THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 255, No. 19, Issue of October 10, pp, 9427-9433, 1980 Printed in U.S.A.

Type I Collagen in Solution STRUCTURE AND PROPERTIES OF FIBRIL FRAGMENTS*

(Received for publication, August 3, 1979, and in revised form, January 14, 1980)

Frederick H. Silver$ and Robert L. Trelstad From the Department of Pathology, Shriners Burns Institute and Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114

We have measured the diffision coefficient and weight average molecular weight of type I collagen fibril fragments obtained by acid extraction of rat tail tendons and neutral extraction of lathyritic chick skin, using laser light scattering techniques. The molecular weight and translational diffusion coefficient were found to be 8.05 f 0.40 X 10’ and 0.450 -C 0.04 X lo-’ cm2/s, respectively, for preparations which contain ag- gregates in 0.01 M HCl, and 2.82 * 0.20 X 10’ and 0.780 f 0.04 x lo-’ cm2/s for collagen single molecules. Using these data, as well as the theoretical difhsion

coefficient for a single collagen molecule, models for various staggering modes and aggregate mixtures were developed in an attempt to understand the structure of the fibril components present in solution. It was found that several models with 1 to 4 D (D = 67 nm) staggers fit the experimental data for the aggregated prepara- tions. Analysis of the end-to-end distance of negatively stained segment long spacing crystallites prepared from solutions containing fibril fragments and the banding pattern of positively stained segment long spacing crystallites suggest that collagen solutions con- tain linear aggregates with 4 D and possibly 4.4 D staggers in agreement with light scattering data. Ther- mal denaturation studies demonstrate that the agpar- ent melting point of cross-linked linear aggregates, 33.5 2 0.5”C, is identical with that of single molecules at pH 2.0. We conclude that a linear filament with predomi- nant 4 D stagger is a basic unit of type I collagen fibrillar structure.

The structure of collagen in solution has been studied extensively both at the molecular (1-6) and supramolecular levels (2-6). Early studies focused on the size and shape of collagen single molecules (1,2), while some of the later studies attempted to focus on supramolecular forms present in solu- tion (3-6). From these efforts, we know that extracts of type I collagen containing tissues are mixtures of single molecules and high molecular weight aggregates. Since these fragments are derived from native collagen fibrils, structural analyses of these components may generate more information which can be used to develop a better understanding of local fibril structure.

* This work was supported by Grant PCM-7818903 from the Na- tional Science Foundation, Grants HL-18714 and GM 20007 from the National Institutes of Health, National Service Award No. 5 T32 HD 07092-01, and by the Shriners Burns Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. + To whom correspondence should be addressed, at the Shriners Burns Institute, 51 Blossom St., Boston, Mass. 02114.

Several models of the substructure of collagen fibrils based on theoretical and experimental considerations have been proposed. The microfibril, the smallest structural unit, con- tains from 4 to 8 collagen molecules in cross-section (7-10). Other evidence suggests that lateral order within a fibril is not long range (11) and may be associated with the asymmetry of the molecule. In vivo collagen aggregation may begin before secretion

since intracellular vacuoles containing aggregates with lengths of 2 or more collagen molecules (12, 13) have been observed. Similar aggregates have been observed both intracellularly and extracellularly in a wide variety of tissues (14,15). In vitro studies of type I collagen fibrillogenesis indicate that initiation of collagen fibril formation occurs by the conversion of single molecules into 4 D staggered dimers and trimers, suggesting that the 4 D stagger may be thermodynamically preferred over other staggering modes (16). Further aggregation into D staggered aggregates is a secondary event (16).

The purpose of this investigation has been to characterize the physicochemical and ultrastructural properties of collagen solutions derived from solubilized type I collagen fibrils and to assess the organization of such fragments in respect to their possible mode of packing in the native fibril.

MATERIALS AND METHODS’ Rat Tail Tendon Collagen-Acid-soluble rat tail tendon collagen

was prepared (see Fig. 1) by dissolving tendons from young rats in 0.01 M HCI at 4°C for 4 h followed by centrifugation for 30 min at 30,000 X g. The supernatant was sequentially filtered through 0.8-, 0.65, and 0.45-micron Millipore filters. SDS’ electrophoresis and amino acid analyses of this preparation showed that it contained predominantly type I collagen a chains and p and y components. The relative proportions of a, p, and y components were determined by chromatographing heat-denatured samples in 1.0 M CaCL on an agarose A-5m column (2.0 X 100 cm) and computing the peak areas of the effluent monitored at 230 n~ in a recording spectrophotometer. Peak identification was achieved by estimating the molecular weight using laser light scattering techniques.

