THE JOIJRNAL OF BIOLOGICAL CHEMX~TRY

Vol. 242, No. 23, Issue of December 10, pp. 54816439, 1967
Printed in U.S.A.

Characterization of Chick Bone Collagen and Compositional

Changes Associated with Maturation
(Received for publication, June 21, 1967)

E. J. MILLER, G. R. MARTIN, K. A. PIEZ, AND M. J. POWERS

From the Laboratory of Biochemistry, National Institute of Dental Research, National Institutes of Health,
Bethesda, Maryland 20014

SUMMARY bonesindicated that during development there was a progres-

Collagen extracted from the diaphyseal region of the tibiae sive decrease in hydroxylysine content, accompanied by an
equivalent increase in lysine content, with no significant
of normal and lathyritic chicks with 0.5 M acetic acid and 5 M

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

changein the content of any other amino acid. This observa-
guanidine hydrochloride at neutral pH has been studied with
tion was verified by amino acid analyses of the total bone
regard to chain organization and amino acid composition.
Decalcification by dialysis against large quantities of 0.5 M collagen of normal chicks of various ages. The hydroxy-
lysine content decreased by nearly one-half, from 11 to 6
acetic acid and subsequent exposure of the decalcified bones
to fresh 0.5 M acetic acid followed by 5 M guanidine resulted residues per 1000, between the 1st and the 49th day after
hatching. This result suggests that the hydroxylation of
in the solubilization of 85% of the total collagen in the bones
bone collagen lysyl residues is dependent in part upon factors
from lathyritic animals. Collagen in the bones of control
animals was almost entirely resistant to extraction with 0.5 extrinsic to the collagen molecule.
M acetic acid, and lessthan 20% of the total collagen could be
extracted in 5 M guanidine. These observations suggest a
high degree of intermolecular cross-linking. Chromatog-
raphy of the collagen extracted by both solvents from the
bones of lathyritic animals on columns of carboxymethyl X-ray diffraction studies (I) indicate that the collagen mole-
cellulose indicated that the extracts were composedalmost cule has a coiled coil, triple chain structure. Further information
entirely of single- and double-chain components. Carboxy- about the structure of various vertebrate collagens has been
methyl cellulose chromatography of the collagen extracted derived by examining the components produced when soluble
with 5 M guanidine from the bones of normal chicks revealed collagens are denatured. The components obtained from salt-
that only 30% of the dissolved collagen (5 % of the total extracted collagen are predominantly of molecular weight of
collagen) could be recovered asthese components. Chroma- about 95,000, and are termed (Y components or (Y chains (2-5).
tography on an agarosemolecular sieve column showed that In most collagens so far examined the amino acid composition of
the remainder could be accounted for in part by high molec- one of the three (Y chains (the (~2 chain) differs significantly from
ular weight aggregates and in part by low molecular weight, that of the other two (the crl chains), making possible the sepa-
degraded material. These results suggestthat bone collagen ration of the two types by chromatography on carboxymethyl
fibrils are stabilized through the formation of intra- and cellulose (4-6). All three (Y chains of codfish skin collagen,
intermolecular cross-links and that lathyrogens exert their however, differ in amino acid composition and can be separated
effects by inhibiting cross-link formation, as previously by CM-cellulose chromatography (7, 8). Other components of
shown for soft tissue collagens. The extractability of normal higher molecular weight than the (Y chains have been found in
bone collagen in guanidine appearsto result from degradation significant amounts in solutions of denatured, acid-extracted
in the presence of the denaturing agent to yield a highly collagens. Covalently cross-linked components containing two
heterogeneous mixture. a: chains are called ,8 components (2-4), and those with three (Y
Carboxymethyl cellulose chromatography of acid-extracted chains have been designated y components (9-11). These high
collagen from the bones of lathyritic animals indicated that molecular weight components have been shown to arise by intra-
the bone collagen molecule contains two chains (al) that are and intermolecular cross-linking of (Y chains (4, 12-14).
identical or similar and one chain ((~2) that is unlike the other The presence of cross-linked components of soft tissue collagens
two. Their compositions are analogous to those of other coincides with a decrease in the ease of extraction of collagen
vertebrate collagens. from the tissue. Collagen that is not highly cross-linked or
Amino acid analyses of the chromatographically separated collagen with largely intramolecular cross-links can be extracted
cychains in successiveextracts of collagen from the lathyritic from soft tissues with solvents such as cold salt (NaCI) solutions

5481
 Chick Bone Collagen Vol. 242, No. 23

at neutral pH and dilute acetic acid (4), while varying amounts bones from both control and lathyritic animals was complete by
of more highly cross-linked collagen can be removed by extrac- the criterion described above when EDTA was the decalcifying
tion with denaturing solutions, such as 5 M guanidine hydrochlo- agent.
ride (6,15). In the latter solvent, the structure of the native col- During the extensive dialysis required for decalcification, the
lagen molecule is destroyed (16). Depending on the species, the solution in the tubing containing bones from lathyritic animals
tissue, and the age of the animal, a considerable amount of colla- became somewhat turbid and viscous, while the solution in con-
gen remains insoluble even after extraction with 5 M guanidine- tact with the bones from control animals remained clear with
HCl.’ little apparent increase in viscosity. The solution remaining in
Since bone collagen is largely insoluble in the solvents that are the dialysis bags after decalcification by dialysis against 0.5 M
used to extract native collagen, it has been more difficult to acetic acid will subsequently be referred to as the first acid ex-
characterize. In this paper, we report our studies on the char- tract. The residues remaining after decalcification were ex-
acterization of collagen extracted with 0.5 M acetic acid and 5 M tracted for 1 week with constant stirring in 200 ml of fresh 0.5 M
guanidine-HCl from the tibiae of chicks fed &aminopropionitrile acetic acid. The collagen solubilized during this extraction pro-
fumarate, a lathyrogen that inhibits cross-linking of collagen cedure is termed the second acid extract. An additional extrac-
and renders it more readily extractable (12, 13, 17). In addition, tion period of 1 week yielded only a small amount of additional
we have studied the molecular organization of collagen extracted collagen.
with 5 M guanidine-HCI from the tibiae of normal chicks. In Following the third acid extraction the bone residues were
several important respects our results differ from those of extracted for 1 week with constant stirring in 200 ml of 5 M
Glimcher and Katz (18)) Glimcher, Katz, and Travis (19)) guanidine-HCl (clarified by filtration through activated charcoal)
Glimcher et al. (20), and Francois and Glimcher (21), who have at pH 7.0. Extraction of the residues for additional weekly

