THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 249, No. 10, Issue of May 25, PP. 32263231, 1974

Printed in U.S.A.

[al (III) I3 Human Skin Collagen

RELEASE BY PEPSIN DIGESTION AND PREPONDERANCE IN FETAL LIFE*

(Received for publication, August 20, 1973)

ERVIN H. EPSTEIN, JR.

From the Dermatology Division, Medical Service, San Francisco General Hospital, and the Department of Der- matology, University of California, San Francisco, California 94110

SUMMARY

Human dermis was digested with pepsin under conditions that maintain the helical conformation of the bulk of the collagen. Molecules containing (rl(II1) chains were sep- arated from [al(I)]& collagen by differential salt precipita- tion at pH ‘7.5, with the use of the peptides produced by CNBr cleavage to monitor the composition of each fraction.

The isolated molecules are composed of three al(IJ1) chains, and these chains are cross-linked by disulfide bonds. Exposure to dithiothreitol liberates the three apparently identical chains, and these elute on carboxymethylcellulose chromatography at a position intermediate between al(I) and (r2. The pepsin-resistant portion of these chains is the same size as that of previously isolated collagen chains.

Human dermis was also digested with CNBr, and resultant al(I) and al(II1) peptides were separated chromatograph- ically. The relative quantities of these peptides indicate that al(II1) chains predominate in early fetal skin, but by birth and in later life al(I) chains are approximately 3 times as plentiful.

Collagen extracted from human skin has been well char- acterized in its chain composition (2) and also at the level of CNBr peptides (3, 4). This molecule is composed of two identical cul chains and a differing, a2, chain and is closely homologous to collagen extracted from skin, bone, and tendon of several species including baboon (4), chick (58), rat (g-11), guinea pig (12), calf (13-15), and rabbit (16, 17).

The collagen extracted from cartilage, by contrast, is com- posed of three identical chains that are similar to, but not identical with the Cal chains of the previously studied tissues (18-21). The cartilage chains are now designated al(II), and those extracted from skin, bone, and tendon are now termed cYl(1).

When the residue of human skin remaining after prolonged extraction with neutral salt and dilute acid solutions is sub- jected to CNBr cleavage, peptides not previously observed are

*This work was supported by National Institutes of Health Grant AM-16478. A portion of this work has been published in abstract form (1).

liberated (22). This suggested that a previously undescribed collagen chain must be present in dermis. Because of the similarity of two of these peptides to well characterized regions of the al(I) and al(I1) chains, it seemed clear that this new chain must be related to the al chains, and hence it was tenta- tively designated (~l(I11). Because of their homologies, the new peptides were designated (~1(111)-CB3 and al(III)-CB(4,5). No intact molecules containing this new chain have been found in extracts of normal or lathyritic tissues.

This report describes the isolation of molecules containing ~ul(II1) chains released by pepsin digestion of human skin. The chains are the same length as the al(I) and (~2 chains isolated from the same digests but differ markedly in their elution in carboxymethylcellulose chromatography and in their CNBr peptides. Unlike collagen molecules previously isolated from the skin of vertebrates, these molecules contain chains that are cross-linked by disulfide bonds. In addition, the relative quantities of peptides derived from CNBr digestion of skin have been studied. The ratio of the two types of crl chains varies with the age of the individual from whom t’he dermis was obtained, crl(II1) chains being predominant in fetal life and al(I) chains being the more plentiful in adult life.

MATERIALS AND METHODS

Skin Samples-Human skin used for pepsin digestion was ob- tained from the trunk and proximal extremities of infants at the time of autopsy. Abdominal skin removed at the time of routine autopsy was used for most determinations of chain ratios at different ages. In one adult, skin samples from both the abdomen and the leg were obtained; in one &year-old girl, skin was ob- tained from a normal appearing area of the calf of a leg amputated for treatment of a soft tissue sarcoma of the thigh, and in small fetuses the skin of the whole trunk and proximal extremities was utilized. The fetuses were obtained at hysterotomy performed for therapeutic abortion; most of the newborn babies died of pulmonary failure, and the adults died of a variety of causes, moat commonly the complications of malignant disease or athero- sclerosis.

The subcutaneous fat was scraped off the dermis, and the skin was cut into pieces 5 to 10 mm2. Skin was either lyophilized directly or was extracted at 4” with two portions of 1 M NaCl-0.05 M Tris at pH 7.5 for 1 week and then with five changes of 0.5 M

acetic acid over 4 weeks. The remaining insoluble residue was lyophilized.

Cleavage with Pepsin and Isolation of Molecular Species-One hundred milligrams of pepsin (Worthington, twice recrystallized, 3000 units per mg) were dissolved in 200 ml of 0.4 M acetic acid at 4’. To this was added 1 g of the finely diced insoluble dermis,

3225

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

3226

and the suspension was stirred at 18” for 22 hours. Subsequent manipulations were performed at 4”. Undigested tissue was re- moved by centrifugation for 1 hour at 16,000 X g and lyophilieed. The supernatant was filtered (Whatman No. 1 filter paper), and the collagen was precipitated twice by dialysis against 0.02 M

NazHP04 and then once by dialysis against 0.5 M acetic acid containing 0.86 M NaCl. Each precipitate was dissolved in 0.1 M acetic acid.