Lathyritic Chick Skin Collagen-Extracts of lathyritic chick skin with 0.4 ionic strength potassium phosphate buffer, pH 7.6, were purified by repeated NaCl precipitation at both neutral and acidic pH. The purified collagen was desalted by dialysis at 4°C after resolubilization in 0.01 M HCI. Evaluation of chemical purity was done by amino acid and hexosamine analysis on HC1-hydrolyzed aliquots.

’ Portions of this paper (including part of “Materials and Methods” and Tables 1 through 7) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biolog- ical Chemistry, 9650 Rockville Pike, Bethesda, Md. 20014. Request Document No. 79M-1568, cite author(s), and include a check or money order for $1.65 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

The abbreviations used are: SDS, sodium dodecyl sulfate; SLS, segment long spacing crystallites.

9427

Type 1 Collagen in Solution

Preparation of Segment Long Spacing Crystallites (SLS)"sLS crystallites of rat tail tendon collagen were prepared by dialysis of collagen solutions in HC1, pH 2.0, at 4°C uersus 0.01 M ATP in HC1, pH 2.0, overnight. A drop of the dialysate was placed on carbon/ Formvar-coated copper grids.

Transmission Electron Microscopy-SLS of acid-soluble rat tail tendon collagen were stained negatively with phosphotungstic acid (PTA), pH 7.6, and positively with saturated aqueous uranyl acetate and viewed with a Philips 300 transmission electron microscope. The end-to-end length of monomeric collagen SLS was determined from negatively stained samples. The locations of the NH2 and COOH termini of negatively stained SLS were determined by measuring a molecular length from each free end. In most cases, the overlap of negatively stained polymeric SLS was obvious by the positions of the NH2 and COOH termini of 2 adjacent molecules, as is seen in Fig. 6.4.

The molecular overlap was also determined by measuring the distance between Bands 3 and 4 and 51 and 52 (17) of positively stained polymeric SLS. Measurements on 10 different preparations of monomeric SLS indicated that Bands 3 and 4 and 51 and 52 are both about 7% from the ends of the molecule, while the distance between these bands is about 86% of the molecular length.

RESULTS

The relative composition of the high molecular weight acid- soluble rat tail tendon as defined by chromatography on A- 5m was as follows: a chains, 33.6%; p components, 42.5%; and y components, 23.9%; a recovery of 88% was obtained. Amino acid analysis was typical of type I collagen; specifically, there were 320 residues of glycine, 2 residues of tyrosine, 93 residues of hydroxyproline, and 128 residues of proline. Noncollage- nous glycoproteins and proteoglycans were less than 1% based on galactosamine and glucosamine analysis. Amino acid anal- ysis of the lathyritic chick skin preparation gave the following composition: glycine, 335 residues; tyrosine, 4 residues; proline, 118 residues; and hydroxyproline, 88 residues. SDS slab gels showed the presence of predominantly type I a chains; how- ever, /3 components were present in the lathyritic preparation.

In a previous study (16), we prepared neutral solutions of single collagen molecules from extracts of lathyritic type I collagen by centrifugation and filtration through a 0.22-micron Millipore filter; here, we want to characterize the physical forms of collagens extracted by dilute acid. To accomplish this, it was essential to develop routine handling procedures to separate and characterize the different solution forms. In order to minimize the number of steps involved in the purifi- cation procedure, it was necessary to study a tissue that was abundant, easily isolated, and as homogeneous as possible. For these reasons, rat tail tendon collagen was chosen.

Molecular Weights-As shown in Fig. 1, about 61% of the dry tendon weight is soluble in 0.01 M HC1 at 4'C. Sequential filtration through OB-, 0.65- and 0.45-micron filters results in a solution with a weight average molecular weight of 8.05 * 0.40 X (see Fig. 2), which represents 55% of the tendon by weight. This preparation will be referred to as a high molecular weight fraction. Further filtration through a 0.22- micron Millipore filter resulted in a single molecule fraction, molecular weight 2.82 * 0.20 X which represented 28% of the initial tendon weight. The plots, for determination of weight average molecular weight, are shown in Fig. 2 for high molecular weight and single molecule fractions. The negative slope of Fig. 2 suggests that attractive forces occur between both single molecules and aggregates in acid solution (20).