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

also studied chick bone collagen. periods resulted in the solubilization of only small amounts of
collagen. The residues remaining after guanidine-HCl extrac-
EXPERIMENTAL PROCEDURE tion were thoroughly washed with distilled water to eliminate all
of the guanidine, and were dried by lyophilization.
Source of Collagen-White Leghorn chicks, obtained from a
All extracts were clarified by centrifugation at 100,000 x g
local hatchery at the age of 1 day, were maintained on a commer-
for 2 hours in a Spinco model L ultracentrifuge. Collagen con-
cial ration for various intervals as required by the experimental
tained in the 0.5 M acetic acid extracts was precipitated by
procedure. Lathyrism was induced by feeding the chicks a
dialysis against large volumes of 0.02 M disodium phosphate.
commercial diet to which was added 1 g of @aminopropionitrile
The precipitate was collected by centrifugation at 40,000 x g
fumarate per kg from the time of arrival until 3 weeks of age.
for 30 min, redissolved in 0.5 M acetic acid, and precipitated again
At this time, the control and lathyritic animals (120 in each
by the addition of sufficient sodium chloride to give a concentra-
group) were killed by decapitation, and the tibiae were immedi-
tion of 5%. The precipitate was collected by centrifugation,
ately removed and rinsed several times in cold (5”) 0.16 M NaCl
redissolved in 0.5 M acetic acid, dialyzed against a large volume
and then in water. After the epiphyseal ends had been carefully
of the same solvent, and lyophilized. No attempt was made to
removed in order to preclude contamination with cartilaginous
further purify the collagen extracted by 5 M guanidine-HCl,
areas, the bones were cleaned of periosteum and endosteal tissue.
since amino acid analyses of this material indicated that collagen
The remaining tissue, consisting of long thin cylinders of dia-
comprised approximately 980/, of the total protein (Table III,
physeal bone, was freeze-dried and weighed. The intact bone
below). Therefore, an aliquot of the guanidine-HCl extracts
cylinders were used for all subsequent procedures.
was dialyzed exhaustively against water, lyophilized, and weighed
Calculation of Bone Collagen and Mineral Content-In each
to determine the approximate collagen concentration. The
experiment 50 dry bone cylinders from each group were de-
remainder was dialyzed directly against a large volume of 0.06
calcified by constant stirring in 500 ml of cold 0.35 M EDTA, as
I’/2 sodium acetate buffer, pH 4.8, in preparation for CM-
the potassium salt, at pH 8.1. Decalcification, as judged by
cellulose chromatography.
the absence of an ash residue when the bones were heated at 600”
CM-cellulose Chromatography-A weighed sample of acid-
for 24 hours in platinum crucibles, was complete within 4 days
extracted collagen (15 to 25 mg) was suspended in 20 ml of start-
provided the decalcifying solution was changed daily. Following
ing buffer, 0.06 P/2 sodium acetate, pH 4.8, and stirred for a few
decalcification, the organic residues were washed repeatedly with
hours at 5” until the collagen was completely dissolved. The
cold distilled water to remove EDTA, lyophilized, and weighed.
solutions were warmed to 45” for 30 min and applied by means
The proportion of collagen and mineral in the bone samples was
of a peristaltic pump (LKB ReCyChrom) to a column, 15 x 75
then determined from the loss of weight during decalcification.
mm, of CM-cellulose (Whatman, microgranular CM-32, capac-
EDTA decalcification was used only for the preparation of total
ity, 1.0 meq per g) that had previously been equilibrated with
bone collagen. All extracted collagens were prepared following
starting buffer. The column temperature was maintained at
acetic acid decalcification, described below.
42”. Fractionation of the material applied to the column was
Preparation of Collagen-All procedures were performed at 5”.
achieved essentially as described by Piez, Eigner, and Lewis (4).
Calcified bone cylinders from each group were placed in cellulose
A linear gradient was obtained by means of a two-chamber con-
dialysis tubing containing 200 ml of 0.5 M acetic acid and dialyzed
stant level device containing 200 ml of starting buffer in the first
extensively against 6-liter volumes of 0.5 M acetic acid which
chamber and 200 ml of limit buffer (0.06 P/2 sodium acetate con-
were renewed twice daily. After 3 weeks, decalcification of the
taining 0.1 M NaCl, pH 4.8) in the second chamber. The column
1 The abbreviation used is: guanidine-HCl, guanidine hydro- was operated at a flow rate of 150 ml per hour. Chromatography
chloride. of the collagen contained in the 5 M guanidine-HCl extracts was
Issue of December 10, 1967 Miller, Martin, Piez, and Powers 5483

performed in the samemanner, except that the sampleswere TABLE I

clarified by filtration through medium porosity sintered glass Analysis of diaphyseal region of control and lathyritic chick tibiae
discsafter being warmed in the starting buffer. for mineral and organic content
The column effluent was monitored continuously at 230 rnp Resultsare expressedas the averageof 50 bonesin eachgroup.
by the useof a flow cell (lo-mm light path, 0.3-ml volume) in a
Weight
Beckman DB spectrophotometer,and absorption was recorded BOJJ.2 Mined
content
by a SargentSRL strip chart recorder equippedwith logarithmic CdCiliCd Decalcified
gearsand operatedat a speedof 6 inchesper hour. The column
mg/bonc, dry wl
effluent was collected in lo-ml fractions. Combinedfractions %

correspondingto various regionsof the chromatogramswereused Control. 74.0 17.1 77.0

for amino acid analysesand disc electrophoresisafter dialysis Lathyritic 63.2 14.2 77.6
againstdistilled water and lyophiliaation.
The areaunder the peaksin eachchromatogramwascalculated TABLE II
by planimetry. A conversion factor for determining sample Percentage of total bone collagen extracted by 0.6 M acetic acid and
weightsfrom areasunder the peaks was obtained by chromato- 5 lli guanidine hydrochloride
graphinga well characterizedsampleof rat skin collagenthat was The values given are averages of data obtained in two sepa-
known to be 100% recoverable as LYand p components. The rate experiments.
percentageof total samplecollagen recovered as various com-
ponents after CM-cellulose chromatography could then be de- solvent Control Lsthyritic
termined.

