscope.




The skin was extracted successively with water, aqueous 1O70 NaCl (adjusted to pH 8.0)) water to remove the chloride, and lastly 0.1 M Siirensen citrate buffer (pH 3.5 f 0.1) until a negligible amount of protein was dissolved. This procedure was essentially that used for ichthyocol from cod swim bladder (1). Separation of the citrate extracts was achieved by spinning at 20,000 r.p.m. for 1 hr. in a Spinco model L centrifuge. The gelatinous residue was next extracted with 0.2 M aqueous Na2HP04 (pH 8.6). This was followed by one extraction with 0.2% Ca(OH)z (pH 13). Finally, after adequate washing, the residue was treated with successive lots of 0.01 M citrate buffer for 4 hr. at 90°C. with mechanical stirring. The grayish black residue was washed with water and dehydrated with absolute ethanol and ether. The citrate extracts obtained at 4°C. were combined and dialyzed at 4°C. against 0.02 M NazHPOl in cellophane tubing until the internal fluid attained pH 7. Fine microscopic needles or fibers of collagen formed at pH 4.9-5.1. The supernatant fluid was dialyzed free of phosphate and freezedried. The moist gelatinous precipitate of collagen was redissolved in citrate buffer (pH 3.6), and the solution was clarified by centrifuging for 1 hr. at 15,000 r.p.m. The collagen was reprecipitated and collected as before. This moist material was kept at 4°C. and put into solution as required. For certain chemical analyses it was washed free of phosphate with water, freeze-dried, and then dried in mcuo over P205 at 23°C. TABLE COMPOSITION


SKIN Percentage


Moisture Lipids Mucopolysaccharides Collagen Other proteins Peptides, amino Ash Total


Natural skin


70.3 0.3 0.2 22.2 3.0 0.9 3.6 100.5

of Total solids

1.0 0.6 74.4

10.0 2.9 12.0



The technique of Gallop (10) for ichthyocol from carp swim bladders was used. It resembled the first procedure except that initial extraction was made with 0.5 M sodium acetate (pH 8.5) in place of chloride. The extracted acid-soluble collagen was precipitated twice. The earlier and later citrate extractions were processed separately.




The second procedure was modified by the use of lactic or tartaric in place of citric acid as solvent. Thus the skin was first extracted with a 0.5 M sodium lactate or tartrate (pH 8.6). After dialysis the precipitated collagen was washed with water and dissolved in a buffer solution of 0.1 M sodium lactate-lactic acid of pH 3.6. These buffers were prepared by adding 167 ml. of 0.2 M NaOH to 333 ml. of 0.2 M lactic or tartaric acid and diluting to 11. They contained no sodium chloride. EXPERIMENTAL COMPOSITION




To follow quantitatively the extraction of collagen and other substancesfrom the skin it was necessary to know its composition. No proximate analysis of whole cod skin could be found in the literature. An analysis was therefore done, and the results are shown in Table I. These figures are based on the following determinations and calculations. Total lipids were extracted from dry skin with anhydrous acetone which gave a value of 0.47 %; with anhydrous ethyl ether, followed by anhydrous ethanol, the value was 1.02 %. Since acetone would not extract most phospholipids and cod flesh is known to contain about half of its total lipids as phospholipids and only 3 % as triglyceride (ll), the figure of 1% is taken as the better value. Ethanol would remove some free amino acids. The mucopolysaccharide was calculated from a determination of total hexosamines on dry skin which had been exhaustively extracted with acetone. The value was 0.24 % from replicate determinations with glucosamine as standard. Maximum color developed after hydrolysis with 2 or 4 N HCl at 100°C. for 16 hr. in a sealedtube. Since some amino acids are known to cause a development, of color in the Elson-Morgan reaction (12), a blank was run with free amino acids