The last solution was adjusted to pH 7.5 by dialysis against 0.05 M Tris, pH 7.5, containing 0.52 M NaCl. Fractions were then collected by dialysis against 0.05 M Tris, pH 7.5, containing progressively higher concentrations of NaCl: 0.86, 1.38, 1.71, 2.05, 2.39, and 2.56 M. The material precipitating at 1.71 M was reprecipitated at the same salt concentration. Each precipitate was collected by centrifugation for 1% hours at 35,04Ml X g, dis- solved in 0.5 M acetic acid, dialyzed against 0.5 M acetic acid, and lyophilized.

Cleavage with CNBr-Samples of unextracted skin or of in- soluble dermis (50 to 500 mg) were suspended in 25 to 50 ml of 70% formic acid. Samples of solubilized collagen (40 mg) were dissolved in 25 ml of 70% formic acid. The suspension or solution was flushed with nitrogen, and a weight of CNBr 2 to 3 times the weight of the collagen was added. The flask was stoppered, and the contents were stirred for 4 hours at 35”. Any insoluble material was removed by centrifugation for 30 min at 35” at 35,006 X g and was lyophilized separately. The liberated peptides were then separated from formic acid and CNBr by passage over a column of Bio-Gel P-2 (50 to 100 mesh, Bio-Rad) and were lyophilized.

CM-Cellulose Chromatography-Collagen molecules were de- natured and chromatographed on a column (9 X 90 mm) of CM- cellulose (Whatman No. CM-52, microgranular, capacity 1.0 meq per g) at 43” using an NaCl gradient in sodium acetate as described previously (4), except that the flow rate and gradient volume were reduced for this smaller column to 50 ml per hour and 450 ml, respectively. In some instances, dithiothreitol (Calbiochem) was added to a concentration of 0.001 M to the deaerated buffers.

Peptides liberated by CNBr digestion were chromatographed on the same column (25- to 50.mg samples) or on a column (25 X 90 mm) (lOO- to 200-mg samples) at 43”. Samples were dissolved in 4 to 10 ml of startine buffer CO.02 M sodium citrate. 0.02 M NaCl. adjusted to pH 3.8 with citric acid). The solutions’were clarified by centrifugation at 35,000 X g at 40” for 20 min and passage through a medium porosity sintered glass filter and were then applied to the column equilibrated with the same buffer. Peptides were eluted from the smaller column with a linear gradient pre- pared from 250 ml of starting buffer and 250 ml of limit buffer (0.02 M sodium citrate, 0.20 M NaCl adjusted to pH 3.8 with citric acid) at a flow rate of 50 ml per hour and from the larger column with a gradient volume and flow rate 4 times as large.

The effluent from all columns was monitored continuously at 226 to 230 nm by means of a Beckman DB-GT spectrophotometer; Appropriate fractions were lyophilized, separated from salts by passage over columns of Bio-Gel P-2, and were relyophilized.

&arose Chromatography-Separation and molecular weight determinations of denatured collagen chains were achieved on calibrated columns (1.8 X 230 cm) of aearose beads (Bio-Gel A-5m or A-15m, each‘200 to 400 mesh, fro& Bio-Rad) at i3” (23). Samples were heated to 43” to ensure denaturation before ap- plication to the column. Dithiothreitol, 0.001 M, was added to the eluting solution of 1 M CaClz-0.05 M Tris at pH 7.5 for some chromatograms. Blue dextran 2000 (Pharmacia) was used to determine the excluded volumes of these columns.

Separation, molecular weight determination, and determination of relative quantities of CNBr-liberated peptidea derived from CM-cellulose chromatography were achieved on a column (1.8 X 230 cm) of agarose beads (Bio-Gel A-1.5,200 to 400 mesh) at 23”.

Calculation of Peptide Ratios-Comparison of the relative quantities of peptides present was performed by tracing the appropriate portions of chromatograms onto paper, cutting out the traced outlines, and weighing the individual pieces. These weights were divided by the molecular weights of the peptides to give estimates of the relative molar ratios of the molecules of each species that were present. In the case of al(I)-CBS, the amount of peptide migrating as the uncleaved peptide al(I)- CB(S-3) was added to that migrating as the fully cleaved peptide.

Amino Acid Analysis-Samples were hydrolyzed as described previously (9). Amino acid analyses were carried out on a Beck- man model 117 amino acid analyzer with a 4-buffer elution system similar to that described recently (24). These were kindly per- formed by Mr. Guy Hawkins in the laboratory of Dr. George Martin, National Institute of Dental Research, Bethesda, Mary- land.

RESULTS

Pepsin Digests and Diflerential Salt Precipitation of Com- ponents-After pepsin treatment, 10 to 14% of the original weight of skin digested was recovered in the 16,000 x g pellet. The peptides liberated from this pepsin-insoluble material by CNBr cleavage were chromatographed on CM-cellulose, but only very poor resolution of peptides could be obtained. This precluded the determination of the chain ratios of the material not solubilized by pepsin treatment.