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

Molecular Sieve Chromatography-Samples containing about
100mg of extracted collagenweredissolvedin 10 ml of 1 M CaCIZ
buffered with Tris (0.05 M, pH 7.5) by warming to 40’ for 15 to
30 min and were placed on a column, 2.5 x 110 cm (549 ml),
of Sepharose4B (Pharmacia)at room temperature. The column
wasequilibrated and eluted with the samesolvent at 40 ml per
hour. The effluent wasmonitored asdescribedfor CM-cellulose
chromatography.
Polarimetry-Collagen obtained in the first acid extract of
bones from lathyritic animals was dissolved in either 0.5 M
acetic acid or 0.15 M potassiumacetate, pH 4.8, at a concentra-
tion of 1 mg per ml. Optical rotation of the samplesin both
solventswasmeasuredat 313 rnp in a Rudolph model80 photo-
electric spectropolarimeterwith a I-dm jacketed cell and with a
mercury arc lamp asthe light source. Denaturation curveswere
obtained by a stepwiseincreasein temperature over the range,
22-50”. At eachtemperature, 30min were allowedfor equilibra-
tion. Collagenconcentrations in the solutions used for polar-
imetry were calculated from the nitrogen concentration (micro-
Kjeldahl), assuminga nitrogen content of 18.6% for collagen.
Acrylamide Gel Electrophoresi-s-Disc electrophoresisof the
fractions obtained during CM-cellulose chromatography of
collagenextracted with 0.5 M acetic acid and 5 M guanidine-HCl
wasperformed accordingto the procedureof Nagai, Gross,and
Piez (22).
Amino Acid Analyses-Samples of bone collagen (2 to 5 mg)
werehydrolyzed in sealedtubesunder nitrogen in 3 ml of constant
boiling 6 N HCl for 24 hoursat 108”. Amino acid analysesof the
hydrolysates were performed on an automatic amino acid ana-
lyzer (23) asmodifiedfor acceleratedanalyses(24). In the final
calculations, correctionswere made for the lossof labile amino
acids (threonine, serine, tyrosine, and methionine) and for in-
complete releaseof valine (25).
RESULTS
In Table I are listed data on the relative amountsof mineral
and organic phasespresent in the bonesused in these studies.
It is seenthat lathyrism causedlittle alteration in the relative
proportion of the mineral and organic phasesin the diaphyseal
regionof chick tibiae. Theseresultsare in agreementwith those
% %
0.5 M acetic acid
First extract (decalcification). 0 7
Second extract. . <l 13
5 M guanidine hydrochloride. 17 64
of Glimcher et al. (20), who studied the mineral and organic
phasesof metatarsal and tibia1 bonesin normal and lathyritic
chick embryos. The decreasein total massobservedfor lath-
yritic chick bone is compatible with the 10 to 15% decreasein
total body weight observed in the experimental animals. The
decreasein body weight wasmost probably the result of an in-
sufficient dietary intake due to immobilization during the last
few days of &aminopropionitrile feeding. Although small
amounts (0.5 to 1.0 mg per bone) of protein were removedfrom
the bonesduring decalcification in the EDTA solutions,amino
acid analysisof this material indicated that no collagenhad been
solubilized. Amino acid analysisalso showedthat collagenac-
counted for 98% of the protein of the organic residuesfollowing
decalcification. These results for avian bones are similar to
those obtained for mammalian compact bone by Eastoe and
Eastoe (26), who found that collagenaccountedfor 94% of the
total organic matter of ox femur diaphyses.
The extractability of the bone collagenin the solventsusedis
given in Table II. The resultswere calculatedfrom the amount
of collagenobtained in the individual extracts and are expressed
as a percentageof the total collagencontained in the calcified
starting material, assumingcollagento be 23% of the dry weight
of the calcified bones. During decalcification and extraction
with 0.5 M acetic acid, lessthan 1y0 of the total collagenin bones
from control chicks was solubilized. Even 5 M guanidine-HCI
was relatively ineffective in removing collagenfrom this tissue.
In contrast, appreciablequantities of collagenin the bonesfrom
lathyritic animalswere solubilizedin acetic acid, and guanidine-
HCl dissolvedthe majority of the rest.
The markedly increasedextractability of collagenfrom the 3-
week-old lathyritic chick bone probably resulted from the fact
that this collagenwassynthesizedfor the most part in the pres-
enceof the lathyrogen. The averageweight of the dry decalcified
diaphysealregion of the tibia of the newly hatched chick is 3.8
 5484 Chick Bone Collagen Vol. 242, No. 23

mg. As seen in Table I, this value increases to 14.2 mg at 3 A chromatogram obtained by CM-cellulose chromatography
weeks of age. Moreover, remodeling during growth would lead of the collagen present in the first acetic acid extract of bones from
to the removal of all or most of the normal bone collagen pres- lathyritic animals is shown in Fig. 1. The peaks were identified
ent when feeding of the lathyrogen was begun. as LYor /3 components by their migration during gel electropho-
resis. Amino acid analyses established that the peak labeled
&Z was composed of an equal mixture of (~1 and (~2 chains (Table
III). Protein in fractions taken across the crl peak (Fractions
I to V, Fig. 1) did not vary significantly in amino acid composi-
tion. This chromatogram resembles previously published chro-
matograms of soft tissue collagens, such as rat skin and tail
tendon (4), shark skin (5), and human skin (6) collagens, in that
it contains three prominent components, crl, ru2, and &. The
fin component frequently observed in chromatograms of other
collagens was not present in sufficient amounts to be readily
apparent. However, material with the mobility of a /3 compo-
nent was observed by acrylamide gel electrophoresis in the trail-
ing edge of the (~1 peak when more highly cross-linked bone
collagen was chromatographed (see below). The ratio of the
area of crl to that of the (~2 component was 2 : 1, consistent with
the presence of two crl and one (~2 chains per molecule, as found