in the amounts known to be the equivalent of the collagen in the sample used for t)he estimat,ion. The blank gave a value equivalent to 0.05% glucosamine. Then since both hyaluronic acid and chondroitinsulfuric acid contain about, 34 % hexosamine, a value of 0.56 % is obtained for mucopolysaccharide. The distribution of nitrogenous mat’erial was calculated from a value of 15.6% for t,ot,al nitrogen, and of 5.95 % for hydroxyproline in t,he dry skin. Since the purified collagen cont,ained 8.0 % hydroxyproline, t,he totma collagen present in the skin would be 74.4 % ; and since it, contained 18.2 % K this mould account for 13.5% of the 15.6% N present. Ot,her nitrogenous matter must, therefore be present. The initial aqueous saline extracts of skin should cont,ain any free amino acids and peptides. Absolute ethanol was added to such ext)racts unt#il a concentration of 75% v/v was attained. The precipitate was filtered off, and the filtrate was found to cont’ain about 3 % of the total X originally present in the skin. Then, by difference, other proteins or complex peptides must account’ for about 10 % of the total solids. This fraction requires furt)her investigation. The value of 12 % for ash in the whole cod skin is higher t.han for descaled skin at 2-5 % and is explicable by the presence of the scales and also probably of the salts of sea wat,er as the skins were not washed. DISTRIBUTION 0F HYDROXTPROLPJE



IN ~~ac~~oss

In t,he procedure of extract,ion t,he dist,ribution of both nitrogen and hydroxyproline was followed quant,itatively as a guide t’o the presence of tot’al protein and of collagen, respect,ively, in t,he extracts and residue. For analytical purposes aliquots of t,he extract,s were clarified when necessary by cent,rifuging. The result,sare shown in Table II. About 87 % of t’he r\’ calculated t,o have been present in the sample of skin was account’ed for in the various fract,ions. The remainder must have been lost in manipulation or the actual sample used contained more K than as calculated. The citrat,e buffer at 4°C. dissolved 57 % of the N present and 62 % of t,he hydroxyproline. Only 17.9 % was dissolved subsequent’ly in hot citrate. About 80 % of






Hypro -_~__

Acetate Water

2 3 14 3 5 -

Citrate Phosphate

Hot nater Insoluble due Total



8.5 7.0

3.4 8.6 7.0


!0.;53 0.031 1.6 1.13056.8'402 0.099 5.0 0.32516.3116 '0.092 4.6

0.0 0.0 61.9 33

5.1 17.9 17.0, 2.6

_----_---_ 1



the hydroxyproline was t,hus ext,racted. The insoluble residue at this &age contained 13.5% N, 2.5% hydroxyproline and 1.6% ash, and constituted only 6.1% of the original solids. On treatment of this material with 40% urea at 100°C. for 24 hr. (13, 14) most of it passedinto solut’ion. This may indicate t’he presence of about 1% elastin in dry cod skin. The initial aqueous extract which contrained about 3 % of the tot,al N showed two boundaries in electrophoresis (Fig. I). This was evident aft,er concentration to small volume and equilibrat,ion against barbiturate buffer (I 0.1, pH 8.6). No collagen was present. The saline ext#ract which contained 6 % of the total N showed one major and two minor boundaries in elect’rophoresis (Fig. I). It contained only small amount*s of hydroxyproline and of hexosamine. The initial citrate extract showed only one asymmetric peak when equilibrated against the citrat.e buffer in which it was dissolved and subjected to elechrophoresisat’ 4°C. The fibrous collagen, after one precipitation, exhibited one sharp major component and also a small percentage of a minor component after 7 hr. of electrophoresis (Fig. 1). ACID-S• LKBLE