The material twice precipitated by dialysis against 0.02 M

Na2HP04 comprised approximately 20% of the original material digested. No attempt was made to recover the remaining material from the large volumes of dialysis fluid. (In experi- ments in which only pepsin was added without skin to the original incubation mixture, no precipitate was obtained upon dialysis against 0.02 M NkHP04.) During subsequent frac- tionation steps, the solutions were cloudy until dialyzed against 1.37 M NaCl-0.05 M Tris at pH 7.5. At this salt concentration, the supernatant was clear and colorless.

Subsequent dialysis against 1.71 M NaCl precipitated most of the (~1 (III), while most of the al (I) and (~2 remained in solution. After reprecipitation under the same conditions, al(II1) com- prised at least 85% of the material. This was determined by digesting an aliquot with CNBr and separating the resultant peptides (Fig. 1, top, Fig. 2, top). This fraction contained approximately 2% of the original skin that was digested with pepsin. Collagen precipitating at 2.56 M NaCl contained only [al(I)]@2 (Fig. 1, bottom, Fig. 2, bottom).

The CM-cellulose chromatographic pattern of CNBr-re- leased peptides of al(II1) (Fig. 1, top) clearly differed from that of [cyl(I)]2~~2 (Fig. 1, bottom). The presence of al(III)-CB3 (80 ml) and al(III)-CB(4,5) (165 ml) as well as the absence of al(I)-CBS (105 ml) was evident. The peak eluting at 220 to 240 ml contained a previously undescribed peptide of 113 amino acids designated here, for convenience, asal(CB113 (Fig. 1, top, Fig. 2, top). It contained hydroxyproline, and one-third of its residues were glycine (Table I). It had no obvious homology with CNBr peptides of human al(I) or (~2 chains (3, 4), chick (rl(I1) chains (19), or the a chain of bovine lens capsule (26). The final large absorbance peak (275 to 350 ml, Fig. 1, top) contained material that elutes from agarose A-1.5m at a site corresponding to a molecular weight of 80,000, a pat- tern that differed from agarose chromatography of the same region from [orl(I)]z~u2 CNBr peptide chromatograms (Fig. 1, bottom), in which peptides of molecular weights of 60,000 [012- CB(3-5)] and 30,000 (&CB5) were separated (4).

Agarose Chromatography of Denatured Molecules-Material precipitated by dialysis against 0.02 M N&HP04 was chroma- tographed on agarose A-15m (Fig. 3, top). Absorbance peaks were present at elution volumes corresponding to species of the size of crl(1) or 012 chains of soluble human skin collagen (450

ml) and of dimers (p components) from these preparations (385 ml). In addition, however, there was a large peak at the site expected for components comprised of three chains or trimers (7 components) (355 ml). No smaller molecular weight species were present. Chromatography of a fraction purified

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

3227

ELUllON VOLUME, ml

El”TlON VOLUME. ml

FIG. 2. Agarose A-1.5m chromatograms of the CNBr peptides eluting between 215 and 245 ml on CM-cellulose as shown in Fig. 1. Top and boltom are as in Fig. 1. The material eluting at 240 ml isal(CB(83), at 270 ml iscul(I)-CB8, and at 330 ml isal(III)- CB113.

by NaCl precipitation and containing approximately 85% cul(II1) collagen yielded trimers and larger species with only very small amounts of a-sized material (Fig. 3, bottom). An- other aliquot of the 0.02 M N&HP04 precipitate was chroma- tographed on agarose A-5m with eluent to which 0.001 M di- thiothreitol had been added. The trimer and larger molecular weight material decreased markedly in amount, and more material of (Y size appeared (not shown in Fig. 3).

CM-Cellulose Chromatography of Denatured Molecules- Material precipitating on dialysis against 0.02 M N&HP04 was chromatographed on CM-cellulose. Although heterogeneity of the crl and 811 regions was evident, no (wl(II1) peptides were detected after CNBr cleavage of that material. All the al(II1) eluted in the position expected for /3~, and arl(II1) purified by salt precipitation eluted at the same site (Fig. 4, top). When the latter material was then chromatographed on agarose A-5m, approximately one-half eluted at the excluded volume and one- half at a site expected for a trimer (Fig. 5). No material eluted at a site corresponding to a molecular weight smaller than 280,000. Each of these two fractions from agarose A-5m was rechroma- tographed on the same column and eluted with the same solution, to which 0.001 M dithiothreitol had been added; in each in- stance most of the material then eluted at a site expected for (Y chains (Fig. 6). Amino acid analyses of the constituent in the two peaks from the initial agarose chromatogram and of material from the second peak that migrated as a chains on rechroma- tography were similar; each contained one-third glycine, high amounts of hydroxyproline, and significant amounts of cysteine (Table I),

A separate sample of the orl(II1) purified by salt precipitation was chromatographed on CM-cellulose, and the material pro- ducing the single large absorbance peak (Fig. 4, top) was re- chromatographed on the same CM-cellulose column with the

FIG. 1. CM-cellulose chromato- grams of the CNBr peptides from collagen released from dermis by pep- sin digestion and precipitated twice by dialysis against 1.71 Y NaCl-0.05 M Tris, pH 7.5, top, and collagen re- leased from insoluble dermis by pepsin digestion and precipitated once by dialysis against 2.56 M NaCl-0.05 M Tris, pH 7.5, bottom.

TABLE I Amino acid compositions

Amino acid Residuesa

al(III)-CBil3 al(m)b

4-Hydroxyproline. . . .

