Vol. 6i DETERMINATION OF GLUCOSAMINE AND GALACTOSAMINE 589 hyde, but the most serious errors arise when small quantities of amino sugar are estimated in the presence of large quantities of mixtures of amino acids with non-nitrogenous sugars (Palmer et al. 1937; Sideris, Young & Krauss, 1938; Lutwak- Mann, 1941; Yasuoka, 1944; Horowitz, Ikawa & Fling, 1950; Immers & Vasseur, 1950; Storey, Yensen, Lisie & Bilin, 1951). For this reason it is advisable to confirm determinations of 'total amino sugar' by alternative means such as that proposed by Gardell (1953). An account of a method whereby such a confirmation can be obtained will be given later. SUMMAlRY 1. A modification of the method described by Elson & Morgan (1933) for the determination of amino sugars is given. 2. The 'best' range, the accuracy and the reproducibility of the determination are illustrated by results obtained by the use of authentic amino sugar samples, and the limitations of procedure, imposed by the instability of the chromophore produced, are given. 3. Factors of importance in the application of the method to naturally occurring substances are discussed. The authors wish to thank the Medical Research Council for the award of a Studentship (1951-4) to C.J.M. R. REFERENCES Adams, R. & Coleman, G. H. (1948). Org. Synth. (Collected), 2nd ed. 1, 214. Aminoff, D., Morgan, W. T. J. & Watkins, W. M. (1950). Biochem. J. 46, 426. Anastassiadis, P. A. & Common, R. H. (1953). Canad. J. Chem. 31, 1093. Annison, E. F. & Morgan, W. T. J. (1952a). Biochem. J. 50, 460. Annison, E. F. & Morgan, W. T. J. (1952b). Biochem, J. 52, 247. Ayres, G. H. (1949). Analyt. Chem. 21, 652. Belcher, R., Nutten, A. J. & Sambrook, C. M. (1954). Analy8t, 79, 201. Blix, G. (1948). Acta chem. scand. 2, 467. Boyer, R. & Fiirth, 0. (1935). Biochem. Z. 282, 242. Brownlee, K. A. (1949). Industrial Experimentation, 4th ed. London: H.M.S.O. Cessi, C. (1952). Boll. Soc. ital. Biol. sper. 28, 858. Dakin, H. D., Ungley, C. C. & West, R. (1936). J. biol. Chem. 115, 771. Dakin, H. D. & West, R. (1935). J. biol. Chem. 109, 489. Dakin, H. D., Young, H. Y. & Krauss, B. H. (1938). J. biol. Chem. 126, 233. Elson, L. A. & Morgan, W. T. J. (1933). Biochem. J. 27, 1824. Gardell, S. (1953). Acta chem. 8cand. 7, 207. Gibbons, R. A. & Morgan, W. T. J. (1954). Biochem. J. 57, 283. Hadidian, Z. & Pirie, N. W. (1948). Biochem. J. 42, 260. Hamasato, Y. & Akakura, K. (1941). J. Biochem., Tokyo, 34, 159. Hewitt, L. F. (1938). Biochem. J. 32, 1554. Horowitz, H. N., Ikawa, M. & Fling, M. (1950). Arch. Biochem. 25, 226. Immers, J. & Vasseur, E. (1950). Nature, Lond., 165, 898. Johnston, J. P., Ogston, A. G. & Stanier, J. E. (1951). Analy8t, 76, 88. Jorpes, E. (1935). Biochem. J. 29, 1817. Jorpes, E. (1942). Biochem. J. 38, 203. Lustig, B. & Ernest, T. (1937). Biochem. Z. 289, 365. Lutwak-Mann, C. (1941). Biochem. J. 35, 610. Masamune, H. & Nagazumi, Y. (1937). J. Biochem., Tokyo, 26, 263. Nilsson, I. (1936). Biochem. Z. 285, 386. Ogston, A. G. & Stanier, J. E. (1950). Biochem. J. 48, 364. Palmer, J. W. & Meyer, K. (1935). J. biol. Chem. 109, lxxiii. Palmer, J. W., Smyth, E. M. & Meyer, K. (1937). J. biol. Chem. 119, 491. Pauly, H. & Ludwig, E. (1922). Hoppe-Seyl. Z. 121, 176. Schloss, B. (1951). Analyt. Chem. 23, 1321. Sideris, C. P., Young, H. Y. & Krauss, B. H. (1938). J. biol. Chem. 126, 233. S0rensen, M. (1938). C.R. Lab. Carlsberg, 22, 487. Storey, Z., Yensen, M., Lisie, S. & Bilien, M. (1951). Rev. Fac. M64. Univ. Itanbul, 14, 392. West, R. & Clark, D. H. (1938). J. clin. Inve8t. 17, 173. Yasuoka, T. (1944). Tohoku J. exp. Med. 46, 260. The Amino Acid Composition of Mammalian Collagen and Gelatin BY J. E. EASTOE The Briti8h Gelatine and Glue Research A88ociation, 2a, Dalmeny Avenue, London, N. 7 (Received 29 April 1955) The name collagen is given to a grqup of related fibrous proteins, found in the mesodermal tissues of animals. The best-characterized forms of collagen occur in the higher animals, where they constitute the main organic components of skin, bone, tendon and loose connective tissue. Gelatin, a derived protein, can be prepared as a breakdown product of collagen by extracting these tissues with hot water (above 400). Extraction with minmal degradation can normally be accomplished in neutral solutions after aprolonged pretreatmentwith cold alkali or by an acidextraction (Ward, 1954). Gelatin istheprinci- pal constituent ofthe commercialproducts 'gelatine' and glue and is responsible for their gel properties.
Transcript
Vol. 6i DETERMINATION OF GLUCOSAMINE AND GALACTOSAMINE 589hyde, but the most serious errors arise when smallquantities of amino sugar are estimated in thepresence of large quantities of mixtures of aminoacids with non-nitrogenous sugars (Palmer et al.1937; Sideris, Young & Krauss, 1938; Lutwak-Mann, 1941; Yasuoka, 1944; Horowitz, Ikawa &Fling, 1950; Immers & Vasseur, 1950; Storey,Yensen, Lisie & Bilin, 1951). For this reason itis advisable to confirm determinations of 'totalamino sugar' by alternative means such as thatproposed by Gardell (1953). An account of amethod whereby such a confirmation can beobtained will be given later.
SUMMAlRY1. A modification of the method described by
Elson & Morgan (1933) for the determination ofamino sugars is given.
2. The 'best' range, the accuracy and thereproducibility of the determination are illustratedby results obtained by the use of authentic aminosugar samples, and the limitations of procedure,imposed by the instability of the chromophoreproduced, are given.
3. Factors of importance in the application ofthe method to naturally occurring substances arediscussed.The authors wish to thank the Medical Research Council
for the award of a Studentship (1951-4) to C.J.M. R.
REFERENCESAdams, R. & Coleman, G. H. (1948). Org. Synth. (Collected),2nd ed. 1, 214.