Microscopic Appearance The fine needle-like fibers which form at pH 5 on dialyzing a citrate solution of col-





completely reversible only with collagen from the first procedure. Reversal was not dependent only on temperature but also on time, especially in passagefrom high to low temperatures. To attain maximum rotat,ion required from one to several days. The rapid change in rotation which occurred on warming to about 15°C. is seenin Fig. 3. Accurate measurement of the temperature of the solution was not achieved. Concentration of collagen was calculat)ed from the value for iY x 5.5. The solvent was 0.1 M citrate buffer (pH 3.4-3.6) in procedures I and II. The specific rotation, [oc]??,for all preparations was -107 f 2” and on cooling to



i oscending



1. Electrophoretic diagrams of extracts of cod skin. (A) Aqueous extract after concentration and equilibration in barbiturate buffer, I 0.1, pH 8.6, temp. 0.5”C., protein 0.30%; 20 ma., 150 v. for 170 min. (B) Saline extract after concentration and equilibration as in (A), protein 0.76%, 15 ma., 110 v. for 120 min. (C) Citrate extract equilibrated against citrate buffer, I 0.1, pH 3.3, temp. 4"C., protein 0.4870,20 ma., 158 v. for 260 FIG.

min. (D) Collagen,after first precipitation, in citrate buffer, I 0.1, pH 3.3, temp. 4'C., protein 0.39’%, 15 ma., 167 v. for 320min.

it was


=I= 12”. The



of [aID at about 5” varied from -289’ to -403” for different preparations in different solvents and averaged -349 A 50” for five preparations. The latter value is comparable with -350” for ichthyocol from carp swim bladder (lo), and - 415” for collagen from calf skin (2). The evidence from optical rotation confirms previous observations (2, 10, 15) that an irreversible change takes place in the collagen molecule at about IO-15’C. This is discussedbelow. Xolubility

An att’empt was made to det’ermine the lagen against water or dilute alkaline phospoint of minimum solubility with purified phate are shown in Fig. 2. These appear to collagen dissolved in 0.1 M citrate buffer. resemble reconstituted acid-soluble collagen A small volume of the collagen solution was from carp swim bladder (10) or from mamadded to a series of standard phthalate or malian tissues (15, 16). On first precipitation about 85 % of the nitrogen and of the hy- phosphate buffers. After 1 hr. the mixtures droxyproline in the solution appeared in the were centrifuged, the protein in solution precipitate. This rose to 98 % on the second measured by micro-Kjeldahl or the biuret precipitation. The dimensions of the needles reaction, and the pH was redetermined. The were approximately 30 X 1 p. In the elec- region of minimum solubility was at pH tron microscope typical banding was seen in 4.8-5.2 by the biuret reaction and 5.0-5.2 most fibers, unshadowed, mounted on Form- by micro-Kjeldahl estimation. Solubility revar film, at magnifications of 10,000 or more, mained low between pH 5.2 and 6.0 and rose with objective aperture in position. This was slowly from pH 6.2 to 8.0. This may be taken confirmed by shadowing (Fig. 2). Solutions as supplementary evidence that the isoelecof collagen in lactate or tartrate buffers tric point for acid-soluble collagen, never tended to form more amorphous, gelatinous exposed to alkali or heat, is about 5. It is precipitates. comparable with 5.4 for ichthyocol from cod Optical Rotation swim bladder by electrophoresis (1). The Change in specific rotation with tempera- marked dependence of this quantity on the of the solvent and t)he ture was observed in all preparations. It was total ionic strength


FIG. 2. Left: shodowed with



“Crystalline” collagen, ordinary microscope, gold and mounted on Formvar film, X10,000.







Chemical Composition


I1 5

FIG. collagen prepared middle:












3. Change of specific optical with temperature and time by three procedures. Left: procedure 2; right: procedure

rotation of for material procedure I; 3.

divergence between isionic and isoelectric poinbs have been shown by Jackson and Neuberger (17) for mammalian acid-soluble collagen.