Aspartic acid. . . . . .

Threonine. . . . . . . . . . .

Serine............ .

Glutamic acid. . . . . . .

Proline. . . . . . . . . . . .

Glycine. . . . . . . . .

Alanine. . . . . . . . .

p$2YS”. . . . . . . . . .

Valine. . . . . . . . . . .

Methionine. . . . . . .

Isoleucine . . . . . . .

Leucine . . . . . . .

Tyrosine . . . . . . . .

Phenylalanine . . . . . .

Hydroxylysine. . . .

Lysine. . . . . . . . .

Histidine. . . . . . .

Arginine. . . . . . . *

Homoserine. . . , . . .

14 6 (5.9)

1 (1.0) 9 (8.9) 9 (8.6)

38 16

2 (2.0)

2 (1.9) 2 (2.1)

1 (1.0) 0.3 4 (4.1)

8 (8.4) 1 (0.7)

121 48 15 41 71

102 355

92 2 (2.4)

16 7 (6.8)

13 21

2 (1.6) 8 (7.8) 5 (5.3)

30 6 (6.1)

46

a Residues per peptide [al(III)-CB113] or residues per 1996 [al (III)], rounded off to the nearest whole number. Actual values are listed in those cases in which fewer than 10 residues were found. Less than 0.2 residues were found of those amino acids not listed.

b These values represent the averages of five separate deter- minations, each of a different preparation. No significant dif- ferences were found for preparations in which the component chains had been separated by exposure to dithiothreitol as com- pared with those in which the disulfide bonds were intact.

c Determination of half-cystine as cysteic acid after performic acid oxidation (25) gave a value of 1.9 residues per 1999.

use of a gradient that differed only by the addition of 0.001 M dithiothreitol. Again, a single peak of absorbance at the same site was seen. No material appeared in the arl(1) or cr2 regions. However, whenthat material was lyophilized, separated from salts by passage over Bio-Gel P-2, and chromatographed on agarose A-5m without dithiothreitol, the reduction in size of most of the material from trimer to a! size was evident.

As a control, acid-soluble human skin collagen containing only [orl(I)]zar2 molecules was chromatographed on CM-cellulose,

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

3228

and the material migrating as /3n was rechromatographed on the same column with the use of a gradient with added di- thiothreitol. The elution position was unchanged, and agarose chromatography of that material revealed only dimer-sized components.

CNBr Digestion of Skin-Recoveries of solubilized peptides released by CNBr digestion of skin averaged 80% of the dried weight of the sample digested. Recoveries from digests of un- extracted samples and from samples of insoluble dermis did not differ significantly, averaging 81 y0 and 79%, respectively. Also, recoveries did not differ significantly among the samples obtained from the various age groups studied: 76% from the fetal skin, 81 y0 from the newborn skin, and 80% from the adult skin. By comparison, CNBr digestion of soluble human skin collagen generally yields a weight of digested peptides 90% of the initial substrate weight. Amino acid analysis of material insoluble after CNBr digestion yielded values for 4-hydroxy- proline of 45 to 76 residues per 1000 residues and for glycine of 122 to 207 residues per 1000 residues, suggesting that approxi- mately one-half of the digestion-resistant material was collagen. This material is also enriched, compared with whole dermis, in the elastin cross-links desmosine and isodesmosine, a finding consistent with the lack of methionyl residues in elastin (27).

Separation and Detection o.f CNBr Peptides of d(I) and cul (Ill)-CM-cellulose chromatograms of the CNBr-derived peptides of insoluble dermis partially resolved the mixture of al(I), cu2, and orl(II1) peptides. As described previously (22),

Fm. 3. Agarose A-15m chromatogram of denatured collagen released from insoluble dermis by pepsin digestion. Top, pre- cipitated twice by dialysis against 0.02 M NasHPOb. Bottom, precipitated twice by dialysis against 1.71 M NaCl-0.05 M Tris, pH 7.5.

0.4

0.3 1

0.2 1

I31 (mu,

/\

the presence of crl(III)-CB3 and al(III)-CB(4,5) was apparent on such chromatograms (Fig. 7, middle, Regions A and B, re- spectively). Agarose chromatography of the peptides eluting in Region C (Fig. 7, middle) separated well al(I)-CB(8-3) and al(I)-CB8 from orl(III)-CB113 (Fig. 8).

Chain Ratios in Various Skin Samples-CM-cellulose chroma- tograms of CNBr peptides from fetal, newborn, and adult skin demonstrated differences in the ratios of crl(1) to cul(II1) pep- tides with a relative abundance of (rl(III)-CB3 and cul(III)- CB(4,5) in the younger specimens and a relative paucity of these peptides in the adult tissue (Fig. 7). This can be seen in Fig. 7 most directly by comparing the size of Peak A with the size of the peak that follows it. The latter contains the homologous peptide al(I)-CB3. The (~1(111)-CB3 peak (A) is

8 Excluded

2 2 02 4 1 oi

Em c-4 0.1 2-

z In., is0

.._ .I 175 200 225 250 275 300

ELUTION VOLUME, ml

FIG. 5. Agarose A-5m chromatogram of collagen released from insoluble dermis by pepsin and purified by differential salt pre- cipitation at 1.71 M NaCl-0.05 M Tris, pH 7.5, and subsequent CM-cellulose chromatography (Fig. 4, top). Excluded, 8, and LY mark the expected elution volumes on this column for blue dextran 2000, p components, and Cal or (~2 chains of 0.5 M acetic acid-ex- tracted human skin collagen.

tsI

$:[ : : r , & c

150 175 200 225 250 275 300

ELUTION VOLUME, ml

Fm. 6. Agarose A-5m chromatogram of material eluting at 175 to 190 ml as shown in Fig. 5. Conditions were the same as in Fig. 5 except that 0.001 M dithiothreitol was added to the eluting fluid.