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

for other collagens (4). Calculation of the area under the peaks
of the chromatogram showed that within experimental error all
of the collagen in the first acid extract was recovered as rul, &,
0 2-5 5’0 75 100 150 I75 200 2i5
EFFLUENT VOLUME, ml.
and 02 components.
FICJ. 1. Elution pattern of 14 mg of denatured lathyritic chick
A chromatogram of the collagen obtained in the second acid
bone collagen (first acid extract) chromatographed on CM-cellu- extract of bones from lathyritic animals is presented in Fig. 2.
lose with a linear gradient of ionic strength from 0.06 to 0.16 at pH The elution pattern resembles that observed for collagen in the
4.8. Protein from Fractions I to VII was taken for amino acid first acid extract, although the double-chain /3rz component com-
analyses and gel electrophoresis. prises a somewhat greater proportion of the total collagen. Of
TABLE III the collagen in the sample, 100% was accounted for as the ob-
Amino acid composition of b-week-old chick bone collagen
served components.
The values given represent averages of duplicate analyses.
In Fig. 3 is presented the chromatogram of collagen extracted
from the bones of lathyritic animals with 5 M guanidine-HCl.
Lathyritic bonea Control bona The forepeak consisted of a strongly absorbing noncollagenous
Amino acid I I I I I I contaminant of the preparation which accounted for only a small
percentage of the weight of the sample. Hydrolysates of
I I I , I I
this material contained hexosamines and no hydroxyproline.
rcsiducs/l0Wtotalresidues
3-Hydroxyproline . ......
Broadening of the peaks, indicative of heterogeneity, was ap-
0.9 1.0 0.8 0.9 0.8 1.0 1.0
4-Hydroxyproline. ...... 102 102 100 98 101 101 100 parent. As shown below, this resulted from the lengthy exposure
Aspartic acid. .......... 42 46 48 46 42 44 48 to the denaturing solvent as well as from the presence of un-
Threonine. ............. 19 19 19 20 19 19 18 resolved high molecular weight components. A part of the
Serine. ................. 29 29 29 28 28 28 28
Glutamic acid ........... 78 73 67 74 78 73 65
Proline. ................ 118 117 118 114 118 119 120 0.8
Glycine.................3 29 329 329 326 332 331 330 c
Alanine................. 29 114 102 121 128 116 104
Valine. ................. 14 21 28 19 14 20 26
Methionine ............. 8.6 7.6 6.0 7.2 8.6 7.7 5.4
Isoleucine. ............. 6.1 11 18 12 6.3 11 18
Leucine ................. 20 26 32 25 20 26 31
Tyrosine ................ 2.1 2.0 2.0 4.9 2.4 2.4 2.5
Phenylalanine ........... 14 14 14 15 14 14 14
Hydroxylysine.. ........ 5.5 6.8 8.0 5.9 5.5 7.0 8.2
Lysine. ................. 29 27 24 28 29 26 24
Histidine ............... 2.8 5.0 7.1 6.5 2.8 4.8 7.0
Arginine. ............... 51 49 49 49 49 49 50
Amide nitrogen=. ....... (39) (43) (45) (45) (39) (40) (44)

a The collagen was that obtained in the first 0.5 M acetic acid
extraction. 0 25 50 75 100 125 150 175 200 225
b The collagen was that obtained by extraction with 5 M guani- EFFLUENT VOLUME, ml.

dine hydrochloride. pii had the same composition as (~1. FIG. 2. CM-cellulose elution pattern of 16 mg of denatured
c Not included in calculation of total residues. lathyritic chick bone collagen (second acid extract).
Issue of December 10, 1967 Miller, Martin, Piez, and Powers 5485

0.7- linked material, and the soluble high molecular weight compo-
nents that are not eluted from the column under the conditions
0.6-
C7I
used account in part for the low recoveries (see below). Similar
05- high molecular weight aggregates have been studied by Veis and
2 Anesey (14), who used elution at high salt concentrations and
4
wo.4-2 2 buffers containing urea.
&-
5 3* $ These results, expressed as recovery of collagen from normal
P 03- 2 s and lathyritic bones as cr and @ components, are summarized in
a 3 B
9 02-G 8 Table IV. Only 5% of the total collagen in normal bone was
recovered, while about 78% was recovered from lathyritic bone.
0.J ( J If only the single-chain al and (r2 components are considered,
these recoveries become 4% and 59% of total bone collagen,
0 25 50 75 100 125 150 175 200 225 250 respectively.
EFFLUENT VOLUME,ml. A possible factor contributing to the broadened peaks seen in
FIN. 3. CM-cellulose elution pattern of 30 mg of denatured CM-cellulose chromatography of guanidine-extracted collagen
lathyritic chick bone collagen (5 M guanidine-HCl extract). Pro- was thought to be degradation occurring in the denaturing
tein from Fractions I to IV was taken for analysis. The forepeak
is noncollagenous.