Translation Difusion-Translation diffusion coefficients, of acid-extracted tendon preparations were determined

by photon correlation of scattered laser light. The correct base-line of the homodyne correlation function was signifi- cantly dependent on the sample time (see Equation la) at concentrations above 1 mg/ml as if a continuous spectrum of aggregates of different sizes were present in solution. At con- centrations below 0.05 mg/ml, the base-line of the normalized

Tendon 1 g I HC1, pH 2.0

4"C, 4 h

Solubyized tendon

Centrifugation - discard pellet

1 30 min @ 30,000 g

Supernatant

Collagen solution Overall recovery of 0.609 g

Recovery 1 Filter through "91% 0.8-, 0.65-, and 0.45-p fdters

High molecular weight fraction

(A?r = 8.05 f 0.40 X lo5, D20.w = 0.45 f 0.04 x 10" cm2/s)

Overall recovery

0.554 g

Recovery I Filter through =M% 0.22-p filter

Single molecule fraction

(ar = 2.82 * 0.20 X lo5, DX).,,, = 0.80 f 0.04 X 10" cm2/s)

Overall recovery

0.277 g FIG. 1. Purification procedure for high molecular weight

and single molecule fractions of acid-soluble (HCl, pH 2.0) rat tail tendon collagen.

40 r

Bi@ 15

lo>

mr - ao5,ooo

5

0 0 0.1 0.2 0.3 0.4 0.5 0.6

coNcENruArloAf /mq/ml /

FIG. 2. Kc/Rs, versus concentration for high molecular weight (0) and single molecule fractions (0) of pH 2.0 HC1- soluble rat tail tendon collagen. All measurements were made at a scattering angle of 4' with respect to the transmitted beam at a temperature of 4°C with a Chromatix KMXS light scattering device. The high molecular weight fraction was filtered through a 0.45-p filter, whereas the low molecular weight fraction was further filtered through a 0.22-p filter.

correlation function was independent of sample time at sample times of 5 and 10 ms. The normalized correlation function at 2, 5, and 10 ms could be superimposed to form a single correlation function which exponentially decayed to a base- line as shown in Fig. 3. The average diffusion coefficient, was determined from the decay rate of the correlation func- tion. The translational diffusion coefficient of the high molec- ular weight fraction at zero concentration was found to be 0.45 f 0.04 X cm2/s (see Fig. 4).

The translational diffusion coefficient of the single molecule preparation was determined as a function of concentration on

Type I Collagen in Solution 9429

samples of known molecular weight (see Fig. 4). D20.,c extrap- olated to zero concentration was 0.780 f 0.04 X 10" cm'/s.

Thermal Stability-Thermal transition temperatures of high molecular weight and single molecule fractions were determined by measuring the temperature dependence of the molecular weight. As shown in Fig. 5, the apparent melting

' O K

P \ : $ 099C \"\x &

0 E ~- B l ' 098C -Q"'""t&*"* 0 .

i '0, \ #\.: A 0-

0 4 0 0 200 3C.J 400 m T I M /mer)

FIG. 3. Normalized experimental and theoretical correlation function of light scattered at an angle of 4' with respect to the transmitted beam at 4°C in HCI, pH 2.0. Experimental curves for lathyritic chick skin (A) and rat tail tendon (0) are composed of data obtained at sample times between 2 and 10 ms. The true base-line was obtainable only at samples times of 5 and 10 ms. The theoretical normalized correlation function (0) was obtained using Equations 4 and 10.

0 0.1 0.2 0.3 0 4 05

CONCENTRATION f m g l m f FIG. 4. Translational difision coefficient back calculated to

standard conditions, Dz0.,, versus concentration for high mo- lecular weight (0) and single molecule fractions (0) of pH 2.0 acid-solubilized rat tail tendon collagen at 4°C. All measure- ments were made at 4' on solutions with Kayleigh factors in agree- ment with those reported in the legend to Fig. 3 using a Chromatix KMX-6 light scattering device modified for photon correlation.

'.O0 r

0 5 40 45 20 25 30 35 40 45

T€MP€RA TURE f "C/

FIG. 5. Molecular weight versus temperature for high molec- ular weight (A) and single molecule fractions (0) of pH 2.0 solubilized rat tail tendon collagen. After denaturation, the high molecular weight and single molecule fractions have molecular weights of 340,000 f 30,000 and 130,000 f 15.000, respectively.

FIG. 6. Transmission electron micrographs. A, negatively stained end overlapped (1) and end-to-end aggregated (2) SLS hased on the end-to-end distance of a single collagen molecule, denoted by the hnr. B, positively stained end overlapped aggregate; overlap zone marked by arrows. C, positively stained end-to-end aggregated SLS (COOH terminus to COOH terminus); arrows, Bands 51 and 52 on adjacent molecules. All electron micrographs shown are a t a magni- fication of x 155.000.

temperature of both preparations was 33.5 f 0.5OC. In addi- tion, this figure shows that the denatured molecular weights are 3.40 f. 0.30 X lo5 and 1.30 f 0.100 X 10" for the high molecular and low molecular weight fractions.