Fw. 4. CM-cellulose chromato- gram of denatured collagen. Top, released from insoluble dermis bv pepsin digestion and precipitated twice by dialysis against 1.71 M NaCl- 0.05 M Tris, pH 7.5. Middle, released from insoluble dermis by pepsin di- gestion and precipitated by dialysis against 2.56 M N&1-0.05 M Tris, pH 7.5. Bottom, 0.5 M acetic acid-ex- tracted human skin collagen. S = sample applied; G = gradient started.

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

3229

TABLE II

Ratios of d (I) to al (III)

06

t II

04

02

0

Frl 7. CM-cellulose chromatograms of the CNBr peptides from insoluble human dermis from (top) a fetus, (middle) a new- born infant, and (bottom) an adult. S = sample applied; 0 = gradient started.

0

'*I N&m

0 '05 325 3x) 375 400 425

EmuEN VOLUME, ml

FIG. 8. Agarose A-1.5m chromatograms of the CNBr peptides eluting in region C from CM-cellulose. Samples are derived from the insoluble human dermis of (top) a fetus, (middle) a new- born infant, and (bottom) an adult.

larger than the al(I)-CB3 peak in the fetus but smaller in the newborn and adult specimens.

Because agarose chromatography of Region C generated relatively large and well separated peaks of absorbance on a flat base-line, the ratio of peptides in this region was calculated for each skin sample. It was assumed that the molar ratio of the peptides al(I)-CB8 to crl(III)-CB113 was the same as the ratio of al(I) to (wl(II1). The molar ratios of crl(1) to al(II1) calculated from tracings of these agarose chromatograms clearly confirmed what was evident on CM-cellulose chromatography: the relative preponderance of crl(II1) in fetal skin and the decline in its amount during intrauterine and postnatal matura- tion (Fig. 8, Table II). This was true whether one considered whole unextracted skin or skin that was insoluble in neutral salt and dilute acetic acid solutions.

Several experiments were performed to assess the repro- ducibility and reliability of these determinations.

1. Two aliquots of skin from one newborn baby were separately digested with CNBr and chromatographed on CM-cellulose. Then several agarose chromatograms were performed on ma- terial derived from Region C of each of the two CM-cellulose

Age of patie&

15-Week fetus l&Week fetus 19-Week fetus 22-Week fetus 32-Week fetus Newborn infant

3 Months 8 Years 20 Years 30 Years 40 Years 43 Years 56 Years 60 Years 63 Years 78 Years

Number of patients

2 1 1 2 1 9

1 1 1 1 2 1 1 1 1 2

Ratiob

0.8 1.0 1.6

2.7 2.5, 2.6, 2.7

2.6 3.6

2.8

3.1 2.7, 4.1

Insoluble dermis

0.6

0.8, 1.1

1.4, 1.5, 1.6, 1.6, 1.6, 1.7, 1.7, 1.8, 1.8, 1.8, 1.8, 1.8, 1.9, 1.9

3.0 3.3 2.8 2.7, 3.0 2.9

2.6

2.7, 3.1

0 Fetal ages were approximate and were estimated both from the time elapsed from the mother’s last menstrual flow, when this was known, and from the length and weight of the fetus. The amount of skin obtained from a single 12-week-old fetus studied was insufficient to permit an assessment of chain ratio.

b Ratios were calculated from agarose chromatograms of pep- tides eluting in region C from CM-cellulose chromatography.

chromatog-ams. The molar ratios of arl(1) to al(II1) calcu- lated from the duplicate agarose chromatograms were 1.4 : 1, 1.6:1, 1.7:1, and 1.8:l from one digest, and the ratios from the separate CNBr digestion, CM-cellulose chromatogram, and duplicate agarose chromatograms were 1.7 : 1, 1.7 : 1, and 1.8 : 1.

2. Peptides derived from agarose chromatography of Region C were collected, desalted, and weighed. Using the measured weights, a molar ratio of al(I) to al(II1) of 1.8: 1 was calculated. By comparison, a molar ratio of 2.0: 1 was calculated from the weights of the appropriately traced peaks of absorbance, thus suggesting that the absorbance was a reliable estimate of the total amount of each peptide present.

3. Duplicate CM-cellulose chromatography was performed on aliquots from a single CNBr digest. In one instance, Region C was limited to 10 9.5-ml fractions (95 ml) ; in the other in- stance, two 9.5-ml fractions on either side of Region C were added to Region C (total volume 133 ml). The molar ratios of al(I) to al(II1) calculated from agarose chromatograms of these samples were 1.7 : 1 and 1.9 : 1, respectively, thus suggesting that variability in selecting the limits of Region C did not cause marked changes in the molar ratios calculated.