Analyses of two homogeneous preparations of collagen after two precipitations a,s described above gave the following results, expressed on a moisture and ash-free basis: K 18.19 and 18.28 %, hydroxyproline 7.95 and 8.13 %, hexosamine 0.02 and 0.10 % (correct,ed). These values are very close t,o those of Doty and Sishihara (2) at, 18.3 % N and 7.0 % hydroxyproline and those of Astrup, Marko, and k’oung (1) for ichthyoco1 from cod swim bladder at 18.3 % N and 8.0 % hydroxyproline. The small amount of carbohydrate found confirms previous results of Moss (12) of 0.05%, and of Astrup et al. (1) of 0.1%. Such low values suggest the presence of a small amount of an impurity. The dist,ribution of amino acids after acid hydrolysis is shown in Table III t,ogether with t.hose for icht,hyocol from cod swim bladder (1). Comparison indicat,es great similarity if not identity. Viscosity All measurements were carried out at 1.4 + O.Ol”C. in Ubbelohde capillary viscometers with flow-times of about 60 sec. and






Amino acid hydrolyzate.


N as percentage

Amino acid

a Corrected 6 Astrup,

N for Marko,

of the


Alanine Arginine Aspartic acid Glutamic acid Glycine Histidine Hydroxylysine Hydroxyproline Isoleucine Leucine Lgsine Methionine Phenylalanine Proline Serine Threonine Tyrosine Valine Amide N Total Total



N in

Swim bladde8

8.6 16.7 3.9 4.2” 25.6 1.7 0.8 4.8 1.2 2.2 4.4 1.0 ? 8.0 4.7” 2.10 ? 1.7 5.06

9.8 15.5 4.0 5.6 27.7 1.3 1.1 4.5 1.1 1.8 3.4 0.8 1.2 8.9 4.8 2.0 0.05 1.3 4.1

96.6 18.2

99.0 18.3

loss by decomposition. and Young (1). TABLE



No. No. No.








12.9 12.8 13.2 12.3

2 (citrate) 3 (lactate) 3 (tartrate)

otropy down to concentrations of about 0.05 %. Two preparations showed intrinsic viscosities of about 0.6 and 1.0 dl./g. Thixotropic behavior was confirmed by measurements in a Stormer viscometer. The gel struct,ure of t,he material was destroyed by very low shear rates, and the reformation was so slow that measurement of the setting time was impracticable. Preparations by procedures 2 and 3 gave solutions of high viscosity which were not thixotropic. Table IV lists values of t.he intrinsic viscosity, [v], dl./g., and the slope 7i given by q,,Jc = [q] + hz. This equation fitted the data up to about c = 0.1%. Above this concentration T&C increased more rapidly with concent’ration. The constants were obtained by t’he met’hod of least, squares. The average value for [q] is 12.8 f 0.3 dl./g. The differences are not significant,. Variations in Ic appear to lie out#sidethe expected experimental error. A series of measurements on three solutions of a single preparation which had been heated was carried out. The concent,rabions employed were all below 0.05 % to avoid t#hixotropic effects. The solutions were heated for 30 min. in the viscometers and then placed at 1.4”C. Flow-times were measured 15 min. later. Intrinsic viscosit’ieswere calculated for each experiment. Figure 4 shows the relation between the temperature of heat’ing for 35 hr., T+ , and



No. 2 (citrate)



186 128 132 97



mean shear rate of about 600 sec.-’ for water. The relative kinematic viscosities were 1.024, 1.021, and 1.020 for the citrate, lactate, and tartrate buffers, respectively, and these were independent of temperature within an experimental error of 0.4 % between 1.4 and 25°C. The values of the reduced specific viscosity, qsp/c, of collagen obtained by procedure 1 increased very rapidly with increasing concentration, and exhibited pronounced thix-

o.2t L I

I 5










4 f

+ 4. Plot of viscosities of collagen soIutions after heating for 34 hr. at various temperatures (abscissa) and the ratio of [q] at T+ to [v] at 1.4’C. (ordinates) for three preparations in different solvents. FIG.