Aminoff, D., Morgan, W. T. J. & Watkins, W. M. (1950).Biochem. J. 46, 426.
Anastassiadis, P. A. & Common, R. H. (1953). Canad. J.Chem. 31, 1093.
Annison, E. F. & Morgan, W. T. J. (1952a). Biochem. J. 50,460.
Annison, E. F. & Morgan, W. T. J. (1952b). Biochem, J. 52,247.
Ayres, G. H. (1949). Analyt. Chem. 21, 652.
Belcher, R., Nutten, A. J. & Sambrook, C. M. (1954).Analy8t, 79, 201.
Blix, G. (1948). Acta chem. scand. 2, 467.Boyer, R. & Fiirth, 0. (1935). Biochem. Z. 282, 242.Brownlee, K. A. (1949). Industrial Experimentation, 4th ed.London: H.M.S.O.
Cessi, C. (1952). Boll. Soc. ital. Biol. sper. 28, 858.Dakin, H. D., Ungley, C. C. & West, R. (1936). J. biol.Chem. 115, 771.
Dakin, H. D. & West, R. (1935). J. biol. Chem. 109, 489.Dakin, H. D., Young, H. Y. & Krauss, B. H. (1938).
J. biol. Chem. 126, 233.Elson, L. A. & Morgan, W. T. J. (1933). Biochem. J. 27,
1824.Gardell, S. (1953). Acta chem. 8cand. 7, 207.Gibbons, R. A. & Morgan, W. T. J. (1954). Biochem. J. 57,
283.Hadidian, Z. & Pirie, N. W. (1948). Biochem. J. 42, 260.Hamasato, Y. & Akakura, K. (1941). J. Biochem., Tokyo,
34, 159.Hewitt, L. F. (1938). Biochem. J. 32, 1554.Horowitz, H. N., Ikawa, M. & Fling, M. (1950). Arch.
Biochem. 25, 226.Immers, J. & Vasseur, E. (1950). Nature, Lond., 165, 898.Johnston, J. P., Ogston, A. G. & Stanier, J. E. (1951).
Analy8t, 76, 88.Jorpes, E. (1935). Biochem. J. 29, 1817.Jorpes, E. (1942). Biochem. J. 38, 203.Lustig, B. & Ernest, T. (1937). Biochem. Z. 289, 365.Lutwak-Mann, C. (1941). Biochem. J. 35, 610.Masamune, H. & Nagazumi, Y. (1937). J. Biochem.,
Tokyo, 26, 263.Nilsson, I. (1936). Biochem. Z. 285, 386.Ogston, A. G. & Stanier, J. E. (1950). Biochem. J. 48, 364.Palmer, J. W. & Meyer, K. (1935). J. biol. Chem. 109, lxxiii.Palmer, J. W., Smyth, E. M. & Meyer, K. (1937). J. biol.Chem. 119, 491.
Pauly, H. & Ludwig, E. (1922). Hoppe-Seyl. Z. 121, 176.Schloss, B. (1951). Analyt. Chem. 23, 1321.Sideris, C. P., Young, H. Y. & Krauss, B. H. (1938).
J. biol. Chem. 126, 233.S0rensen, M. (1938). C.R. Lab. Carlsberg, 22, 487.Storey, Z., Yensen, M., Lisie, S. & Bilien, M. (1951). Rev.
Fac. M64. Univ. Itanbul, 14, 392.West, R. & Clark, D. H. (1938). J. clin. Inve8t. 17, 173.Yasuoka, T. (1944). Tohoku J. exp. Med. 46, 260.
The Amino Acid Composition of Mammalian Collagen and Gelatin
BY J. E. EASTOEThe Briti8h Gelatine and Glue Research A88ociation, 2a, Dalmeny Avenue, London, N. 7
(Received 29 April 1955)
The name collagen is given to a grqup of relatedfibrous proteins, found in the mesodermal tissues ofanimals. The best-characterized forms of collagenoccur in the higher animals, where they constitutethe main organic components of skin, bone, tendonand loose connective tissue. Gelatin, a derivedprotein, can be prepared as a breakdown product of
collagen by extracting these tissues with hot water(above 400). Extraction with minmal degradationcan normally be accomplished in neutral solutionsafter aprolonged pretreatmentwith cold alkali orbyan acidextraction (Ward, 1954). Gelatin istheprinci-palconstituent ofthe commercialproducts 'gelatine'and glue and is responsible for their gel properties.
J. E. EASTOE
There are no rigid distinguishing features ofproteins belonging to the collagen group. Bear(1952) used certain characteristics of the wide-angle X-ray diffraction pattern as the main criteria,supported, except in the collagen of the simplestorganisms (e.g. spongin), by the existence of longspacings (approx. 6001), detectable either by small-angle X-ray diffraction or electron microscopy.Proteins falling within this classification are verywidely distributed in many animal phyla, and forsome purposes it appears useful to subdividefurther this large group, on a chemical basis,distinguishing those members which containsignificant numbers of hydroxyproline residues.Hydroxyproline and hydroxylysine are not knownto occur in substantial quantities in proteins otherthan those of the collagen type. The restricteddefinition would include the characteristic meso-dermal protein of vertebrates but might excludemany of the invertebrate types, since none of thesehas yet been shown to contain hydroxyproline. Inaddition, only vertebrate collagens have, so far,been shown to be capable of yielding a degradationproduct, a solution of which gels on cooling. Thisprovides further- support for distinguishing verte-brate collagens as special members ofa wider group.
Complete amino acid analyses by chemicalmethods are available only for mammalian collagenand gelatin. Chibnall (1946), with his co-workersand later Bowes & Kenten (1948), have built upaccurate data for gelatin and ox-hide collagen,using a wide variety of analytical techniques. Thesevalues, with recent minor adjustments, have beensummarized by Tristram (1953). The amino acidcomposition of collagen and gelatin from a widerrange of vertebrates has been studied by Neuman(1949), using microbiological methods. Theseresults as a whole represent a higher order ofaccuracy than the earlier work, details of whichhave been summarized by Bowes & Kenten (1947).
The present survey of the composition of mam-malian collagens and gelatins has been made afterusing the ion-exchange chromatographic method ofMoore & Stein (1951). Work is still in progress onthe remaining classes of vertebrates.
EXPERIMENTAL
MaterialsBone collagen (ox and human). Collagen, substantially
free from mucopolysaccharide-protein complex, wasprepared from compact bone from the centre of the dia-physes of femora by the method of Eastoe & Eastoe (1954).(Moisture content: ox collagen, 17-0, human collagen,15-7 %; ash: 0-08 and 0.04% respectively.)