4. One skin sample was divided into two portions. One aliquot was lyophilized directly; the other aliquot was stirred in 0.5 M acetic acid for 1 month, following which the residue and the acid-extracting solution were lyophilized together. Each aliquot was subjected to CNBr digestion, CM-cellulose chroma- tography, and agarose chromatography of Region C. The calculated molar ratios agreed closely, being 2.5: 1 for the directly lyophilized sample and 2.2:1 for the acid-treated skin, thus suggesting that prolonged exposure to dilute acetic acid alone did not of itself change the observed chain ratios.

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

3230

5. Insoluble dermis from the abdomen and thigh of a 40.year- old woman gave calculated molar ratios of 3.0: 1 and 3.3:1, respectively. Unextracted skin from the back and chest of one newborn infant gave molar ratios of 2.2 : 1 and 1.8 : 1, respectively. These results suggest that the ratios of chains do not vary markedly between at least these different sites. Also, no dif- ference was observed in chain ratios in dermis from males or females of the same age.

DISCUSSION

Isolation of [oll(ZZZ)lB-This paper details the isolation from pepsin digests of human dermis of molecules containing cul(II1) chains. Although experiments reported here were performed with material released by digestion at 18”, similar results were obtained when the digestion was performed at 4” for the same length of time.

It has been evident for some time that the bulk of the collagen molecule is resistant to pepsin digestion so long as the triple helix is intact. Pepsin does cleave nonhelical terminal segments from the molecule (28). The liberation of most of the collagen from otherwise insoluble tissue by pepsin digestion is compatible with current concepts that much of the covalent cross-linking between collagen molecules occurs at nonhelical sites (29-33). Similar liberation of collagen from basement membrane (34) and cartilage (24) has been effected by pepsin digestion of otherwise insoluble preparations.

The preservation of the helix in the material released by pepsin digestion of skin is indicated by the resistance to the initial pepsin digestion and by precipitation on dialysis against 0.02 M N&HP04 (35). Also indicative are the ability to form segment long spacing (SLS) crystalloids,’ and the typically high optical rotation with loss upon heating to a temperature (36”) that is t,he same as for [al (I)lza2 isolated from pepsin digests2

The finding of significant amounts of cysteine was unexpected. This amino acid has been found in vertebrate collagens only in basement membrane (34) and in the nonhelical NHz-terminal extensions of procollagen (36-40). It is unlikely that the (rl(II1) could be derived from basement membrane of epidermis or epidermal appendages or of dermal blood vessels, since it comprises approximately one-quarter to one-half of the total skin collagen. This quantity is more than could be ascribed to the amount of basement membrane recognizable histologically. Also, the basement membrane preparations studied have been very rich in hydroxylysine, whereas (rl(II1) contains only as much of this residue as is found in [~ul(I)]zcu2 (3, 4). Lastly, the peptides derived by CNBr digestion of basement membrane do not resemble those derived from cul(II1) (26).

It is unlikely that the cysteine is present in a nonhelical procollagen peptide, since the latter are removed by pepsin (37, 41). The presence of some /3 components of [0~1(1)]2012 isolated from the same pepsin digests, however, suggests the possibility that some nonhelical portion may survive this diges- tion, because the intramolecular bond linking two chains in soluble collagen is located in the nonhelical NHz-terminal region. Alternatively, the persistence of 0 components may be due to intermolecular cross-linking at a helical site. Since the mo- lecular weight of the al(II1) collagen falls from approximately 280,000 to approximately 93,000 on addition of 0.001 M dithio- threitol to the eluting buffer on agarose chromatography, some

1 Ervin H. Epstein, Jr., and Kimie Fukuyama, unpublished observations.

2 Ervin H. Epstein, Jr., and Jen T. Yang, unpublished ob- servations.

of the cysteine residues must form disulfide cross-links. It is tempting to speculate that the inability to ext,ract oil (III) from the skin even of lathyritic animals may be due to the persistence of disulfide bonds that are intermolecular. This possibility is supported by the liberation of a-sized moieties on exposure to dithiothreitol of the material larger than trimers (Fig. 5, ex- eluded material).

Cysteine is also present in the collagen of Ascaris lumbricoides cuticle. The chains of that collagen form an intrachain helix by folding back on themselves and are linked together by disul- fide bonds (42, 43). The chains are, however, smaller than the (rl(II1) chains isolated from human skin, having a molecular weight of 62,000.

The chain composition of this collagen is probably [orl(III)]~. This is suggested by the failure to liberate chains migrating in the al(I)-crl(I1) or a2 regions on CM-cellulose chromatography performed with dithiothreitol added to the gradient buffers. As judged by subsequent agarose chromatography, trimers are cleaved to LY chains under these conditions. The CR&cellulose chromatogram of the CNBr-released peptides from the al (III)- containing fraction does contain material that elutes in the position expected for ol2-derived peptides. However, the mo- lecular weight of 80,000, with absence of 60,000 and 30,000 mo- lecular weight peptides, suggests that this material is not derived from the known (r2 chain. Since only (rl(II1) chains seem to be present, and since most of the cul(II1) is present in these prep- arations in a moiety of a molecular weight of 280,000 and can be dissociated by dithiothreitol to a molecular weight of approxi- mately 93,000, the most likely chain composition for this mole- cule is [al(III)],.