Ultrastructural Properties-Transmission electron micro- graphs of SLS formed from the high molecular weight fraction are shown in Fig. 6, A, B, and C. This preparation contains linearly aggregated SLS with both 9% overlaps and end-to-

9430 Type I Collagen in Solution

end arrangements of adjacent molecules. Positively stained preparations indicate that most, if not all, of the overlapped SLS have common polarities of the overlapping molecules, whereas the end-to-end aggregates were predominately anti- pardel with the associating ends being the COOH termini.

Acid-soluble versus Lathyritic Collagens-Weight average molecular weight and average translational diffusion coeffi- cient of neutral extracted lathyritic chick skin collagen in 0.01 M HCI are shown in Table 1. The lathyritic solutions were filtered through a 0.45-p fiter prior to examination but, as noted earlier, were subjected to a different purification process from that shown in Fig. 1.

DISCUSSION

In an attempt to determine the staggering mode(s) present at the macromolecular levels in the type 1 fibril, we have solubilized fibrils into constituent soluble forms and examined their physical properties in solution. The acid-extracted rat tendon collagen preparations were not salt-precipitated in an effort to preclude aggregation induced by the precipitations used in typical protocols. Because of the purity of the tendon as a tissue, the acid-soluble material obtained from it revealed typical type I collagen properties by SDS-gel electrophoresis and amino acid analysis. Laser light scattering studies dem- onstrate that the weight average molecular weight, Br, of the high molecular weight fraction was 8.05 -t 0.40 X lo5 at 4°C. Using 2.82 X lo5 for the light scattering molecular weight of the single molecule preparation, we calculate that the average aggregate contains 2.85 molecules. From this nonintegrd value of the molecular weight and the sample time dependence of the correlation function, it is apparent that this preparation contains at least two components. From experimental and theoretical values of M ? and Dm+,, models were generated of the several possible aggregate structures present in solution. The first models generated were for single component solu- tions of 0, 1, 2, 3, and 4 D staggered collagen molecules and are listed in Table 2. The value of the diffusion coefficient of single molecules at zero concentration was found to be 0.780 +- 0.04 X 10-7 cm2/s, which is close to the theoretical value of 0.858 X cm2/s for a 300 nm long, 1.5 nm wide rod calculated using Equation 4. Previously, Fletcher (5) found an anomalous dependence of the diffusion coefficient on concen- tration at acid pH with a maximum value of 0.86 f 0.2 X cm2/s near a concentration of 0.4 mg/ml. Our data also appear to maximize near that concentration. We have used the the- oretical diffusion coefficient 0.858 X 10-7 cm2/s for generating staggering models which falls between Fletcher’s (5) maxi- mum experimental value and our experimental value at zero concentration.

In order for the diffusion coefficient to be 0.450 X lo-? cm’/ s and the weight average number of molecules in the aggregate to be 2.85, the aggregate cannot be 0 D staggered since the diffusion coeffkient of all 0 D staggered models up to a hexamer is higher than the observed value. On the other hand, using Table 1, it is apparent that a two-component mixture of monomers and tetramers, monomers and pentamers, and di- mers and trimers with 1 to 4 D staggers can quantitatively explain these experimental observations. The best two-com- ponent models can be obtained by finding the staggering pattern which results in the experimentally observed values of weight average number of molecules and average diffusion coefficient, As model screening criteria, we used 0.45 -+ 0.06 X IO-’ cm2/s for the average diffusion coefficient and 2.85 f 0.35 for the weight average number of molecules. Tables 3 to 6 illustrate that mixtures of monomers and either tetramers or pentamers and mixtures of dimers and either trimers, tetramers, or pentamers with 1, 2, 3, or 4 D staggering are

models. From these calculations, it is clear that other infor- mation is necessary to define the exact structure of collagen aggregates in acid solutions. In order to determine which of the staggering modes prevail in collagen solutions, we pre- pared SLS of solubilized collagen from the 0.45-p filtered preparation. Electron micrographs of these solutions revealed many monomeric, dimeric, trimeric, tetrameric, and even pen- tameric SLS. Careful analysis of negatively stained prepara- tions demonstrated the presence of two types of SLS.

The first type (Fig. 6A) consisted of crystallites with over- laps of about 9% of the length of a monomeric SLS; the second type were end-to-end or 4.4 D staggered SLS. In fact, 4.0 and 4.4 D staggered SLS interactions accounted for 92% of all interactions observed in negatively stained specimens. Posi- tively stained samples were analyzed by determining the fraction of the molecular length between bands 3 and 4 and 51 and 52 to the NH2- and COOH-tenninal ends (see “Materials and Methods”). As shown in Fig. 6B, 4.0 D polarized staggered aggregates at least 2 molecules long as well as 4.4 D staggered antiparallel aggregates are observed (Fig. 6 0 .