Wallaby tail tendon. The tail was skinned and the fourmassive tendon bundles were pulled from the attachedvertebrae. Small fragments of bone and adhering musclefibres were removed mechanically. The separated tendonswere cut into 1 cm. lengths and washed for 48 hr. in fourchanges of 10% aqueous NaCl at 00. The tendon waswashed with siix changes of water, dehydrated with acetone,transferred to ether and finally air-dried. (Moisture, 17-7;ash, 0-09 %.)Human Achilles tendon. The tendons were treated with
ethanol and the fat was removed with changes of ether.Adhering muscle tissue was cut away, the tendon bundleopened out and as much as possible of the yellow sheathdetached. Smaller fibre bundleswere separated and cut into1 cm. lengths. (Total N, 17-88; ash, 0.82%.) The tendonwas then treated successively with 10% NaCl (two changes,24 hr. each at 00) and saturated aqueous Na2HPO4 (threechanges, 24 hr. each at 0°). The material was washedthoroughly with distilled water, transferred successively toacetone and ether and air-dried. (Total N, 18-14; ash,0-05%.)The phosphate-treated tendon was extracted for 1-5 hr.
with four times its weight of 0-1N-HC1 at 1000. Most of thetendon went into solution. The solution was filtered throughglass wool and evaporated to dryness in vacuo. The residuefrom evaporation was allowed to reach equilibrium withmoisture in the atmosphere and used for the amino acidanalysis. (Moisture content, 15-0; ash, 0.04%.)
Table 1. Moisture and ash contents, isoionic points and physical properties of gelatins
PretreatmentOx hide I
(a) First extraction; lime (3 months)(b) Third extraction; lime (3 months)
* Determined according to B.S. 757:1944 at 6-67% (w/w) concentration of gelatin with specified moisture and ashcontent and at pH 6-0-6-5.
t Gel matured at 10-00 for 17 hr.
590 I955
MAMMALIAN COLLAGEN AND GELATINReticulin. Reticulin from the mid-cortical zone of
human kidney, prepared by the method of Windrum, Kent& Eastoe (1955), was analysed.
Commercial gelatins. Six samples of commercial high-grade gelatin, manufactured from known raw materials,were selected for this study. The gelatins designated ox-hide gelatin I a and I b were from first- and third-extractionliquors respectively of a single batch of alkali-treatedprecursor. The bone gelatin had been prepared from decalci-fied 'Indian bone', which probably consisted mainly ofcattle bones. The raw materials, methods ofpretreatment,moisture and ash contents, isoionic points and character-istic physical properties of these gelatins are summarized inTable 1.
Purified gelatin. Ox-hide gelatin II was purified byfractionation with ethanol (Stainsby, 1955), followed bydeionization with mixed-bed resins (Janus, Kenchington &Ward, 1951). Approximately 12% of the material from thehigh-viscosity end was removed from solution as a co-acervate in three separations. The bulk of the material(83 5 %) was then separated, leaving some 4-5% of low-viscosity material in solution. The middle fraction wasdissolved in more water at 600 and deionized; the solutionset to a gel, which was dried in a current of air at roomtemperature.
MethodsMoisture. The loss of weight on drying the finely divided
collagen or a thin ifim of gelatin for 18 hr. at 1050 was usedto calculate the moisture content.
Ash. The sample was slowly incinerated, at a low temper-ature, in a platinum dish over a Bunsen flame and theresidue heated for 2 hr. at 5500 in an electric mufflefurnace.
Total N. This was determined by the micro-Kjeldahlmethod of Chibnall, Rees & Williams (1943), as modified byEastoe & Eastoe (1954).
Isoionic point. The isoionic points of gelatins were deter-mined by the deionization method of Janus et al. (1951).
Hydrolysis. The protein (0-2 g.) was hydrolysed for24 hr. or 48 hr. in 20 ml. of 20% (w/w) HCI in a sealedresistance-glass tube immersed in water at 1000. The acidwas removed by evaporation in vacuo at room temperatureover P205 and NaOH. The dry residue could be stored at00 over P205 for at least 2 months without change.
Chromatographic analysis. Separation of the amino acidswas carried out on columns of Dowex 50 by the method ofMoore & Stein (1951). The 1 ml. effluent fractions wereanalysed by the method of Moore & Stein (1948), 2 ml. ofninhydrin reagent being used per fraction. Details ofminor alterations to the published procedures are describedin the Appendix (p. 601).
Hydroxyproline. The hydroxyproline peak overlappedthat of aspartic acid, the deep-blue colour of the latterpreventing measurement of the pale yellow produced bythe imino acid. Hydroxyproline was therefore determinedindependently, without preliminary chromatographicseparation by the method of Neuman & Logan (1950), bymeans of the Martin & Axelrod's (1953) modification forremoving hydrogen peroxide. To minimize errors due tolack of reproducibility between replicate samples orstandards, determinations were carried out in duplicate ineach of two separate experiments. The calibration wascarried out on a chromatographically purified sample ofhydroxyproline.
CorrectionsColour yields. The values for 'colour yield' used in the
present calculations were those of Moore & Stein (1948,1951), which were confirmed under the present experi-mental conditions, except for those of lysine and proline,which were found to be 1-08 (at 570 m,.) and 3-52 (at440 miz.) respectively. A value of 1-08 was found forhydroxylysine, and the value 1-02 (Schram, Dustin, Moore& Bigwood, 1953) was assumed for methionine sulphoxide.
Methionine. The methionine content was calculated bycombining values for the methionine peak and the smallpeak emerging immediately before hydroxyproline; thislatter peak was assumed to be methionine sulphoxide,formed from methionine during the acid hydrolysis(Schram et al. 1953). No evidence was obtained for theidentity of this peak, other than its reported position ofemergence.
Olutamic acid. The values for glutamic acid have beencorrected for an assumed 3% loss, probably resulting frompyrrolidonecarboxylic acid formation during passagethrough the column (Moore & Stein, 1951).
Serine, threonine and amide nitrogen. The values forserine and threonine were corrected for decomposition by5 and 3% respectively with hydrolysis for 24 hr. and by10 and 5% respectively with hydrolysis for 48 hr. at 1000.These corrections were adopted after experiments inwhich serine or threonine was heated with 20% (w/w) HCIat 1000. The results obtained were in agreement with thedifferences in the weights of these amino acids recoveredfrom gelatin after 24 and 48 hr. hydrolysis at 1000. Lossesof both serine and threonine after 24 hr. were almostexactly one-half of those found by Rees (1946), who usedHCI of the same concentration, and heated for the sametime but at a higher temperature, 108.50. This indicatesthat the rate of decomposition increases by a factor ofslightly more than 2 for a rise in temperature of 100.The value for amide nitrogen was calculated from the
ammonia peak after correcting for ammonia formed bydecomposition of serine and threonine during hydrolysis.Rees (1946) has shown that the decomposed hydroxyamino acids give a quantitative yield of ammonia.