The separation of [al(I)lz(r2 from [al( by differential salt precipitation at pH 7.6 in phosphate buffer has been reported (44). The former precipitates at 2.2 M NaCl; the latter remains in solution and can be precipitated at higher NaCl concentration. Using Tris buffer at pH 7.5, it has been demonstrated here that pepsin-digested [(rl(III)13 precipitates at the lower salt con- centration of 1.71 h% NaCl. This suggests that it might be pos- sible to separate all three of these types of collagen by differential salt precipitation were such a separation necessary.

al (III) Preponderance in Fetal Skin-These studies demon- strated that ~ul(II1) chains are the most plentiful type of col-

lagen chain in the skin of early fetuses, but comprise a much smaller fraction of skin collagen in children and adults. The reproducibility of the methods used seems quite adequate for this conclusion.

The relative paucity of (rl(III)-CB113 in adult skin is not due solely to a dilutional effect with maintenance of the same total quantity of crl(II1) with an increasing amount of al(I), because the quantity present in a small fetus is many times less than the quantity of al(II1) present in even a small part of the skin of a child or an adult. It is unknown whether this ratio is controlled by differing rates of synthesis, of degradation, or both. The possibility that different collagen molecules may be degraded in- dependently is suggested by the demonstration that [(~1(11)]~ molecules of cartilage resist attack by collagenase preparations that are active against [cr1(I)lza2 molecules (45).

Trelstad et al. found in neutral salt extracts of skin of lathyritic chick embryos a collagen species containing no a2 chains (46). They postulated on the basis of amino acid analyses that this was composed of chains differing from al(I) and (rl(II) and proposed the designation al(II1). This species was not found in extracts from skin of lathyritic 3-week-old chicks. The exact identity of the molecule they described is not certain, but it is

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

likely that it is not the same as the al(II1) described herein be- cause there are significant differences in the amino acid com- position of the two preparations3

The conclusion that skin collagen varies with age was also reached by study of the reducible aldimine cross-links of skin from fetuses and newborns (47). In fetal skin from humans, chicks, or calves, the major such cross-link is derived from 2 hydroxylysine residues, whereas in more mature skin the major such cross-link is derived from 1 hydroxylysine and 1 lysine residue. One explanation suggested for this difference in cross- links was a difference in primary structure of the collagen at different ages, but age-related differences in hydroxylation of the same lysine residue could not be excluded. No evidence is available, however, that the two cross-links are indeed derived from [(ul(I)lza2 and [arl(III)], molecules.

The realization that various tissues may contain genetically distinct collagen molecules is a relatively recent one. At this time, no clear data have been presented linking differences in collagen molecules to differences in normal tissue structure or function. The finding of a preponderance of al(III) chains in early fetal skin and a relative paucity of these chains postnatally might suggest that their presence could be important in facilitat- ing remodeling and growth of the skin. However, other tissues including tendon and deep muscle fascia contain only [al(I)lpa2 collagen molecules in children, thus suggesting that al(II1) chains are not necessary for growth at least in these tissues.*

If molecules containing (rl(II1) functioned in maintenance of the epidermis, one might also expect a preponderance of these molecules in the thinner skin of the fetus, where relatively more of the dermis lies close to the epidermis. That explanation would also account for its absence in tissues such as tendon that do not support an epithelium, as well as its presence in tissues with an epithelium or endothelium, including rat lung, human placental membranes, and human aorta.4 The same considera- tions would apply if [arl(III)] 3 were produced by the epidermis, it now being clear that epithelia are capable of secreting collagen (48). Studies to localize the al(II1) chains within the dermis are now in progress.

REFERENCES

1. EPSTEIN, E. H., JR. (1971) Clin. Res. 19, 359 2. BORNSTEIN, P., AND PIEZ, K. A. (1964) J. Clin. Invest. 43,

1813-1823 3. CLICK, E. M., AND BORNSTEIN, P. (1970) Biochemistry 9,4699-

4706 4. EPSTEIN, E. H., JR., SCOTT, It. D., MILLER, E. J., AND PIEZ,

K. A. (1971) J. Biol. Chem. 246, 17181724 5. MILLER, E. J., LANE, J. M., AND PIEZ, K. A. (1969) Bio-

chemistry 8, 30-39 6. LANE, J. M., AND MILLER, E. J. (1969) Biochemistry 8, 2134-

2139 7. KANG, A. H., PIEZ, K. A., AND GROSS, J. (1969) Biochemistry

8, 15061514 8. KANG, A. H., IGARASHI, S., AND GROSS, J. (1969) Biochemistry

8, 32063204 9. BUTLER, W. T., PIF,Z, K. A., AND BORNSTEIN, P. (1967) Bio-

chemistry 6, 3771-3780 10. FIETZEK, P. P., AND PIEZ, K. A. (1969) Biochemistry 8, 2129

2133

3 Robert L. Trelstad, personal communication. 4 Ervin H. Epstein, Jr., unpublished observations.

3231

11. BORNSTEIN, P. (1969) Biochemistry 8, 63-71 12. CLARK, C. C., AND BORNSTEIN, P. (1972) Biochemistry 11,

1468-1474 13. RAUTXRBERG, J., AND K~~HN, K. (1971) Eur. J. Biochem. 19,

398-407 14. FIF,TZEK, P. P., M~~NcH, M., BREITKREUTZ, D., AND KUHN, K.