It is apparent from these SLS that the most common staggering patterns observed are 4.0 and 4.4 D alignment of neighboring collagen molecules. It is not clear whether 4.4 D interactions are induced during SLS formation. Other reports of head to tail or end overlapped SLS have been made (24, 25) supporting these observations. lt is important to note that 4.0 and 4.4 D staggered aggregates would have about the same diffusion coefficients and therefore either model or a mixture of these aggregates would fit the light scattering data. As noted earlier, most of the 4.4 D staggered preparations con- tained an antiparallel arrangement of the molecules. Although antiparallel components can be present in fibrillar forms, it is not known whether such components can be present in typical native fibrils without disrupting fibril structure. Of interest is the observation that antiparallel SLS with 0 D overlap have been described in odontoblast-derived tissues as well as near 4.0 to 4.4 D dimeric SLS in chick embryo tendon cells whose polarities have not been determined (13, 14).

Refinement of the 4 D staggered model can be made by including additional components in the model and by fitting the theoretical correlation function to experimental measure- ments. Multicomponent 4 D models (Table 7) fit the diffusion and molecular weight data better but, in many cases, do not significantly change the fit of the correlation function. We have chosen the model which has weighting factors of 0.2,0.3, and 0.50 for monomers, dimers, and trimers to illustrate how well these models fit the correlation function. Fig. 3 compares the experimental and theoretical normalized correlation func- tions for a typical experiment. We conclude from these results that solubilized fibril fragments are predominantly monomers, linear dimers, and trimers, the weight fraction of each of which is probably dependent on collagen concentration and other solution conditions. An estimate of the weight fraction of each species can be made from the weighting factors (see Equations 6 and 7), which we have calculated to be 0.387 monomers, 0.291 dimers, and 0.321 trimers. Multiangle studies on these solutions are needed to determine more accurately the weight fraction of each component.

To determine whether the process of linear aggregation thermally stabilizes collagen in acid solution, denaturation studies were conducted on rat tail tendon solutions of single molecules (I& = 0.282 X lo6) and linear aggregates (nr = 0.805 X lo6). The process of denaturation was followed by measuring the apparent molecular weight (Re/&, see “Ma- terials and Methods”) as a function of temperature. Both solutions behaved similarly in that the molecular weight dropped significantly at a temperature of 33.5 f 0.5”C, which

Type I Collagen in Solution 9431

DISSOLUTION

\ n+

GELATION ' HE AT

\-- " t" - - - - .- FIG. 7. Schematic representation of dissolution and gelation

of type I collagen from a collagen fibril. The relationship between monomers and 4 D staggered dimers, trimers, and tetramers in the standard two-dimensional presentation of the collagen fibril is high- lighted. The 0.4 D overlap or 4 D stagger of molecules reflects a particularly stable form of aggregation in solution for both lathyritic and nonlathyritic collagens. Dissolution of the fibril renders such aggregates soluble as indicated by the fragments in the lower portion. Gelation of such solutions apparently involves the same 4 D staggered intermediates (16).

is close to the melting point of collagen at pH 2.0 determined by viscometry (26, 27) and optical methods (26). After dena- turation, the single molecule solution had a weight average molecular weight of 1.30 X lo5, suggesting that the weight fraction of a chains must be at least 0.6, assuming a molecular weight of 0.95 X lo5 and 1.80 X lo" for 01 chains and ,8 components, respectively. In comparison, the molecular weight after denaturation of the high molecular weight frac- tion containing linear aggregates was found to be 3.4 x lo5, which is greater than the molecular weight of a y component and therefore must reflect the fact that at least some of the linear aggregates are covalently cross-linked together. This fact is significant since, in the absence of cross-linking, it could be argued that linear aggregates are a quasi-stable solution state and do not accurately represent fibril fragments.

In addition to the experiments on acid-soluble rat tail tendon collagen, we have studied a purified neutral extract of lathyritic chick skin collagen in 0.01 M HCl. This preparation was salt-precipitated several times so that induced aggregation during purification was a distinct possibility. Table 1 and Fig. 3 compare the weight average molecular weight, average diffusion coefficient, and the experimental normalized corre- lation function to the values of these parameters obtained for the acid-soluble rat tail tendon solution at 4°C. These data indicate that the lathyritic skin collagen solution and the rat tail collagen in acid appear to be quite similar and that purification or the lathyritic state does not influence the state of aggregation. These results suggest that the type I native fibril from at least two different vertebrate tissues probably contains 4 D staggered filaments at least 3 or 4 molecules in length which appear to be identical with filaments which form during the initial stages of collagen heat gelation (16). The specificity for the 4 D staggered interaction probably involves the imino acid-poor regions located about 0.4 D from each end of the molecule which from stereochemical considerations may be more flexible than the neighboring parts of the colla- gen helix. The importance of the nonhelical ends on fibril formation in vitro (28-30) probably reflects the necessity for the proper alignment of the telopeptides and imino acid-poor regions (30, 31).