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

material chromatographing with /3iZ in Region III failed to enter
the separating gel during the disc electrophoresis, indicating the
presence of material at least as large as the y component. A part
of the material in Region II (Fig. 3) displayed the mobility of a
/3 component, and is listed as pii since this region had the same
amino acid composition as Region I. The /3 components com-
prised a greater proportion of the collagen than in the second
acid extract. Planimetry of the area under the peaks on three
chromatograms of guanidine-extracted collagen from bones of
lathyritic animals indicated that about 90% of the extracted
collagen was recovered under the conditions used.
The collagen in the bones from normal animals that could be
extracted by 0.5 M acetic acid was insufficient to allow purifica-
tion and fractionation by CM-cellulose chromatography. A 0 i5 50 75 100 125 150 175 200 225 250
EFFLUENT VOLUME.ml.
representative chromatogram of the collagen extracted from
control bones by 5 M guanidine-HCI is shown in Fig. 4. The FIN. 4. CM-cellulose elution pattern of 30 mg of denatured
elution pattern showed the expected components but again in- normal chick bone collagen (5 M guanidine-HCl extract). Protein
from Fractions I to IV was taken for analysis. The forepeak is
dicated marked heterogeneity, similar to the results for collagen noncollagenous.
extracted from bones of lathyritic animals by guanidine-HCI
(Fig. 3). The forepeak was found to represent a noncollagenous
TABLE IV
contaminant, similar to that seen in guanidine-HCl extracts of
Percentage of collagen sample recovered after
bones from lathyritic animals, although it was relatively larger,
CM-cellulose chromatography
perhaps because the proportion of collagen solubilized was
The numbers are averages from two or three chromatograms
smaller. Gel electrophoresis of material chromatographing with
except for the second 0.5 M acetic acid extract, which was chro-
pi2 in Region III likewise indicated the presence of unresolved
high molecular weight components, and /3ii was clearly observed
matographed once.
- - -
in Region II (Fig. 4). The amino acid compositions of the total
guanidine-HCI extract and the al, (~2, and /3iZ components of the
Bonesand
extraction solvent
I Component
- - I 1 ‘otal *s per-
centa
total %z
01 a2 BlS collagen
collagen extracted by guanidine-HCl from control bones are -- _-
listed in Table III. The components obtained from guanidine- % % % % %
extracted collagen and acid-extracted lathyritic collagen do not Normal5
differ significantly in composition. The amino acid composition 5~guanidine...... 176 5 8” 30 5
of the total extract from control bone indicates the presence of a Lathyritic
small percentage of contaminating protein. First 0.5 M acetic
Calculations of the area under the peaks observed in chromat- acid. . 62 31 7 100 7
ograms of collagen from control bones showed that only about Second 0.5 M acetic
acid. . . 57 29 14 100 13
30% of the collagen originally extracted by 5 M guanidine-HCl
5~guanidine...... 40b 23 27” 90 58
was actually recovered. In contrast to collagen extracted by - -
guanidine-HCl from lathyritic bones, some of the guanidine- 6 Less than 1% of the collagen in normal bones could be ex-
extracted collagen from control bones would not dissolve on tracted by 0.5 M acetic acid (Table II).
warming in the starting buffer and therefore could not be ap- b Includes j3ii; see text.
plied to the column. This fraction, presumably highly cross- c Includes some y component; see text.
5486 Chick Bone Collagen Vol. 242, No. 23

solvent. To test this hypothesis, acid-extracted lathyritic lyophilization of the effluent fractions. It is clear that the very
collagen was dissolved in 5 M guanidine-HCI, allowed to stand low recoveries after CM-cellulose chromatography must be
for 1 week at 5”, dialyzed against strating buffer, and then accounted for not only by high molecular weight aggregates but
chromatographed on CM-cellulose. As illustrated in Fig. 5 also by low molecular weight degraded products. The latter may
(compare with Fig. 2), this treatment caused a marked broden- appear throughout the effluent from CM-cellulose columns and
ing of the peaks, indicative of heterogeneity. Recoveries were may not be apparent as discrete peaks. This would result in an
about 70 %. elevated base-line, which would be difficult to detect and would
To characterize further the guanidine-extracted collagen from not be included in the measurement of peak areas.
normal bone, samples were chromatographed on 4% agarose Amino acid analysis and the weight of protein in the fractions
beads (Sepharose 4B, Pharmacia), a molecular sieve material from the agarose column indicated that the absorbance was
with a fractionation range between about 30,000 and 2 x 1O8. a reasonably good measure of collagen concentration. The
The collagens were kept in the denatured state at room tempera- noncollagenous, ultraviolet light-absorbing protein that appeared
ture by the use of 1 M CaCh as the solvent. Fig. 6 shows a as a forepeak in the effluents from CM-cellulose columns was
chromatogram of guanidine-extracted collagen from normal largely confined to the excluded peak and the tail of the low
chick bone compared to a chromatogram of acid-extracted molecular weight material in the effluent from the agarose column.
collagen from lathyritic chick bone. As in the case of CM- Its presence in these regions was indicated by measurable
cellulose chromatography (Fig. l), the lathyritic collagen could amounts of hexosamine and slightly low values for glycine and
be entirely accounted for as LYand /3 components. However, the hydroxyproline in hydrolysates of the proteins.
guanidine-extracted normal collagen showed extreme hetero- To rule out the possibility that the difficulty in redissolving
geneity, with molecular weights ranging from the millions to guanidine-extracted collagen from normal bone and the poor

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

values considerably below the 95,000 expected for a! chains. A recoveries from CM-cellulose columns might be explained in part
major peak in the position of (Y chains (but considerably broad- by reaggregation of the denatured collagen, protein from various
ened), a smaller amount of /3 component, and a peak that may portions of the effluent from the agarose column was isolated by
represent y component were apparent. Recoveries were about dialysis against water, lyophilized, and rechromatographed on the
80% as measured by weighing the protein after dialysis and same column. In each case the fraction redissolved readily in
1 M CaCh and chromatographed in the original position with,
however, a pronounced shifting of the protein toward lower
molecular weights, again indicative of degradation. Recoveries
were complete within experimental error, and no more than a
small percentage of any fraction chromatographed in a position
suggesting reaggregation.
The collagen to gelatin transition of collagen extracted from
the bones of lathyritic animals with 0.5 M acetic acid was studied
in order to determine some of the physical properties of collagen
from a tissue that normally calcifies. Melting curves for chick
bone collagen in solution, as determined by the change in optical
EFFLUENT VOLUME, ml. rotation with increasing temperature, are shown in Fig. 7. The
FIG, 5. CM-cellulose elution pattern of 16 mg of denatured midpoint between maximal and minimal specific optical rotation
lathyritic chick bone collagen (second acid extract) that had been (T,,,) was 39.5” when the collagen was dissolved in 0.5 M acetic
exposed to 5 M guanidine-HCl at 5” for 1 week prior to chromatog- acid, and 41’ when 0.15 M potassium acetate, pH 4.8, was the
raphy. Compare with Fig. 2. solvent. Identical values for the denaturation temperature of
soluble chick skin collagen have been observed.* The high
--1-- ’ I / I / I
0.7 initial specific rotation of the collagen in both solvents, as well as
the sharp transition from collagen to gelatin, indicates that the
0.6 t-
material had been extracted in its native state.
The denaturation temperature of chick bone collagen is about
3” higher than that of mammalian collagens as represented by
rat skin collagen (27). The specific rotations of both the native
and the denatured forms are also higher. These differences are
consistent with the higher total content of imino acid in chick
bone collagen and are expected in view of the known correlations
among body temperature, collagen stability, and imino acid
content (16). The normal body temperature of the chick is about
700 200 300 400 500 600
41”, compared to about 37” for the white rat. The change in
specific rotation on denaturation appears to be slightly greater
EFFLUENT VOLUME, ML
for chick bone collagen than for rat skin collagen but is probably
FIG. 6. Molecular sieve chromatograms (4% agarose, 540-ml within experimental error, suggesting a similar content of helix.
column) of 106 mg of denatured normal chick bone collagen (5 M These data are summarized in Table V. Samples of rat skin and
nuanidine-HCl
a---
extract. -) and 28 mn of denatured lathvritic
chick bone collagen (f&t acid extract,-- - -). The eluent was chick bone collagen were hydrolyzed at the same time and
1 M CaCla, pH 7.5 (0.05 M Tris). The molecular weight of LYis
about 95,069, and that of fl about 190,000. 2 A. H. Kang and K. A. Piez, manuscript in preparation.
Issue of December 10, 1967 Miller, Martin, Piez, and Powers 5487