Overlapping peaks. Where slight overlaps occurredbetween adjacent peaks, as in the pairs aspartic acid andthreonine, threonine and serine, and isoleucine and leucine,the colour intensity ofthe fraction at the minimum betweenthe peaks was divided equally between them, all othervalues being placed in their respective peaks. The con-siderable overlap between tyrosine and phenylalanineinvalidated this method of calculation, and a procedureinvolving graphical reconstruction of the separate peakswas substituted (see Appendix). No attempt was made toestimate separatelythe small amount of ornithine occurringin the ox-hide gelatins.
RESULTSThe values obtained for the amino acid contents ofmammalian collagen, gelatin and reticulin samplesare given in Tables 2-4. Results have been calcu-lated on the weight of dry, ash-free protein(Tables 2 and 3) and on the total nitrogen content(Table 4). The amino acid composition of humanrenal reticulin (Windrum et al. 1955) has been
Vol. 6I 591
J. E. EASTOE
oo c oo
00 CO 0 N -
N 0
00
C <: Co- :c
00 t
C. .) . ..6c 646
00 QCit: 0 N 0
66466
o 0sNt N'
~ 00-
- 0
NN 0-
- t
0 o
0: +
0 CO
o COCO-
0
-0
01
~ -
01
01
CO-
0oCO N
-0
46
CO r46.
01 00
00CO00 00~~~(CO
N 0 01. -
o esC oo _)
4 _ i_ : 0
-" - 0- 0 00
i >o0
0 _ i: _N 00o
00 C
o 0 0C 0 -_t i: CO - o t
0: 0
0 01
CO 0C Ni:0N
CO
0- 0
>
0
-: 00z_-
1 _
I: - 0 N i:z
01
-0- 0 0
-> - 0 - 0 0
s - -N 0 0-- - 0 - 0
=t=
n_10141C;t
_ _~ri6 6; (:: 6:it o: I os:
- - O0-
0C1; s4 -
;>*c_e ce_ -4._c > X O X C no X X
- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
0 Z Ca~~~~~~~~~
5: )
a ., c.. -l .5 Q o E f &,, E _ c 3
592 I955
ce 0
0 -~
Ca0
0 C0
zV01
C0)
Cl
0 Q"0.C
= t.
0 X
- I~
0 ;
br
r 1r- 01
0 t1
_ N
- 01
C 01
- N- 01
- N
- N1
01- N
000
0 t1
_ *e- 01
- oi
OOit C
01 CO - n 01
- it it 01 sCO CO _ it 01
N 0 0 0
01 CO - i: 01
01C CO 01
' i0rN 01
01 CO - 00 01
1) CO CO 1 it:
01 CO - it: 01
r CO 01i: cOit CO _ CO 01
00 CO N: : 01
0>1 CO - 00 01
i: CO ^.' 01 01v:
01 CO _ ^1
it : oCOr
CO1- 00 1
Xit 01 N 01-01 CO - 00 01
oi:00 COX0 1
it: 01 it
01 _ 01
N: N -
O CO 0
CO CO 00
01 N CO
0 - 01
01 01 010 .I 0
itNo
N 1 0
it: 0 it
CO CO 01
Int?1o
0
cr
CO
0
;
0
1
I?
0o=
C
00
x
Co
c.
00
0)0.0
-. p
0
0)00 0).1.
0)C)o -.
0 ?-.0)
0
0
0 o
000 -Cl
C)00) C)O
Cl
0
0)C)
01 -Cl
.0
H
0)0
o50
-0)
.0ii. +.0)0) .A ii)
0) Cl 00) 0)"01) ..0)0
"0 0.z.i*i .00-. 0.- 0.i. .0 rICO"0 .
.0 w00)0)1).. .
.
."0 ,.00)0 .o 0)4.. .- 0
001.1
0.1. 0)000
". -. 0 0 .
01. -. Cl i. -
"0 .0)n) Cl0).0) 00)0).0)0)
0) il*.0 01 C) -. 0 .0QHCOHOO* *1..1..I.0QI.1=.
Cl
0)-
._
=1
Clz06;
oc,It
11 I
MAMMALIAN COLLAGEN AND GELATIN
a °)
6
a 0
Ca00
4Ca)
" 5 -
0 (Da 0c
5 co
Ca)
S lP. rO bi
CDMX @ .
bO X
CO
CO
CO
CO
CO
N
CO
CO
N
10
CO
0110
CO
cO0o
CO
00
co
CO
Ce
co
00
00coCO
4 oo (6
0101
CP10 C
N-O 0
01 01_
00 10 C
N CO
O O C
0-1 01
q 0 CO
CO 10 aO
0 01
o 0o -CC_10
01
cq
0-
01CO
COCO~
01~rCcq-
0
CO
0
P-
CO
0
6
C-,-d
P-
P-
P-
P-
Oe04
CO CO 10 CO- COb 0
10 4 CO 0-
10-10 P--
- . . 1
e CO
t- CO
ch 6 c
- ~-
10 0f 10 to
Co 4 ob ob
- CO
CO - 0)C
- C5)O-
- t -
CO- - C
CO_ CO CO
_ C O-o
0oC mCO
I N o10 6 o
;4 C> 6
100 01 10p< o010 00 1
100
10 10 01 0-C>4n Ca eB do ce
0000cqt
4 tb 4o6
CA O 10 aq CO in
I CO 00 00
0010) m
I6 4 CQ
10 C O O
0)00 C> 1-
o m 1
COI 0ON
x 4 q
CO _)
N 0)
00 Ot 0)
COO
NP CO
-CO
CS O6._ 000 O
P-
* O
NO
00
NO
CO4
N N
N:O
N 0)N 0)
oo 0
00 0)4
N 0
CO0
ON 0
N 0
N 0
0
O
r-
CO CO
Cp
000O-
CC6-4
00
COO
c-c
N
60
0-
O.
6400
-
N CO
tbdD
CO
0
P--
00 0)
0)
O
r-
P-
00
0)
. 01
P-
O 1
CO CO
0 CO
0)C0O
0
0.
CO
CO01 o5.
-a)
0 -i0)
a a45
-P
a) 0
C o
o~ oCa
000 00
.. t
N w
-10* *I++00
-a
0
-~
co
a)~~~~~ 4
"4B E5 61
Bioch. 1955, 6138
Vol. 6i 593
R00
a)
..q
0
Z*0
co0
0
0)
.S
0
Q)
.190
P,
a)114E
.4
ea)
*40
00a~)
0
ae
O
a)
0
a)0.a)
a)
594 J. E. I
included in Table 4 to facilitate comparison withvalues for human collagen, which were not pre-viously available. The reticulin figures are basedsolely on nitrogen content owing to the sub-stantial proportion of fatty acid and carbohydratepresent.
Table 4. Amino acid composition of humancollagen and reticulin
Results are expressed as g. of amino acid nitrogen/100 g.
* Corrected for decomposition during hydrolysis (seetext).
t Trace (see text).i Sum of methionine and methionine sulphoxide peaks.§ Corrected for ammonia formed by decomposition of
serine and threonine.11 Including 0 07% of glucosamine nitrogen.