The presence of aggregates in acid solutions agrees with

previous observations by Obrink (3) that the molecular weight by light scattering of citrate solutions (pH 3.7) of collagen from lathyritic and nonlathyritic rat skin are 7.40 X lo5 and 8.10 to 18.0 X lo5, respectively, and by Fletcher (5) of the two- component nature of acid solutions.

In summary, our findings suggest that type I fibrils are at least partially composed of 4 D staggered filaments as dia- grammed in Fig. 7 and that the 4 D stagger is probably more stable than other types of interactions in acid or neutral solutions at 4°C. The ability of collagen to reassemble into native fibrils probably derives from sequence information contained in the nonhelical ends and in the imino acid-poor regions about 0.4 D from each end. How or if this information facilitates fibril assembly in uzuo is uncertain; however, in uiuo morphological observations that tendon fibroblasts and cor- neal epithelial cells (12,13,15) contain intracellular aggregates 2 and 3 molecules in length may reflect this preference.

REFERENCES 1. Boedtker, H., and Doty, P. (1956) J. Am. Chem. SOC. 78, 4267-

2. Davison, P. F., and Drake, M. P. (1966) Biochemistry 5,313-321 3. Obrink, B. (1972) Eur. J. Biochem. 25,563-572 4. Yuan, L., and Veis, A. (1973) Biopolymers 12, 1437-1444 5. Fletcher, G. C. (1976) Biopolymers 15, 2201-2217 6. Thomas, J. C., and Fletcher, G. C. (1979) Biopolymers 18, 1333-

7. Smith, J. W. (1968) Nature (Lond.) 219, 157-158 8. Veis, A., Anesey, J., and Mussell, S. (1967) Nature (Lond.) 215,

9. Hosemann, R., Dressig, W., and Nemetschek, TH. (1974) J. Mol. Biol. 83,275-280

10. Piez, K. A. (1975) in Extracellular Matrix Influences on Gene Expression (Slavkin, H. C., and Greulich, R. C., eds) pp. 231- 236, Academic Press, New York

11. Hukins, D. W. L., and Woodhead-Galloway, J. (1977) Mol. Cryst.

4280

1352

931-934

Liq. Cryst. 41, 33-39 12. Trelstad. R. L. (1971) J. Cell Biol. 48.689-694 13. 14. 15.

16.

17. 18.

19. 20. 21. 22.

23.

24.

25. 26.

27.

28. 29.

30.

31.

Trelstad; R. L., and Hayashi, K. (1979) Deu. Biol. 71, 228-242 Weinstock, M. (1977) J. Ultrastruct. Res. 61, 218-229 Bruns, R. R., Hulmes, D. J. S., Therrieu, S. F., and Gross, J.

(1979) Proc. Natl. Acad. Sci. U. S. A . 76, 313-317 Silver, F. H., Langley, K. H., and Trelstad, R. L. (1979) Biopoly-

mers 18, 2523-2535 Bnms, R. R., and Gross, J. (1973) Biochemistry 12,808-815 Tanford, C. (1961) Physical Chemistry of Macromolecules, Chap.

Schwartz, D., and Veis, A. (1978) Connect. Tissue Res. 6,185-190 Doty, P., and Edsall, J. T. (1951) Adu. Protein Chem.6, 50-55 Ford, N. C., Jr. (1972) Chem. Scr. 2,193-206 Berne, B. J., and Pecora, R. (1976) Dynamic Light Scattering,

Chap. 8, John Wiley & Sons, New York Cohen, R. J., and Benedek, G. B. (1976) J. Mol. Biol. 108, 151-

178 Rauterberg, J., von Bassewitz, D. B. (1975) Hoppe-Seyler's 2.

Physiol. Chem. 356,95-100 Veis, A., and Yuan, L. (1974) Biopolymers 14,895-900 Hayashi, T., and Nagai, Y. (1973) J. Biochem (Tokyo) 73, 999-

Dick, Y. P., and Nordwig, A. (1966) Arch. Biochem. Biophys. 117,

Comper, W. D., and Veis, A. (1977) Biopolymers 16,2133-2142 Hayashi, T., and Nagai, Y. (1973) J. Biochem (Tokyo) 74, 253-