the ratio of [q] at temperature T; to [v] at 1.4”C. for several preparations. The temperature of transformation, T, , was taken to be that value of T+ which corresponds to the mean of t’he values of [q]/[qo] at T+ = 1.4 and 25°C. for t’he citrate preparations. The values for T, were 12.7”C. (cit’rate), 13.3”C. (tartrate) and 14.6’C. (lactate). The viscosities of t.he heated samples increased wit#h time at 1.4”C., so that the above procedure does not give intrinsic viscosities that are strictly correct. The decrease in viscosity near the temperature of transformation is so rapid that the value of T, , and more particularly differences in T, values, are lit,tle affected. .Sedh.entation Collagen prepared by procedure 1 showed two peaks in t,he ultracentrifuge when examined at, 69”C., although the material had never been exposed consciously to temperatures above 10°C. These t)wo peaks could be resolved completely at concentrations below 0.3% by centrifuging at’ 60,000 r.p.m. for 4 hr. Collagen prepared by procedures 2 and 3 showed a single sharp peak, although two preparations comained a small quantity of material with a high sedimentation rate. Sedimentation coefficient,s szoWwere calculated by the usual met’hods (19). Corrections were computed from the viscosities of t’he buffer solutions, density increments (meas-





ured pycnometrically) of 0.0069,0.0049, and 0.0025 g./cc. for citrate, tartrate, and lactate buffers, respectively, and an assumed partial specific volume of 0.700 cc/g. (18). Collagen from procedure 2 was examined over the concentration range of 0.02-0.37 %. Plot’s of l/snow and ~20~. where rlr is the relat,ive viscosity of the solution were linear below 0.15% and gave a mean value of szoow = 3.17 f 0.10 s. After heating at 25”C., collagen from procedure 2 showed t,wo peaks at 6-9°C. in t’he ultracentrifuge. They remained unchanged on storage at 3°C. for 1 week. The change upon heating thus appears t,o be irreversible. A series of experiments in which samples of solution were heated in the ultracentrifuge cell for 30 min. at predetermined temperatures and then cent,rifuged at once at 6-9°C. was carried out. Figure 5 shows the effect of heat,ing at 12.4, 13.7, and 24.8”C. It, is evident that T, lies between 12.4 and 13.7”C., in agreement with the value obt,ained from change in viscosity. Values of l/szoW for the two peaks obtained with mat,erial by procedures 1 and 2 lay on two curves when plotted against, concentration. By extrapolat’ion, the values for sio,,, were 3.57 S (fast>er component) and 1.9, S (slower component). The t’wo components appeared to be present in approximately equal amount,s. The total area under t#he peaks changed considerably and errat.ically

Fro. 5. Photographs of boundaries of collagen preparations dissolved in citrate buffer at pH 3.5 in the ultracentrifuge, spinning at 59,780 r.p.m. (11) At 69°C. after 5 14 hr. Top. C = O.l58LT,. Bott,om. C = 0.106cj0. (H) After heating for 30 min. at 12.4”C. Top. Collagen soln. Bottom. Buffer soln. (C) Top. After heating for 30 min. at, 13.7”C. Bottom. Aft,er heating for 30 min. at 21.8”C., and spinning for 414 hr. at full speed.




with time, probably due to thermal convection or reaggregation on cooling. Calculation of Molecular Dimensions The strong concentration dependence of both the viscosity and the sedimentation coefficient suggestst#hat acid-soluble collagen is a highly asymmetric molecule. The magnitudes of [TJ]and sioware similar to those for carp ichthyocol (18). The axial ratio of the equivalent hydrodynamic ellipsoid was first found to be 188 by insertion of the values 1~1= 12.8 dl./g. and assumed ii = 0.700 ct./g. into Simha’s viscosity equation (20). With this value of the axial ratio and .s& = 2.94 S as the sedimentation constant in 0.1 M citrate buffer, the length of the minor axis was found to be 14.9 A. and of the major axis 2810 A. from the combined Svedberg (19) and Perrin (21) equations. Since the dimensions and density of the ellipsoid are now known, the molecular weight can be calculated to be 280,000 f 20,000. Evidence from light-scattering suggeststhat carp ichthyocol is a rod-shaped molecule (18). A similar molecule of cod collagen with length and specific volume the same as that of the equivalent. hydrodynamic ellipsoid would have a diameter of 12.1 A. The limits of error ascribed to the molecular weight have been computed from the errors m s,“,,and [VJ]given above, and an uncertainty of 0.005 ct./g. in C (10). It has been suggested (18) that Vfor carp ichthyocol may be as much as 0.05 ct./g. greater than the value for carp gelatin which is used here. However, recent dilatometric investigations (22) indicate that there is an increase of in on transformation of collagen to gelatin of a magnitude of only 0.005 ct./g. From the relationship of molecular weight to sedimentation constant found by Williams et al. (23), the faster component has a molecular weight of 2 10,000 and the slower sedimenting component a molecular weight of 110,000 after thermal decomposition of the collagen. If these two components are present in equal proportions, and are each monodisperse, the sum of their molecular weights should be equal to the molecular weight of the parent collagen. This appears to be the case.