Complete recovery of nitrogen (100-1-101-8%),within the experimental error, was obtained for all
collagen and gelatin samples (Table 2). The
tendency for this figure to be persistently high hasalso been reported for other proteins (Smith &
Stockell, 1954; Smith, Stockell & Kimmel, 1954).The effect in the present study is probably attribut-able to small errors in the colorimetric factors, andpossibly a slight underestimation of the totalnitrogen content. The recovery of weight rangedfrom 95-6 % for human Achilles tendon to 100.5%for ox-hide gelatin II. The difference between theweight and nitrogen recovery figures represents thenon-nitrogenous constituents, but it can only beregarded as approximate since it reflects, in
EASTOE I955addition, the combined experimental errors ofthe moisture, ash and total nitrogen determina-tions.The chromatograms of all samples qualitatively
resembled those for ox-hide gelatin II on both 100and 15 cm. columns of Dowex 50 (Figs. 1 and 2).Good separations ofpeaks were obtained except forthe pairs of hydroxyproline, aspartic acid andtyrosine, phenylalanine. The optical densities ofthetest solutions, measured against water, are shownplotted against eluent volume to illustrate theextent of variations in the ninhydrin blank duringthe course of a chromatogram.With most samples, a trace of ninhydrin-positive
material emerged from the 100 cm. column at apoint corresponding to zero retention volume(22 ml.); this may have been cysteic acid (Moore &Stein, 1951) but it was not definitely identified. Theninhydrin blank in the region between alanine andvaline was sometimes slightly erratic, but in noinstance was a clear indication of a much-flattenedcystine peak obtained. These results suggest thatthe cystine content of all samples examined wasless than 0-05 %.The position of emergence of hydroxylysine was
checked, using a synthetic mixture of the diastereo-isomers kindly supplied by Dr J. R. Weisiger of theRockefeller Institute for Medical Research, NewYork. Asymmetry of the leading edge of the lysinepeak with hydrolysates of lime-processed gelatin,but not with acid-processed gelatin or collagen, wasattributed to the presence of ornithine, whichemerges at this point (Moore & Stein, 1951). Noseparate identification of ornithine was made. Nopeak occurred in the position reported for trypto-phan, although this cannot be regarded as evidencefor the complete absence of this amino acid owingto the extensive destruction of tryptophan that isfrequently reported as occurring during acidhydrolysis. A large body of evidence suggests,however, that tryptophan is absent from gelatin(Bowes & Kenten, 1947), and no attempt was madeto determine it separately.
DISCUSSION
i Values for duplicate determinations on 24 hr.hydrolysates of ox-hide gelatin Ib (Table 2) showa reproducibility of better than + 4% for all aminoacids other than tyrosine and better than + 2% foramino acids present in amounts greater than 3 % ofthe gelatin.
Values for amino acids present in amounts lessthan 5 % have been given to two decimal places inTable 2. This is intended to imply not that thesecond place is accurately known but that it may betaken as an indication for comparative purposes. Ithas also been used in computing totals.
-- -- --- -a---
lMAMMALIAN COLLAGEN AND GELATIN
Hydroly8siRecently 6N [20 % (w/w)] hydrochloric acid has
been used by many workers at temperatures of 1050(Smith & Stockell, 1954), 108-5' (Macpherson, 1946)and 1100 (Harfenist, 1953; Hirs, Stein & Moore,1954) for timnes varying from 20 to 140 hr. for
hydrolysis of proteins before amino acid analysis.Close control of temperature appeared importantfor obtaining reproducible hydrolysis, a point thathas been recently emphasized by Hirs et al. (1954).To avoid the inaccuracies of control and the longtime required for the sample to reach the operatingtemperature of an electric oven, a temperature of
Fig. 1. Elution curve of the hydrolysate from 5B56mg. of untreated ox-hide gelatin II, chromatographed on a100 cm. x 0-9 cm. column of Dowex 50. - , Optical density/cm. optical path at 570 mp.; ...... opticaldensity/cm. for imino acids at 440 mu.; ----, the tops of the aspartic acid, glycine and alanine peaks at one-quarter vertical scale. The buffer was changed from pH 3-42 to 4*25 and the temperature from 37.50 to 60° at thepoint marked X.
Fig. 2. Elution curve of the hydrolysate from 11-13 mg. of untreated ox-hide gelatin II, chromatographed on a15 cm. x 0-9 cm. column of Dowex 50 to separate the basic amino acids. The buffer was changed from pH 6-75,0*1M sodium phosphate, to pH 6-5, 0-2M sodium citrate, at the point marked Y.
38-2
Vol. 6I 595
J. E. EASTOE100°, maintained by immersion in boiling water, wasemployed. Superheating and exposure of thehydrolysate to atmospheric oxidation, such asmay occur when the hydrolysate is boiled directly(Macpherson, 1946), were also avoided by thisprocedure.The decreased hydrolysis rate at this lower
temperature did not result in the need for an in-conveniently long period for the complete hydro-lysis of collagen and gelatin. Synge (1953) con-sidered that gelatin does not contain significantproportions of those amino acid sequences whichare extremely stable to acid hydrolysis. The releaseof valine and isoleucine by hydrolysis with 6Nhydrochloric acid has been shown to be incompleteafter 20-24 hr. for certain other proteins such asinsulin (Harfenist, 1953), carboxypeptidase (Smith& Stockell, 1954), papain (Smith et al. 1954) andribonuclease (Hirs et al. 1954) at temperatures of105-110°. However, despite the decreased hydro-lysis rate at 1000, release of amino acids from ox-hide gelatin was found to be substantially completeafter 24 hr. (Tables 2 and 3), the increase in valineand the leucines on extending the heating period to48 hr. being of the same order as the experimentalerror. Collagen, like gelatin, may be supposed to becompletely hydrolysed after 24 hr. at 100°, since thepeptide bonds broken during the conversion ofcollagen into gelatin are labile under very mildconditions. Protein impurities in collagen maycontain sequences of amino acids which requiremore prolonged hydrolysis, but the proportion leftin peptide combination after 24 hr. will be verysmall compared with those already released fromboth collagen and impurity.The mild hydrolytic conditions employed mini-
mized the breakdown of amino acids. Decomposi-tion of serine and threonine with formation ofammonia was observed, but there was no indica-tion of decomposition of aspartic and glutamicacids in periods up to 48 hr., as was observed bySmith et al. (1954) and Hirs et al. (1954) at highertemperatures.The asymmetry of the hydroxylysine peak
(Fig. 2) has been shown by Piez (1954) to be due tothe presence of diastereoisomers, which he separ-ated on a 30 cm. Dowex 50 column. Recently,Hamilton & Anderson (1955) have obtainedevidence that only hydroxy-L-lysine residues arepresent in collagen, allohydroxy-D-lysine beingformed during hydrolysis by racemization at thea-carbon atom.