(1970) Fed. Eur. Biochem. Sot. Letl. 9, 229-231 15. VOLPIN, D., AND VEIS, A. (1971) Biochemistry 10, 1751-1755 16. BORNSTEIN. P.. AND NESSE:. R. (1970) Arch. Biochem. BioDhvs.

138, 4433450 ’ ‘ L 1

17. BECKER, U., AND TIMPL, R. (1972) Fed. Eur. Biochem. Sot. Lelt. 27, 85-88

18. MILLER, E. J. (1971) Biochemistry 10, 1652-1659 19. MILLER, E. J. (1971) Biochemistry 10, 30363035 20. TRELSTAD, R. L., KANG, A. H., IGARASHI, S., AND GROSS, J.

(1970) Biochemistry 9, 4993-4998 21. STI~AWICH, E., AND NIMNI, M. E. (1971) Biochemistry 10,

3905-3911 22. MILLER, E. J., EPSTEIN, E. H., JR., AND PIEZ, K. A. (1971)

Biochem. Biophys. Res. Commun. 42, 1024-1029 23. PIEZ, K. A. (1968) Anal. Biochem. 26, 305312 24. MILLER. E. J. (1972) Biochemistrv 11. 4903-4909 25. HIRS, C: H. W.’ (1967) Methods Ekzy&oZ. 11, 59-62 26. KE:FALIDES, N. A. (1972) Biochem. Biophys. Res. Commun. 47,

1151-1158 27. SANDBERG, L. B., WEISSMAN, N., AND SMITH, D. W. (1969)

Biochemistry 8, 294G-2945 28. DRAKE, M. P., DAVISON, P. F., BUMP, S., AND SCHMITT, F. 0.

(196(i) Biochemislry 6, 301-312 29. ZIMMERMANN, B. K., PIKKARAINEN, J., FIETZEK, P. P., AND

KUHN, K. (1970) Eur. J. Biochem. 16, 217-225 30. MILLER, E. J. (1971) Biochem. Biophys. Res. Commun. 46,

444-451 31. DIXIT, S. N., AND BENSUSAN, H. B. (1973) Biochem. Biophys.

Res. Commun. 62, 1-8 32. KANG, A. H. (1972) Biochemistry 11, 1828-1835 33. EYRE, D. R., AND GLIMCHER, M. J. (1973) Biochem. Biophys.

Res. Commun. 62, 663-671 34. KEFALIDES, N. A. (1971) Biochem. Biophys. Res. Commun. 46,

226234 35. LAYMAN, D. I,., MCGOODWIN, E. B., AND MARTIN, G. R.

(1971) Proc. Nat. Acad. Sci. U. 8. A. 68, 454458 36. BORNSTEIN, P., VON DER MARK, K., WYKE, A. W., EHRLICH,

H. P., AND MONSON, J. M. (1972) J. Biol. Chem. 247, 280% 2813

37. BURGESON, R. E., WYKE, A. W., AND FESSLER, J. H. (1972) Biochem. Biophys. Res. Commun. 48,892-897

38. UITTO, J., JIMENEZ, S. A., DEHM, P., AND PROCKOP, 1). (1972) Biochim. Biophys. Acta 278, 198-205

39. VON DER MARK, K., AND BORNSTEIN, P. (1973) J. Biol. Chem. 248, 2285-2289

40. FURTHMAYER, H., TIMPL, R., STARK, M., LAPIERE, C. M., AND KUHN, K. (1972) Fed. Eur. Biochem. Sot. Lett. 28, 247- 250

41. BELLAMY, G., AND BORNSTEIN, P. (1971) Proc. Nat. Acad. Sci. U. S. A. 68, 1138-1142

42. MCBRIDE, 0. W., AND HARRINGTON, W. F. (1967) Biochemistry 6, 1484-1498

43. MCBRIDE, 0. W., AND HARRINGTON, W. F. (1967) Biochemistry 6, 1499-1514

44. TRELSTAD, It. L., KANG, A. H., TOOLE, B. P., AND GROSS, J. (1972) J. Biol. Chem. 247, 6469-6473

45. ROBERTSON, P. B., AND MILLER, E. J. (1972) Biochim. Biophys. Acta 289, 247-250

46. TRELSTAD, R. L., KANG, A. H., AND GROSS, J. (1971) Fed. Proc. 30, 1196

47. BAILEY, A. J., AND ROBINS, S. P. (1972) Fed. Eur. Biochem. Sot. Lelt. 21, 330-334

48. NEWSOME, D. A., AND KENYON, K. It. (1973) Develop. Biol. 32, 387400

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Ervin H. Epstein, Jr.PREPONDERANCE IN FETAL LIFE

Human Skin Collagen : RELEASE BY PEPSIN DIGESTION AND31(III)]α[

1974, 249:3225-3231.J. Biol. Chem. 

  http://www.jbc.org/content/249/10/3225Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

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

  http://www.jbc.org/content/249/10/3225.full.html#ref-list-1

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

by guest on April 26, 2020

http://ww

w.jbc.org/

Dow

nloaded from