The small amount of mueopolysaccharide in cod skin under the conditions of our examination suggeststhat its function may be external rather than internal and not related to the collagen matrix or combined chemically with collagen and possibly in the form of a glycoprotein (24). The characteristics of the acid-soluble C& lagen in cod skin indicate its great similarity, and probable identity with ichthyocol from the swim bladder. Hydroxyproline in cod skin could be extracted with cold citrate buffer, pH 3.5, to the extent of 62 % of that present. An appreciable amount was liberated by subsequent treatment with hot water and a small amount appeared to be present as elastin. This suggeststhe presence of hydroxyproline in three protein entities in cod skin, but the large amount extractable by citrate negatives its suggestedrole as a true procollagen or precursor of the insoluble collagen. Bald, Banga, and Szabo (25) have designated as “metacollagen” the portion of the collagen fiber which is not extractable by acid citrate buffer and have pointed out its similarity to elastin. A difference between cod and carp collagensis indicated by differences in sedimentation constants (s!& of 3.17 vs. 2.85 S), thermal stability, (Tu of 13” vs. 29’) and content of hydroxyproline (8 % vs. 11%). A greater difference is apparent between collagen from fish and from mammalian tissues in content of hydroxyproline, the value of T, , and the molecular dimensions, although the latter are still in some uncertainty. The change in cod collagen which takes place at 13 f 0.5” is similar to that observed in carp collagen (10) and mammalian collagen (26). This change has been referred to as denaturation and is considered to be irreversible. It is accompanied by marked lossin specific optical rotation and intrinsic viscosity. Two components are detectable in the ultracentrifuge and by electrophoresis. These facts are confirmed in the present investigation. Acid-soluble collagen by procedure 1 was already in this state although it had never been exposed knowingly to a temperature above 10”. This may have been due to contact with 1.7 M NaCl as in the case of


mammalian procollagen (27). The temperature of change is astonishingly sharp, similar to a melting point. Two separable components appear with the disappearance of the original collagen. The isolation of the two components by Shpikiter (28) should lead to a better understanding of this phenomenon. The preparations of collagen made with citrate, tartrat’e, and lactate were very similar except that’ the values for [vJ], [v]/[v]o , and TD indicated a slight difference for that prepared with lactate buffer. The value for [a$ was another difference since that’ with lactate was -403”, with tartrate - 389”, and with citrate - 354”. The reversible change in t’he specific rotation was reasonably consistent between [a];’ at -223 & 12” and [cx]~” at - 107 f 2”. This is in substantial agreement with the observations of Cohen (29) on carp ichthyocol. The time required to attain t,he maximum rotat’ion on cooling indicates a slow molecular orientation as in crystBallisation or gel formation. This phenomenon has been known for many years as applied to gelat,in. ACKNOWLEDGMENTS We gratefully acknowledge the able technical assistance of F. G. Mason, D. Johnson and Miss J. Macpherson, and the stimulating interest of Dr. H. M. Sonnichsen. We are indebted to PermacelLePage’s Inc. for a financial grant, to Mr. E. I’. Sidaway of General Sea Foods Corp. for a supply of cod skins, to Mr. D. G. Smith for the determination of the distribution of amino acids, and to Mr. E. Copps for the electron microscopy. REFERENCES 1. ASTRUP, H. N., MARKO, A. M., AND YOUNG, E. G., in “Recent Advances in Gelatin and Glue Research” (G. Stainsby), p. 76. Pergamon Press, London, 1958. 2. DOTY, P., AND NISHIHARA, T., in “Recent Advances in Gelatin and Glue Research” (G. Stainsby), p. 92. Pergamon Press, London, 1958. 3. GUSTAVSON, K. H., Nature 176, 70 (1955). 4. TAKAHASHI, T., AND YOKOYAMA, W., Rull. Japan Sot. Sci. Fisheries 20, 525 (1954). 5. GUSTAVSON, K. H., Acta Chem. Stand. 8, 1298 (1954).





6. BOWES, J. H., ELLIOTT, R. G., AND Moss, J. A., Biochem. J. 61, 143 (1955). 7. GORNALL, A. G., BARDAWILL, C. J., AND DAVID, M. M., J. Biol. Chem. 177,751 (1949). 8. RONDLE, C. J. M., AND MORGAN, W. T. J., Biochem. J. 61, 586 (1955). S., AND STEIN, W. H., J. Biol. Chem. 9. MOORE, 192, 663 (1951). P. M., Arch. Biochem. Biophys. 64, 10. GALLOP, 486,501 (1955). J. A., AND OLLEY, J., 11. GARCIA, M. D., LOVERN, Biochem. J. 62, 99 (1956), J. A., Biochem. J. 61, 151 (1955). 12. Moss, 13. HALL, D. A., Nature 168, 513 (1951). R. E., 14. HALL, D. A., REID, R., AND TUNBRIDGE, Nature 170, 264 (1952). V. N., TOUSTANOVSKI, A. A., 15. OREKHOVITCH, OREKHOVITCH, K. D., AND PLOTNIKOVA, N. E., Biokhimiya 13, 55 (1948). 16. JACKSON, S. F., AND RANDALL, J. T., in “Nature and Structure of Collagen” (J. T. Randall), p. 181. Butterworth’s Scientific Publications, London, 1953. A., Biochim. 17. JACKSON, D. S., AND NEUBERGER, et Biophys. Acta 26,638 (1957). 18. BOEDTKER, H., AND DOTY, P., J. Am. Chem. Soc.78,4267 (1956). 19. SVEDBERG, T., AND PEDERSEN, K. O., “The Ultracentrifuge,” pp. 5, 36. Oxford Univ. Press, Oxford, 1940. 20. SIMHA, R., J. Phys. Chem. 44, 25 (1940). 21. PERRIN, F., J. phys. radium [7] 7, 1 (1936). 22. FLORY, P. J., AND GARRETT, R. R., J. Am. Chem. Sot. 80, 4836 (1958). 23. WILLIAMS, J. W., SAUNDERS, W. M., AND CICIREXI, J. S., J. Phys. Chem. 68, 774 (1954). 24. WESSLER, E., ANU WERNER, I., Acta Chem. Stand. 11, 1240 (1957). 25. BALK, J., BAN(:A, I., AND SZAB~, D., Magyar Tudomzinyos Akad. Biol. Bs Orvosi I’udomdnyok OsztCllyClnak Kiizlemenyei 7, 385 (1956); C. A. 63,467 (1959); ibid. 8, 229 (1957); C. A. 63, 1428 (1959). 26. OREKHOVITCH, V. N., AND SHPIKITER, V. O., Doklady Akad. Nauk S.S.S.R. 101(3), 529 (1955); Science 127, 1371 (1958). 27. BRESLER, 8. E., FINOGENOV, P. A., AND FRENKEL, S. I., Doklady Akad. Nauk S.S.S.R. 72 (3),555 (1950). 28. SHPIKITER, V. O., Doklady Akad. ~Yauk S.S.S.R. 116, 179 (1957). 29. COHEN, C., J. Biochem. and Biophys. Cytol. 1, 203 (1955).