Purity of coUagenIt is difficult to prepare collagen free from im-
purities (Bowes & Kenten, 1948; Kendrew, 1954).No methods are available for the removal ofelastin and other less well-defined proteins (Eastoe
& Eastoe, 1954) which resist solution in hot water,without risk of simultaneous modification ofcollagen. Further, it is not easy to find validcriteria ofpurity for collagen, nor to define whethera given minor component is strictly part of thecollagen structure or adventitious. Chemical com-position appears, at present, to be the most sensi-tive tool for examining the purity of collagensystems.Two approaches have been used for isolating
samples of collagen. The first, reviewed by Baker,Lampitt & Brown (1954), was aimed at dissolvingimpurities with suitable reagents, leaving intactcollagen fibres. Treatments include the use of:10% sodium chloride to dissolve albumins andglobulins; disodium phosphate, which removesother connective tissue proteins, possibly collagenprecursors (see Harkness, Marko, Muir & Neuberger,1954); alkaline trypsin, which breaks down muscle;lime water, to dissolve mucopolysaccharides andmucoproteins; and fat solvents. The chief difficulty,already mentioned, is the presence of proteins moreresistant than collagen itself.The second type of method involves dissolving
the collagen in a slightly acid solution, possiblywith intermediate purification (Salo, 1950), beforeprecipitating the collagen, in a form that still showsstructure in the electron microscope (Schmitt, Hall& Jakus, 1942), by dialysis or addition of sodiumchloride. The main objection to this method is thatonly some 10% of collagen (Nageotte & Guyon,1934) dissolves, leaving a residue of the bulk of thecollagen, whose chemical composition is close tothat of the soluble collagen (Bowes, Elliott & Moss,1953) but which clearly differs from it in solubility.It appears that fractionation of a heterogeneouscollagen population occurs, the basis of which isnot understood. Orekhovitch (1952) has supposedacid-soluble collagen, termed 'procollagen', to bethe metabolic precursor of relatively insolublecollagen, but more recent work casts doubt on this(Harkness et al. 1954).
Commercially, collagen can be converted intogelatin in 80-90% yield, so that gelatin, unlikeacid-soluble collagen, is derived from the bulk ofthe collagen.In the present study, native collagenous tissue
has been analysed after comparatively mild treat-ment with salt solutions. Use of enzymes has beenavoided, owing to risk of alteration of the collagen(Bowes & Kenten, 1948).
Differences in composition betweencollagen and gelatin
The amino acid data for collagen and gelatin inthe recent literature (Tristram, 1953) show closesimilarities (Table 5). This has been confirmed in thepresent study for collagens and gelatins from a
596 I955
MAMMALIAN COLLAGEN AND GELATIN
number of mammalian sources (Tables 2 and3). Certain differences have, however, beennoted:
(1) The amide-nitrogen content of gelatin pre-pared from limed collagen is less than that of theoriginal collagen and falls to a low value afterprolonged liming (Table 2). There is a resulting fallin the isoionic point of the gelatin to a limitingvalue of approximately pH 4-8 for completeremoval of amide groups (Ames, 1944). Gelatinextracted after short acid pretreatment, however,may contain substantially all the amide nitrogenof the original collagen (Table 2) and have anisoionic point in the region pH 8-5-9-5, whichpresumably approximates to that of collagen,although the isoelectric point of acid-solublecollagen measured in the presence of sodiumacetate appears to be in the region of pH 5-8(Brown & Kelly, 1953). A more prolonged acidtreatment of the precursor results in the removalof a proportion of the amide groups, indicated bya fall in the isoionic point to pH 6-8 (Rousselot,1944).
(2) Treatment of collagen with lime results in aslow conversion of arginine residues into ornithine.In the present study an accurate determnination ofornithine was not possible, since the ornithine peakwas superimposed on the leading edge of the lysinepeak. The presence of a small quantity of ornithinecould be detected, however, by the shape of thecurve in this region, positive indications beingobtained only for the two ox-hide gelatins, theprecursors of which had been limed for 3-4 months.The upper limit for the ornithine content of thesegelatins was estimated to be about 0-2% of aminoacid/100 g. of gelatin, corresponding to about 3%conversion of arginine.
(3) The two differences set out above are welldefined; in addition, there is a less definite tendencyfor gelatin to contain very slightly more of thoseamino acids, e.g. glycine and proline, that arepresent in large amounts in both gelatin andcollagen, and the collagen samples tend to haveslightly larger amounts of amino acids such astyrosine, leucine and phenylalanine, which arepresent in both proteins in small proportion only.These differences could be explained by assumingthe presence, in the collagen preparations, of smallproportions of protein impurities that are notpresent, or present only in smaller proportion, ingelatin, with amino acid compositions markedlydifferent from collagen and gelatin. Bowes et al.(1953) considered that conversion of citrate-solublecollagen into insoluble collagen fibres in vivo mustinvolve the addition of a protein fraction rich intyrosine, histidine, lysine, proline and amidenitrogen and low in hydroxyproline, alanine andserme.
Relationship of gelatin and collagenThe possibility that gelatin is relatively free from
protein-degradation products, derived from con-taminants of native collagen and not otherwiseeasily removed, suggests that gelatin may representhighly purified collagen in amino acid composition.This applies especially to lime-processed gelatin,since removal of mucopolysaccharide and muco-protein constituents (Eastoe & Eastoe, 1954) wouldbe expected to take place during the pretreatmentof raw material and subsequent washing.The extraction itself results in the conversion of
insoluble precursor into soluble gelatin, by aprocess whose mechanism is not understood, butwhich probably involves, among other factors,thermal shrinkage of the precursor. Much of theoriginal inter- and possibly intra-chain structure,dependent on hydrogen bonding, is lost or altered.In addition, occasional peptide bonds are probablybroken, since the molecular weight of collagen, asdetermined by end-group methods, exceeds 1-5 x 106(Bowes & Moss, 1953), and gelatins have a number-average molecular weight of 5-7 x 104 (Pouradier &Venet, 1950; Courts, 1954). After extraction ofgelatin an insoluble protein residue, which is clearlya contaminant of the original collagen, remains.
This theoretical picture of the collagen-gelatinconversion does not indicate any changes in aminoacid composition except those resulting fromremoval of impurities. In assuming the amino acidcomposition of a gelatin as representative of apurified collagen, amide-nitrogen content shouldclearly be determined on native collagen, andornithine present in gelatin which has been sub-jected to alkaline pretreatment should, be inter-preted as an equivalent proportion of arginine.
(ompoaition of 8erilly extracted gelatinDifferences in amino acid composition of gelatin
extracted in the first and third liquors from limedox hide did not greatly exceed the experimentalerror of the method (Table 2). The average mole-cular weights of these were identical as judgedfrom the viscosity data (Table 1), although themolecular-weight distributions could be expectedto be different (Stainsby, private communication);but the mechanical rigidity of gels of the sameconcentration, matured under identical conditions,were markedly different, as indicated by the Bloomjelly strength value, which is an arbitrary measureof gel rigidity. Unexplained anomalies of this typeoften occur in successively extracted gelatinsamples. It is possible that, though there are nolarge differences in amino acid composition, theremay be an overall difference in sequence withrespect to the ends of the polypeptide chainsresulting from a shift in the site of bond-breaking.