Helseth, D. L., Jr., Lechner, J. H., and Veis, A. (1979) Biopolymers

Silver, F. H., and Trelstad, R. L. (1979) J. Theor. Biol. 81, 515-

5 and 6, John Wiley & Sons, New York

1006

466-468

262

18,3005-3014

526

9432 Type I Collagen in Solution

By. F . H . S~lvel and R.L. Tlrelrtad

EXPERIMENTAL AND THEORETICAL METHODS

Laser Llght Scattering

a , MoleCUlaT Welght Determ~nat~on

where

component

OD O D OD on I n on 10 1D 10 2D 2D 2D 2D

3n

30 40

30

3n

4 0 40

4n

0 . 8 5 8 0.718 0.758

0.717 0.717

0.717 0.642

0.471 0 . 5 5 9

0.425

0.446 0.559

0 . 3 1 6

0.555 0.308

0.337 0.415

0.281 0.498

0.281 0.357

0.233

1 2 3 4 5 6 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5

Calculated u r l n g equarlon 4 ~ bee Methods .. n IS the number Of molecules 1" fhe aggregate

""" D = b7 nm 1s used for all calculations

Table 3

0 . 2 0.4 0 . 6 0.8

,I . 2 0.4 0.6 0.8

r r lners

0.8 0.6 0.4 0.2

0 . 2 0.4 0 . 6 0 . 8

0.8 0.0 0.4 0.2

0 . 8 0 . 6 0.4 0.2

2a = 300. Zb = l.5nm 2a = 300. 2b - 3.0nn 2a = 300. 2h - 3.0nm 2a - 300, Zb = 4 Onm La - 300, 2h = 4.Onm

2a ~ 368, Zh = 3.0nm 2a - 300, 2b = 4.Onm

2a - 4 3 6 , 2h = 3.Onrn 2 1 - 504, 2h - 4.0nm l a = 972, 2b - 4.0nm 2a = 436, 2h = : . O m 2a = 572, 2b = 3.0nm Za = 708. 2b = 4.0nm

2a - 504, ?h - 1.5nm 2a - 844, 2b = 4.Onm

Za = 9 1 2 , 2b = 1 5nm 2a = 708, 2h = l.5nm

2a = 1116, 2h = I S n m La = 572. 2h ~ 1.Sna

Za = 1116, 2h = 1.1nm 2a = 844. 2h = 1.5nm

2a - 1388, 2h = 1.5nm

1.667 1.439

1.111 1 . 2 5 0

2.14 1.667

1.154 1.364

2 . 5 1.818

1.240 1.750

1.923 2 . 7 8

1.470 1.279 2.14 2 . 3 1 2.50 2.73

x c d / s e c

0.685 0.728

0.815 0.772

0.b19 0.079

0 . 7 9 8 11.548 0. 6 2 6 0 . 7 0 3

0.734

0. 7 8 0

u . s v n 0 . ~ 1 2

0.685 0.771 0.625 0 . 0 0 9 0 . 1 9 2 0.576

0.8 3 33 0 . 5 0 1 O.b 2.86 0.539 0.4 2.50 0.574 0.2 2 2 2 0.608

0 . 8 3.85 a 408

0 . 2 2 . 2 7 0.599

O.b 3.13 0 . 1 1 2 0.4 2 b3 0 5 5 5

Type I Collagen in Solution 9433 T a b l e 6 T a b l e 4

Trlrner

0 . 8 0 . 6 0 4 0 . 2

0 . 8 O b 0 . 4 0 . 2

MO"Olller

0 . 2 0 . 4 0 . 6 0 . 8 0 . 2 0 . 4 0 . 6 0 . 8

0 . 2 0 . 4 0 . 6 0 . 8

0 . 2

O h 0 8

0 4

Models

nlmer

0 . 8 0 . 6 0 . 4 0 . 2

0 . 2 0 . 4 0 . 6 0 . 8

0 . 2 0 . 4

0 . 8 0 . 6

0 . 2 0 . 4 0 . 6 0 . 8

% O , w crn2/se,

0 . 3 8 5 0 . 4 1 3 0 . 4 4 2 0 470 0 . 3 2 4 0 . 3 6 8 0 . 4 1 1 0 . 4 5 5 0 . 2 8 5 0 . 3 3 9 0 . 3 9 2 0 . 4 4 5 0 . 5 7 0 0 . 6 4 2

0 . 7 8 0 0 . 7 5 8 0 . 6 5 8 0 5 5 7

0 . 3 9 6 0 . 4 5 7

0 . 5 1 2 0 . h ? 7

0 . 3 5 8 0 . 4 8 3 0 . 6 0 8 0 . 7 3 3

0.714

0 . 7 4 3

1 . 4 2 9 1 .667

1 . 2 5 0 1.111 2 . 1 4 1 . 6 6 7 1 . 3 6 4 1 . 1 5 4

2 . 5 0 1 . 8 1 8 1 . 4 2 9 1 . 1 7 6

2 . 7 8 1 . 9 2 3

1 . 2 7 9 1 . 4 7 0

2 . 7 3 2 . 5 0

2 .14 2 . 3 1

3 . 3 3 2 . 8 6 2 . 5 0 2 . 2 2

3 . 8 5 3 .13 2 . 6 3 2 .27

1 . 6 6 7 1 . 4 2 9

1 . 1 1 1 1 . 2 5 0

1 . 6 6 7 2 . 1 4

1 . 3 6 4 1.154

1 . 8 1 8 2 . 5 0

1 . 4 2 9 1 . 1 7 6

1 . 9 2 3 2 . 7 8

1 . 4 7 0 1 . 2 7 9 2 . 7 3 2 . 5 0 2 . 3 1 2 . 1 4 3 .33 2 . 8 6 2 . 5 0 2 . 2 2 3 . 8 5 3 . 1 3 2 . 6 3 2 . 2 7

Tetramer Tetramer Pentamer

2 . 7 3 2 . 5 0 2 . 3 1

3 . 3 3 2 . 1 4

2 . 8 6 2 . 50 2 . 2 2

0 . 8 3 . 8 5 0 . 6 3 . 1 3 0 . 4 2 . 6 3 0 . 2 2 27

1 , 6 6 7 1 . 4 2 9 1 . 2 5 0 1.111 1 . 1 5 5 1 . 3 6 4 1 . 6 6 7 2 . 1 4 4 2 . 5 0 0 1 . 8 1 8 1 . 4 2 9 1 . 1 7 6

0 . 8 2 . 7 8 0 . 6 1 . 9 2 3 0 . 4 1 . 4 7 0 0 . 2 1 279

0 . 2 0 . 8 0 . 4 0 . 6 0 . 6 0 . 4 0 . 8 0 . 2 0 . 2 0 . 4 0 . 6 0 . 8 0 . 2 0 . 4

0 . 6 1 9 0 . 6 7 9 0 . 7 2 3 0 . 7 9 8 0 , 5 2 8 0 . 6 1 1 0 . 6 9 3 0 . 7 1 6

0 . 8 0 . 6 0 . 4 0 . 2

0 . 8 0 . 6 0 . 4 0 . 2

0 . 8 0 . 6 0 . 4 0 . 2

0 . 4 5 6 0 . 5 5 7 0 . 657 0 . 7 5 8

0 . 8 0 . 6 0 . 4 0 . 2

0 . 4 0 . 6 0 . 8 0 . 2 0 . 4

0 . 3 9 7 0 . 4 3 7 0 . 4 7 8 0 . 5 1 8

0 . 3 5 8 0 . 4 0 8 0 . 4 5 9 0 . 5 0 9

t i20 Y

C d l S e C

0 . 8

0 . 4 0 . 6

0 . 2

0 . 8 0 . 6 0 . 4 0 . 2

0 . 6 0 . 8

Table 7

4D Staggered F l b r l l Fragmenf Models

Tahle I Weighrlng F a c t o r s Olner Trlmer

0 . 5 0 . 5

0 4

0 . 3 0 . 5

0 . 4 0 . 5

0 . 3

0 . 3 0 . 2 5

0 . 2 0.25

0 . 4 0 . 2 5

retramer D~~ x IO'? .sl * c ( ~ A ~ ) C.j/SeC a, I

0 . 4 2 8 2 . 5 1 . ~ 2 0

0 5 0 . 4 Z 7 2 . 9 0 . 9 4 6

0 . 4 7 5 2 . 8 1.1180

0 4b5 2 . 4 1 064

0 . 5 0 . 5 0 2

0 25

2 3 1 . 2 1 0

U.489 2 . 5 5 1 . 1 0 0

0 . 2 5 O . 5 l b 2 . 4 5 1 . 3 5 0

0 . 2 5 0 . 4 4 6 2 , b S 1 . 0 9 0

Weighting F a c t o r s 3~ D S t a g g e r

rTlmer

0 . 8 0 . 0 0 . 4 0 . 2

0 . 8 0 . 6 0 . 4 0 . 2

0.1

0 . 2

0 . 1

0 . 2

0 . 2

0 . 3

0 . 1

0 . 2 0 . 4 0 0

0 . 2 0 8

<I . 6 !I . 4

0.:

(I . h U.8 (I. 2 0 4 !J h

u . n

0 . 4

0 . 2

11 . 8 O.b 0 . 4 0 . 2

0 2 0 . 4 D , b 0 . 8

0 . 4 0.6 0 8 0 2

0 . 0

0.2

0 . 4

0 . n

0 . 8 0 . 6 0 . 4 0 . 2

0 . 8

0 . 4 0 . 6

0 . 8