analyzed in successiveruns to obtain the values for proline and TABLE VI

hydroxyproline content. Otherwisethe rather small differences Content of lyeine and hydroxylysine in fractions of lathyritic chick
would be within experimentalerror. bone collagen
The amino acid composition of the collagen in successive First acid
extracts of lathyritic boneswasidentical except that the content extract
Amiio acid
of hydroxylysine increased in succeeding extracts and was
highest in the insoluble residue. The increasedhydroxylysine
content wasaccompaniedby an equivalent lossin lysine content,
assummarizedin Table VI. Since the more readily extractable Lysine............... 29
collagen is generally of more recent origin than the insoluble Hydroxylysine.......
collagen, theseresults suggestedthat lysine and hydroxylysine Total.
levelsin bonecollagenmight vary normally with age. To inves-
tigate this possibility, diaphysealareasfrom groupsof 24 normal
chicksranging in agefrom 1 to 70days weredecalcifiedin EDTA

2600-

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

l 0.5M Acetic Acid
I I t
o0 0.
O.f5M
f5M Potassium Acetate
0 20 40 60
AGE (days)

Fro. 8. Content of lysine and hydroxylysine in chick bonecol-

lagenasa function of the age of the animal.
I OOOI
22 26 30 34 38 42 46 50
TEMPERATURE, “C
solutions. The decalcified residues were then analyzed for
FIG. 7. Thermal denaturation curves of acid-extracted bone aminoacid compositionafter hydrolysis in 6 N HCl. As shown
collagenfrom lathyritic chicks. The curveswereobtainedduring in Fig. 8, the hydroxylysine content underwent a progressive
stepwiseincreasesin temperature of collagensolutionsat a con- decreasefrom 11 to 6 residuesper 1000total residuesbetween
centration of 1 mg per ml in the solventsindicated. the 1st and the 49th day after hatching. Concomitantly,
lysine content increasedfrom 23 to 23 residuesper 1000 total
TABLE V residues. Little further change could be observedto Day 70.
Comparison of collagens from chick bone and rat skin
DISCUSSION
Property 1 Chick bone’ 1 Rat skin
The experiments described here are interpreted by us as
Proline content?.. .I 115 114 indicating that normal chick bonecollagenis highly cross-linked.
Hydroxyproline content’. 102 93 Only about 4% of the total collagen could be identified as a!
Total imino acid content?. 217 207 chains. This figure assumesthat the collagen that was not
T,b 41.0° 38.0"0 extracted by 5 M guanidine-HCl contained little or no single-
[a]:$0 for denatured collagen. -1050° -825'0
chain component. However, even if this fraction contained as
[a& for native collagen. . -2650' - 2300"~
much a! chain as the extractable fraction, which is highly un-
0 Residues per 1000 residues. likely, the total content of a! chain would be only about 20%.
bAt pH 4.8 in 0.15Mpotassiumacetate. Furthermore, the data showthat the observedCYcomponentsare
c Data are taken from Reference 27. heterogeneousand may have arisen by degradation of larger
5488 Chick Bone Collagen Vol. 242, No. 23

units. In any case, we find a much smaller proportion
chain than reported by Glimcher and Katz (18), who found that
80 to 85% of the total collagen consisted of more than 75% QI
chain, indicating a very low degree of cross-linking.
important to reconcile these data, since the interpretation
Glimcher and Katz and of Glimcher et al. (18, 19) is that bone
collagen is stabilized in the fibril largely by noncovalent forces,
unlike most soft tissue collagens, in which covalent cross-links
appear to be an important factor in intermolecular stabilization.
Probably the major difference can be ascribed to the different
procedures used to prepare the samples. Glimcher and Katz
(18) used stronger denaturing agents and longer periods of
extraction than we used, and they were able to extract 80 to
85% of the collagen. It seems likely that extensive degradation
may have occurred under these conditions. In support of this
possibility are (a) the experiments described here showing that
collagen kept in 5 M guanidine-HCl for 1 week becomes markedly
heterogeneous and (6) the presence of material in the effluent of
the agarose column having molecular weights less than (Y chains.
Furthermore, other experiments have shown the lability
of (Y thawing can mechanically rupture a sufficient number of covalent
bonds to solubilize or disperse a part of the collagen in the form
of high molecular weight aggregates.
It is We do not mean to imply here that covalent cross-links are the
of only, or even the major, force stabilizing fibrous collagen. We do
not have sufficient data or understanding to determine the rela-
tive importance of the many factors involved. However, we
believe that the evidence strongly favors the conclusion, not only
that hard and soft tissue collagens do not differ qualitatively at
the level of structure so far examined, but also that quantitative
differences are in the direction of a greater degree of covalent
cross-linking in bone collagen than in most soft issue collagens.
Chick bone collagen has some unusual properties, as Glimcher
and Katz (18), Glimcher et al. (19, 20), and Francois and Glim-
cher (21) have shown. For example, it does not swell in dilute
acid but it can be extracted by denaturing agents. This is
compatible with a high degree of covalent cross-linking but an
unstable primary structure. Also, the guanidine-extracted
fraction of chick bone collagen has a low content of /3 components,
of while mammalian skin collagen obtained in the same manner

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

/3 components (5,28) and LYchains (8) when kept under denatur- has a high content of /3 components (6, 15). This again suggests
ing conditions. It appears that the polypeptide chains, when instability of denatured bone collagen, and suggests further
not stabilized by secondary or tertiary structure, cleave at weak that intermolecular cross-linking predominates in chick bone
points under the stresses produced by thermal and mechanical collagen while mammalian skin collagens are relatively richer
agitation. We suggest that some of the products of degradation in intramolecular cross-links. This property of chick bone
of cross-linked components have molecular weights and prop- collagen is shared by chick skin collagen,* and is therefore a char-
erties sufficiently close to those of unaltered (Y chains to be con- acteristic of the species rather than of the tissue.
fused with them. The chain structure of chick bone collagen has been examined
A second important difference may be the method of quantita- by Francois and Glimcher (21), with the use of molecular sieve
tive determination. Glimcher and Katz (18) examined their chromatography, free flow electrophoresis, and CM-cellulose
samples by sedimentation velocity in the ultracentrifuge. This chromatography. They concluded that this collagen contains
method does not readily lend itself to precise quantitative three different (Y chains analogous to the al, 0~2, and (~3 chains of
determination with the kind of samples studied; the method is codfish skin collagen (7, 8). Their criteria were first, a differ-
also relatively insensitive to heterogeneity, owing to the large ence in electrophoretic and chromatographic behavior and second,
concentration dependence of sedimentation. The content of (Y an apparent difference in amino acid composition. Inasmuch as
chain may therefore have been overestimated and its very heterogeneity of the (Y chains can be present normally in at least
marked het,erogeneity may not have been apparent. two ways (31, 32) or can arise artifactually as shown here (by
Our observations on lathyrism also offer strong evidence that guanidine treatment and perhaps in other ways), reliance must
chick bone collagen normally is highly cross-linked. It has been be placed on the amino acid analysis.3 In our opinion the ap-
shown that the collagen of animals given a lathyrogen such as p- parent differences in composition between “al” and “a3”
aminopropionitrile becomes readily extractable (17) as a result reported by Francois and Glimcher (21) are within the normally
of an inhibition of cross-linking (12, 29-31). Since it has been expected experimental error. They are certainly much smaller
shown in the present study, as well as elsewhere (17, 20), that than previously shown for the ~yl and a3 chains of codfish skin
the bone collagen of lathyritic animals is readily extractable and collagen (7, 8). Furthermore, the al fraction of acid-extracted
has properties that are very similar to those of soft tissue col- chick bone collagen from lathyritic animals as studied here
lagen from lathyritic animals, it is reasonable to conclude that shows no evidence of heterogeneity. We believe there is not
both hard and soft tissue collagens are stabilized by the same or sufficient evidence to conclude that the (~1 fraction of chick bone
similar cross-links. The only alternative is to ascribe two altera- collagen contains two fundamentally different species. The
tions to the action of @aminopropionitrile, as Glimcher et al. same arguments apply to calf skin collagen, studied by Francois
(20) have done. This seems unnecessarily complex in the light and Glimcher (34) and by Heidrich and Wynston (35).’
of present evidence. An additional finding of the present study is that the lysine to
Glimcher et al. (19) have reported that although bone collagen hydroxylysine ratio of chick bone collagen varies with age while
is not normally extractable by dilute acid, a large proportion can
be solubilized as high molecular weight, but still native, aggre- 8 More generally, a difference in primary structure must be
shown, which may or may not be evident in the amino acid com-
gates after freezing and thawing. They interpreted this result in position. In addition, stoichiometry between the different a
terms of strong noncovalent forces and their rupture by the chain species should be shown to rule out the presence of minor
mechanical introduction of water and acid between molecules. variants (33).
This explanation seemed to be required by their conclusion that 4 The suggestion by Piez (7,s) that the two Cal chains of collagen
might in general be found to be different chains, as shown for cod-
bone collagen has relatively few covalent cross-links. However, fi& skin collagen and suggested for rat skin cdllagen by prelimi-
if one accepts the suggestion offered here that bone collagen is nary data (36), was clearly premature (33), and may in some
highly cross-linked, it is at least as plausible that freezing and measure have led to the conclusions criticized here.
Issue of December 10, 1967 Miller, Martin, Piez, and Powers 5489

the total content of the two amino acids is constant. This ratio 2. OREKHOVICH, V. N., CHPIKITER, V. O., MAZOUROV, V. I., AND
is known to vary in collagens from a number of vertebrate species, KOUNINA, 0. V., Bull. Sot. Chim. Biol.. 42,505 (1960).
3. PIEZ, K. A:, LEWIS, M. S., MARTIN, G: R., AND GROSS, J.,
as well as in collagens extracted from different tissues of the Biochim. Biophys. Acta, 63, 596 (1961).
same species (37). Moreover, a progressive loss of hydroxylysine 4. PIEZ, K. A.. EIONER, E. A., AND LEWIS, M. S., Biochemistw, _
and an equivalent increase in lysine content accompanying 2, 58 (1965).
advancing mineralization of the collagen of turkey tendon have 5. LEWIS. M. S.. AND PIEZ. K. A.. J. Biol. Chem.. , 239. 3336
(1964). ’ ’ ’
previously been observed (38). In the latter study, no altera- 6. BORNPTEIN, P., AND PIEZ, K. A., J. Clin. Invest., 43, 1813
tion in the content of lysine and hydroxylysine in tendon collagen (1964).
of noncalcified areas was observed. The significance of these 7. PIEZ, K. A., J. Biol. Chem., 239, PC4315 (1964).
observations may be found in the carbohydrate which is bound to 8. PIEZ; K. A.; Biochemistry, 4, 2590 (1965).
collagen through the side chain of hydroxylysine, probably in O- 9. ALTOELT, K., HODGE, A. J., AND SCHMITT, F. O., Proc. Nat.
Acad. Sci. U.S.A., 47, 1914 (1961).
glycosidic linkage (39). 10. GRASSMAN, W., HANNIG, K., AND ENQEL, J., Hoppe-Seyler’s Z.
It is conceivable that during early development a different Phusiol. Chem.. 324.284 (1961).
collagen is produced, which gradually replaces the collagen of the 11. VI&A., ANESE;, J.,.AND COHEN, J., Arch. Biochem. Biophys.,
94, 20 (1961).
less mature tissue. If the collagen synthesized after hatching
12. MARTIN, G. R., GROSS, J., PIEZ, K. A., AND LEWIS, M. S.,
contained a greater proportion of its lysyl residues in sequences Biochim. Biophys. Acta, 63, 599 (1961).
unavailable to the hydroxylating enzyme, then less lysine would 13. MARTIN, G. R., PIEZ, K. A., AND LEWIS, M. S., Biochim. Bio-
be hydroxylated. However, if this were the case, it would be phys. Acta, 69, 472 (1963).
expected that the different sequences would cause a recognizable 14. VEIS, A., AND ANESEY, J., J. Biol. Chem., 240.3899 (1965).
15. BORNSTEIN, P., MARTIN, G. R., AND PIEZ, K. A., Science, 144.
difference in composition. A more likely suggestion is that the

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

1220 (1964).
collagen synthesized after hatching is the same as that elaborated 16. HARRINQTON, W. F., AND VON HIPPEL, P. H., Advance. Protein
in the embryonic bone and that the increasing ratio of lysine to Chem., 16, 1 (1961).
hydroxylysine results from an alteration in the ability of the 17. LEVENE, C. I., AND GROSS, J., J. Ezp. Med., 110,771 (1959).
18. GLIMCHER, M. J., AND KATZ, E. P., J. Ultrastruct. Res., 12,
lysine hydroxylase to act on potentially available lysyl residues.
705 (1966).
The latter explanation is supported by the sequence studies of 19. GLIMCHER, M. J., KATZ, E. P., AND TRAVIS, D. F., J. Ultra-
Bornstein (32) on a peptide derived by cyanogen bromide cleav- struct. Res., 13, 163 (1965).
age of cr chains from rat skin and tail tendon collagen. In this 20. GLIMCHER, G. J., FRIBERO, U. A., ORLOFF, S., AND GROSS, J.,
J. Ultrastruct. Res.. 16. 74 (1966).
work evidence was obtained that although a prolyl residue may
21. FRANCOIS, C. J., AN; GLIMC~ER,’ M. J., Biochim. Biophys.
be present in a sequence suitable for hydroxylation, other factors Acta, 133, 91 (1967).
determine the extent of hydroxylation at a given site. Hydroxy- 22. NAGAI, Y., GROSS, J., AND PIEZ, K. A., Ann. N. Y. Acad. Sci.,
lysine, like hydroxyproline, is formed in peptide linkage by the 121, 494 (1964).
same or a similar enzyme (40). 23. PIEZ, K. A.; AND MORRIS, L., Anal. Biochem., 1, 187 (1960).
24. MILLER, E. J.. AND PIEZ. K. A. Anal. Biochem.. 16.320 (1966).
In the present studies on bone collagen, amino acid analyses of 25. PIEZ, K: A., ~~IXSS, E., AND LEWIS, M. S., J. lkol: Chek, 236,
the total collagen showed no difference in over-all prolyl hydroxy- 1987 (1960).
lation as a function of age. These differences would certainly 26. EASTOE, J. E., AND EASTOE, B., Biochem. J., 67,453 (1954).
have been detected had they occurred in the same proportion as 27. PIEZ, K. A., AND CARILLO, A. L., Biochemistry, 3.908 (1964).
28. ENGEL, J., AND BEIER, G., Hoppe-Seyler’s Z. Physiol. Chem.,
that observed for lysyl residues, but would not have been de-
334, 201 (1963).
tected if the change involved only a few residues per cr chain. 29. BORNSTEIN, P., KANG, A. II., AND PIEZ, K. A., Biochemistry, 6,
The amino acid compositions of al and a2 chains differ in a 3803 (1966).
manner similar to that of the corresponding chains from other 30. BORNSTEIN, P., AND PIEZ, K. A., Biochemistry, 6.3460 (1966).
vertebrate collagens (4). The major differences are larger 31. PIEZ, K. A., MARTIN, G. R., KANG, A. H., AND BORNSTEIN, P.,
Biochemistry, 6, 3813 (1966).
amounts of valine, isoleucine, leucine, and histidine and smaller 32. BORNSTEIN, P., J. Biol. Chem., 242, 2572 (1967).
amounts of glutamic acid and alanine in the ar2 chain. 33. BUTLER,W.T.,BORNSTEIN,P.,ANDPIE~,K.A.,F~~.P~~~.,~~,
The denaturation curves of acid-soluble lathyritic chick bone 670 (1967).
collagen obtained in these studies clearly show that 20% of the 34. FRANCOIB. C. J.. AND GLIMCHER. M. J.. Biochem. J.. 102. 148
(1967). ’ ’
total bone collagen can be extracted in the undenatured state. 35. HEIDRICH, H., AND WYNSTON, L. Ii., Hoppe-Seyler’s Z. Physiol.
Since bones from lathyritic animals were as heavily calcified as Chem., 342, 166 (1965).
those from control animals, it is apparent that the deposition of 36. BORNSTEIN, P., AND PIEZ, K. A., Science, 148.1353 (1965).
mineral in the collagenous matrix does not necessarily lead to 37. PIEZ, K. A., AND LIKINS, R. C., in R. F. SOGNNAES (Editor),
denaturation of the protein. Thus, the insolubility of collagen C&i&&m in biological syitems, American Association
for the Advancement of Science. Washineton. 1960. D. 411.
in bones from normal animals in cold 0.5 M acetic acid cannot be 38. LIKINS, R. C., PIEZ, K. A., AND KuN&, k. i.; in R.
attributed to prior denaturation of the protein resulting from F. SOGNNAES (Editor), Calcijkation in biological systems,
mineralization. American Association for the Advancement of Science,
Washington, 1960, p. 143.
REFERENCES 39. BUTLER, W. T., AND CUNNINGHAM, L. W., J. Biol. Chem., 241,
1. RAMACHANDRAN, G. N., Inl. Rev. Connective Tissue Res., 1, 127 3882 (1966).
(1963). 40. UDENFRIEND, S., Science, 162, 1335 (1966).
 Characterization of Chick Bone Collagen and Compositional Changes
Associated with Maturation
E. J. Miller, G. R. Martin, K. A. Piez and M. J. Powers
J. Biol. Chem. 1967, 242:5481-5489.

Access the most updated version of this article at http://www.jbc.org/content/242/23/5481

Alerts:
• When this article is cited
• When a correction for this article is posted

Click here to choose from all of JBC's e-mail alerts

This article cites 0 references, 0 of which can be accessed free at

Downloaded from http://www.jbc.org/ by guest on April 26, 2018

http://www.jbc.org/content/242/23/5481.full.html#ref-list-1