Vol. 6I 597
J. E. EASTOE
Composition of a purified gelatinOx-hide gelatins I and II, prepared from
different batches of hide, do not differ markedly incomposition. Gelatin II was purified by a processdesigned to remove (1) high- and low-viscosityfractions known to have abnormally low nitrogencontents and slight brown colours, (2) small ionsand certain neutral molecules, e.g. glucose (Eastoe,unpublished work), absorbed on a mixture ofstrongly acid and strongly basic resins. Afterpurification, only a slight change in the amino acidcomposition, of the same order as the experimentalerror, was observed, indicating a further slighttrend, in the direction already described, on passingfrom collagen to gelatin.The mean value for five to six determinations of
the amino acid composition of ox-hide gelatin hasbeen calculated (Tables 2 and 3) by ignoringdifferences in raw material, extraction, time ofhydrolysis and purification treatment, all of whichwere small.
Comparison with previous results
The mean values, obtained in the present study,for the amino acid composition of ox-hide gelatinare compared in Table 5 with those published
values which have appeared more recently than theearlier summary of Bowes & Kenten (1947). Thevalues for ox-hide collagen compiled by Tristram(1953) are also included.Agreement between the various methods is good
both in the amino acid composition as a whole andin many of the individual values. The most markeddiscrepancies between the present values andprevious ones occur for alanine and serine. Thepresent value for alanine exceeds that of Tristram(1946) by 18% and those of Neuman (1949) by24%. It is unlikely that all the present series ofdeterminations are in error by such a large amountunless an unknown ninhydrin-positive substancecoincides almost exactly with the alanine peak,which shows no abnormality in shape. A similarcomment holds for serine, where the present valueexceeds by some 32% that ofRees (1946), who usedperiodate oxidation. The present values are in goodagreement with those of Bowes (private communi-cation) for ox-hide collagen, obtained by usingion-exchange chromatography.The tyrosine content of different gelatins varies
a great deal. The present results, which indicate a
distinctlylower tyrosine content for gelatin than forcollagen preparations, are lower than most of theliterature values, but are in good agreement with
Table 5. Comparion of the mean of present values for the amino acid composition of ox-hide gelatinwith recent literature values
Values are expressed as g. of amino acid/100 g. of protein.
0-92* Based mainly on Macpherson (1946) for basic amino acids, Rees (1946) for serine and threonine, Tristram (1946) for
alanine, valine, leucines, proline, phenylalanine and tyrosine, and Bowes & Kenten (1948) for methionine and acidic aminoacids, with later revisions.
t Neuman & Logan (1950).
I955598
V6MAMMALIAN COLLAGEN AND GELATINdeterminations on the same samples made in thisLaboratory (Kenchington, private communication)by the colorimetric method of Udenfriend &Cooper (1952). Harkness et al. (1954) found thatthe tyrosine contents of the various collagenouscomponents of skin differ appreciably from oneanother; gelatin prepared from insoluble collagencontained more tyrosine than either alkali- or acid-soluble collagen. Fractionation of this gelatin byprecipitation with trichloroacetic acid, however,reduced the tyrosine content from 1 to 0-5 %. Acorresponding reduction from 0-29 to 0-18% forox-hide gelatin, after fractionation with ethanol,was found in the present study. This last value fortyrosine is close to that of 0.14% recorded byNeuman (1949) for a carefully purified calf-skingelatin made by Eastman Kodak Ltd. It is possiblethat the main protein constituent of gelatin andcollagen contains no tryosine, but the questioncannot yet be regarded as settled.The present values for valine are in agreement
with those obtained by microbiological methods(Table 5), but are significantly lower than thoseobtained by Tristram (1946). The proline valuesobtained in this study are somewhat higher thanthose of Tristram (1946). The sample of prolineused for the present calibration gave a 4% higherintensity of colour than that reported by Moore &Stein (1951) and was therefore considered to besubstantially pure. The values for lysine in gelatinreported by Tristram (1953) may possibly beaugmented by the presence of a small amount ofornithine, as is the present value. The high valuesreported by Neuman (1949) may result from theinability of the microbiological method to distin-guish hydroxylysine from lysine. The present valuesfor hydroxylysine in gelatin are slightly lower thanmost ofthose in the literature, obtained by chemicalmethods (Ramachandran, 1953).
ReticulinHuman renal reticulin, which contained myristic
acid, carbohydrates and inorganic matter (Windrumet al. 1955), was hydrolysed without preliminaryseparation ofthe protein. The general picture oftheamino acid composition of reticulin is quite similarto that of collagen; there are, however, somenotable differences. The low values for alanine,glycine and proline, together with the high valuesfor the leucines, phenylalanine and methionine,suggest that reticulin contains collagen accom-panied by other proteins in smaller amounts. Thisis contradicted, however, by high values forhydroxyproline, hydroxylysine, serine and threo-nine. Examination of reticulin prepared fromother sites in the body appears to be necessary fora full appraisal of its range of chemical com-position.
Variation of collagen compo8itionin mammalt8
The mammalian gelatins and collagens form acompact group as regards their amino acid com-position. The placental land mammals man, oxand pig show few differences in composition,except for the low value for isoleucine in pig-skingelatin.The whale, on the other hand, shows striking
increases in the hydroxyamino acids serine andthreonine, compared with the land mammals.Even higher levels of these two amino acids arefound in fish collagens (Neuman, 1949; Gustavson,1955), combined with decreased levels of hydroxy-proline and proline. Whales, which are descendantsof land mammals, have probably lived in the seasince Eocene times or earlier (Young, 1950). Thehigh content of hydroxyamino acids in whalecollagen may perhaps therefore be attributed to thelong-term effect of a marine diet.The wallaby also differs from the other land
mammals in having increased levels of serine andthreonine. The effect is only slight compared withthe whale, but may be of evolutionary significance,since the wallaby belongs to the marsupials, agroup which has been separated from the placentalsfor a long period.
Achilles tendon is not a satisfactory material forthe preparation of human collagen that is suffi-ciently pure for determinations of amino acidcomposition. It appears to contain substanceswhich render it difficult to convert into gelatin. Thestrong acid extraction finally employed evidentlydissolved considerable quantities of non-nitro-genous constituents, shown by the low recovery ofmaterial by weight, despite the complete recoveryof nitrogen (Table 2). The increased hydroxylysinecontent and correspondingly diminished lysinevalue of this material is difficult to explain. Thesame effect is shown, but to a smaller degree, inwallaby tendon.The specialized variations in composition of
collagen from widely different mammalian species,superimposed on an otherwise fixed composition,suggests that, although collagen is built up accord-ing to a definite pattern, its structure is notabsolutely fixed and amino acids in certain posi-tions may be replaceable within limits byothers.Examination of the amino acid composition of
collagen from a much wider range of anils mayindicate that a proportion of the residues form aninvariable framework throughout the series. Suchinformation might be an important guide to theinterpretation of X-ray-diffraction data and thusassist in resolving the complete picture of collagenstructure.
Vol. 6I 599
600 J. E. EASTOE I955SUMMARY
1. The amino acid composition of collagen orgelatin from five species of mammals (man, ox, pig,whale and wallaby) has been determined by ion-exchange chromatography.
2. The results obtained for ox-hide gelatin are ingood agreement with previously published values,except for alanine and serine.
3. The composition of gelatin is closely similarto that of the collagen preparations, suggestingthat gelatin is representative of the main proteinconstituent of collagenous tissues in amino acidcomposition.
4. The collagen of the land placentals studied didnot show any marked variation between species.The content of serine and threonine in whalegelatin exceeded that in land mammals, but wasnot so high as that from other classes of marinevertebrates. Wallaby collagen also contained veryslightly higher amounts of serine and threoninethan the other land mammals.
5. There were no significant differences in aminoacid composition between two gelatins, seriallyextracted from limed ox hide. These gelatinsdiffered in rigidity properties, although viscositymeasurements indicated that they had the sameaverage molecular weight.
6. Gelatin contained less tyrosine than thecollagenous precursors. Fractionation of gelatinwith ethanol to remove impurities led to a furtherreduction in tyrosine content. It has not yet beenproved that tyrosine is a true constituent of gelatin.
The author is very grateful to Dr G. M. Windrum of theDepartment of Pathology, Radcliffe Infirmary, Oxford, forproviding the samples of human material; to Professor S.Kikuchi of Tokio, Japan, for a sample of whale-skingelatin, to Dr S. M. Partridge of the Low TemperatureResearch Station, Cambridge, for a sample of chromato-graphically purified hydroxyproline, and to the Superin-tendent of the Zoological Society of London's Park,Whipsnade, Bedfordshire, for obtaining the wallabymaterial used in this study. This paper is published bypermission of the Director and Council of The BritishGelatine and Glue Research Association.
REFERENCES
Ames, W. M. (1944). J. Soc. chem. Ind., Lond., 63, 200, 234,277, 303.
Baker, L. C., Lampitt, L. H. & Brown, K. P. (1954).J. Sci. Fd Agric. 5, 226.
Bear, R. S. (1952). Advanc. Protein Chem. 7, 69.Bowes, J. H., Elliott, R. G. & Moss, J. A. (1953). In
Nature and Structure of Collagen, p. 199. London:Butterworth.
Bowes, J. H. & Kenten, R. H. (1947). Summary ofPublished Results (1900-46). On Amino Acid Composi-tion of Gelatin and Collagen. London: British LeatherManufacturers' Research Association.
Bowes, J. H. & Kenten, R. H. (1948). Biochem. J. 43,358.
Bowes, J. H. & Moss, J. A. (1953). Biochem. J. 55,735.
Brown, G. L. & Kelly, F. C. (1953). In Nature and Structureof Collagen, p. 169. London: Butterworth.
Chibnall, A. C. (1942). Proc. Roy. Soc. B, 131, 136.Chibnall, A. C. (1946). J. int. Soc. Leath. Chem. 30, 1.Chibnall, A. C., Rees, M. W. & Williams, E. F. (1943).
Biochem. J. 37, 354.Courts, A. (1954). Biochem. J. 58, 70.Eastoe, J. E. & Eastoe, B. (1954). Biochem. J. 57, 453.Graham, C. E., Waitkoff, H. K. & Hiers, S. W. (1949).
J. biol. Chem. 177, 529.Gustavson, K. H. (1955). Nature, Lond., 175, 70.Hamilton, P. B. & Anderson, R. A. (1955). J. biol. Chem.
213, 249.Harfenist, E. J. (1953). J. Amer. chem. Soc. 75, 5528.Harkness, R. D., Marko, A. M., Muir, H. M. & Neuberger,
A. (1954). Biochem. J. 56, 558.Hirs, C. H. W., Stein, W. H. & Moore, S. (1954). J. biol.
Chem. 211, 941.Janus, J. W., Kenchington, A. W. & Ward, A. G. (1951).
Research, Lond., 4, 247.Kendrew, J. C. (1954). The Proteins, vol. 2B, p. 845.New York: Academic Press.
Macpherson, H. T. (1946). Biochem. J. 40, 470.Martin, C. J. & Axelrod, A. E. (1953). Proc. Soc. exp. Biol.,
N.Y., 83, 461.Moore, S. & Stein, W. H. (1948). J. biol. Chem. 176, 367.Moore, S. & Stein, W. H. (1951). J. biol. Chem. 192, 663.Nageotte, J. & Guyon, L. (1934). C.R. Ass. Anat. 29, 408.Neuman, R. E. (1949). Arch. Biochem. 24, 289.Neuman, R. E. & Logan, M. A. (1950). J. biol. Chem. 184,
299.Orekhovitch, V. N. (1952). 2nd Int. Congr. Biochem.,
Paris. Communications, p. 106.Piez, K. A. (1954). J. biol. Chem. 207, 77.Pouradier, J. & Venet, A. M. (1950). J. Chim. phys. 47, 11.Ramachandran, L. K. (1953). J. sci. industr. Res. 12B, 9.Rees, M. W. (1946). Biochem. J. 40, 632.Robson, W. & Selim, A. S. M. (1953). Biochem. J. 53,431.Rousselot, A. (1944). C.R. Acad. Sci., Paris, 219, 62.Salo, T. P. (1950). Arch. Biochem. 28, 68.Schmitt, F. O., Hall, C. E. & Jakus, M. A. (1942). J. cell.
comp. Physiol. 20, 11.Schram, E., Dustin, J. P., Moore, S. & Bigwood, E. J.
(1953). Analyt. chim. acta, 9, 149.Smith, E. L. & Stockell, A. (1954). J. biol. Chem. 207, 501.Smith, E. L., Stockell, A. & Kimmel, J. R. (1954). J. biol.
Chem. 207, 551.Stainsby, G. (1955). Disc. Faraday Soc. no. 11, 288.Synge, R. L. M. (1953). Ciba Foundation Symposium. The
Chemical Structure of Proteins, p. 193. London: Churchill.Tristram, G. R. (1946). Biochem. J. 40, 721.Tristram, G. R. (1953). The Proteins, vol. lA, p. 181.New York: Academic Press.
Udenfriend, S. & Cooper, J. R. (1952). J. biol. Chem. 196,227.
Ward, A. G. (1954). Chem. & Ind. p. 502.Windrum, G. M., Kent, P. W. & Eastoe, J. E. (1955).
Brit. J. exp. Path. 36, 49.Young, J. Z. (1950). The Life of Vertebrates. London:
