/. Biochem. 84, 245-249 (1978)
Time-Dependent Increase in the Stability of Collagen Fibrils Formed In Vitro I.
Effects of pH and Salt Concentration on the Dissolution of the Fibrils 1 Toshihiko HAYASHI Laboratory of Chemistry, Jichi Medical School, Minamikawachimachi, Tochigi 329-04, and Department of Tissue Physiology, Medical Research Institute, Tokyo Medical and Dental University, Kandasurugadai, Chiyoda-ku, Tokyo 101 Received for publication, September 17, 1977
The time-dependent increase in stability, as measured in terms of the rate of dissolution, of collagen fibrils formed in vitro from pepsin-treated collagen was significantly affected only by temperature, and not by either ionic strength or pH. This is in contrast with collagen fibril formation, a process which is greatly affected by ionic strength and pH. Within the range of temperature 29-37°C, lower temperature caused slower fibril formation and faster fibril stabilization. These results suggest that the intermolecular interactions involved in stabilizing collagen fibrils are entirely different from those involved in fibril formation. Based on kinetic analysis of the dissolution and stabilization of the fibrils, it is proposed that collagen molecules first form unstable fibrils which become gradually stabilized on prolonged incubation, without necessarily introducing covalent cross-links.
Collagen fibrils formed in vitro by warming solutions of either pepsin-treated or acid soluble collagen will redissolve upon cooling. In the previous study (7) it was reported that these fibrils become less soluble on cooling with increasing time of incubation at the temperature used for fibril formation. In that study it was concluded that not only covalent cross-links but also non-covalent types of interactions between collagen molecules were involved in the time-dependent increase in stability. 1
This study was supported in part by a Scientific Research Grant from the Ministry of Education, Science and Culture of Japan (No. 837008). Vol. 84, No. 2, 1978
245
These conclusions were based on three principal grounds. First, the covalent cross-links in collagen so far reported (2) are located in the nonhelical regions of the molecules which are split off on treatment with a protease such as pepsin. Thus, fibrils formed in vitro from pepsin-treated collagen contain no pre-existing covalent cross-links and it is unlikely that such cross-links are formed during the incubation. However, these collagen fibrils did increase in stability, as measured in terms of the rate of redissolution, when the incubation time was prolonged. Second, although the fibrils formed from pepsin-treated collagen gradually stabilized as mentioned above, they always redissolved completely at 0°C, even after prolonged
T. HAYASHI
246
incubation, e.g. for several weeks, at the fibrilforming temperature. That is, stabilization of collagen fibrils as measured in terms of the reduced rate of redissolution is not accompanied by irreversible insolubilization as it is in the case of stabilization by covalent cross-links. Third, if only a cross-linking reaction is responsible for stabilization of the collagen fibrils, the fibrils should be stabilized faster at 37°C than at 29°C, since collagen fibril formation is completed much earlier at 37°C than at 29°C. However, the experimental results (1) showed that fibrils were stabilized faster at 29°C than at 37°C. The previous study (1) thus suggested that it was necessary to take into account non-covalent interactions between the molecules to explain the stabilization of collagen fibrils as measured in terms of the rate of redissolution. Non-covalent interactions include, for instance, electrostatic interactions, hydrogen bonds and hydrophobic bonds. In order to investigate the role of electrostatic interactions in the stabilization of collagen fibrils, the effects of pH and ionic strength were studied. It was found that collagen fibril stabilization essentially did not depend on either pH or ionic strength. MATERIALS AND METHODS Pepsin-treated collagen from calf skin, prepared as reported previously (5), was used throughout these experiments. A collagen solution (final concentration, 2 mg/ml) was prepared by mixing a chilled buffer of appropriate pH and concentration with a chilled solution of collagen (4 mg/ml) in 0.05% acetic acid. Collagen fibril formation was followed by measuring the turbidity of the solution at 430 nm using a Shimadzu model 20 spectrophotometer. Redissolution of collagen fibrils was determined by measuring the decrease in turbidity at 430 nm upon transferring .the tubes containing the fibrils to a lower temperature. RESULTS AND DISCUSSION Collagen fibrils formed by incubation at 37°C for 6 h began to disappear upon transferring them to a lower temperature (Fig. 1), At 0°C, 5°C, or 10°C, collagen fibrils disappeared completely in a few minutes and gave a clear solution. The higher .the cooling temperature, the more slowly the fibrils
08 33>37>41°C) as described previously (1). This is in contrast with the formation of collagen fibrils, a process which is
Vol. 84, No. 2, 1978
50 BO (hr) Fig. 3. Effect of temperature on the stabilization of collagen fibrils. Collagen fibrils were formed at 29°C (O), 33°C ( • ) . 37°C (A), or 41°C (A), for the period indicated on the abscissa, and then cooled to 20°C. Values of Fc/Ft were obtained as descnbed in the legend to Fig. 2.
0-, Q5 turbidity of Fc turbidity of R
10
Fig. 4. Correlation between an arbitrary measure of stabilization and the degree of stabilization, FcjFt, defined in this report. The arbitrary measure of stabilization is the ratio of the turbidity remaining 60 min after cooling to 20°C to the turbidity before cooling. promoted by elevated temperature up to the temperature of denaturation of the collagen molecules (4). In the previous study (/), the degree of stabilization of collagen fibrils was expressed as the ratio of the turbidity remaining 0.5 h after cooling to that before cooling. However, as shown in Fig. 4, there is no essential difference between the previous and present ways of expressing the degree of stabilization. The rate of collagen fibril formation is known to depend strongly on pH (4). The effect of pH on the stabilization of collagen fibrils was therefore
T. HAYASHI 1.0
-JO • • •
•
lion (! he/Ft
$
Jtion (Fci'Ft)
248
o
05
1
65
7.0
D 4
• • •
•
g 5 A
A ^
D
°
°2
D D
01
0
60
°
B
o o °
V)
6 A
§05
o o
8.-'
n
75
PH Fig. 5. Effect of pH on the stabilization of collagen fibrils. Collagen fibrils, formed by incubation either at 29°C ( • ) or at 37°C (O) for 6 h from solutions of various pH values in 0.3 M potassium phosphate, were cooled to 20°C. Values of Fc/Ft were obtained as described in the legend to Fig. 2.
examined (Fig. 5). Although the pH of the collagen solution greatly affected the stabilization of collagen fibrils formed at 37°C, little dependency of the stabilization of collagen fibrils on pH was observed at 29°C. This again is in contrast to the case of collagen fibril formation, which is more affected by pH at 29°C than at 37°C. These experimental results indicate that the main driving force involved in fibril formation in collagen is essentially different from that involved in the stabilization of the collagen fibrils. Experimental data concerning the degree of the stabilization below pH 6.7 at 29°C were not obtained because under these conditions collagen solutions formed no fibrils within 6 h. The contribution of electrostatic interactions to the stabilization of collagen fibrils can be tested by changing the ionic concentration. The effect of varying the NaCl concentration on the stabilization of collagen fibrils was studied (Fig. 6a). Redissolution of collagen fibrils formed at 37°C occurs more readily as the concentration of NaCl is increased. The results are consistent with those of Gross (5). However, redissolution of collagen fibrils formed at 29°C was not affected by changes in NaCl concentration at any pH tested. Since stabilization of collagen fibrils proceeds faster at 29°C than at 37°C (Ref. 1 and Fig. 3), these results indicate that ionic interactions are not primarily responsible for the stabilization of collagen fibrils. Formation of collagen fibrils is more dependent on phosphate concentration than on NaCl concen-
0
0] 02 03 QA 05 cone, of Nad added (M)
10-
a
5 05
•
"1
02 03 QA 05 06 cone of phosphate (M)
Fig. 6. Effect of salt concentration on the stabilization of collagen fibrils. Collagen fibrils were formed either at 29CC (filled symbols) or at 37°C (open symbols) for 6 h in various concentrations of NaCl and of phosphate buffer, and were cooled to 20°C. Values of Fc/Ft were obtained as described in the legend to Fig. 2. (a) Stabilization as a function of NaCl concentration in 0.05 M phosphate buffer at three pH's (circles: pH 6.2; triangles: pH6.8; squares: pH7.1). (b) Stabilization as a function of phosphate concentration at the same three pH values in the absence of NaCl.
tration (4). Thus, it is possible that even though the concentration of NaCl does not have a marked effect on fibril stabilization, that of phosphate might. However, the data in Fig. 6b show that the stabilization of collagen fibrils formed at 29°C is essentially unaffected by the concentration of phosphate, though that of fibrils formed at 37°C is affected. This again suggests that the molecular interactions involved in the stabilization of collagen fibrils at 29°C are not ionic in nature. Thus, all the results indicate that a type of force primarily controlled by temperature is the predominant factor influencing the stabilization of collagen fibrils and that ionic interactions are either absent or of secondary importance to the stabilization. /. Biochem.
EFFECTS OF pH AND IONS ON COLLAGEN FIBRIL STABILIZATION 10 05 02
0
100
50
(hr) Fig. 7. Change in Fb(=Ft—Fc) as a function of time. Collagen fibrils formed at 37°C for the time indicated on the abscissa from a solution containing 0.3 M potassium phosphate, pH 6.8, were cooled to 20°C. Values of Fc were obtained as described in the legend to Fig. 2. Stabilization of collagen fibrils can be considered to proceed in the following way: Monomers
> Unstable fibrils Fb • Stable fibrils (without cross-links) Fc
The model is based on the finding that the fibrils have at least two different stabilities as measured in terms of the rate of dissolution of the fibrils at 20°C (Fig. 2). The amount of stable fibrils increased with increasing time of incubation used for fibril formation (Fig. 3). If the rate of fibril formation is much faster than that of stabilization, then virtually all of the collagen molecules may be assumed to be in the unstable form upon completion of fibril formation. Given this, the following equation should hold in the later stages of the stabilization: dFcldt=
-dFbldt=kBFb.
Thus \nFb=\n(Ft—Fc)= where Ft=Fb+Fc=constant
—kBt+const. (3)
As shown in Fig. 7, Eq. 3 holds for collagen solution incubated at 37°C. This means that at 37°C the collagen fibril formation and stabilization
VoL 84, No. 2, 1978
249
proceed as shown in the scheme and that the rate constant k± is much greater than ke- However, at 29°C, where the fibril formation was slower and the fibril stabilization was faster than at 37°C, Eq. 3 is not valid because fibril stabilization proceeded to a significant degree before all the free molecules had been incorporated into fibrils. Thus, the kinetic constants k*. and ks are influenced in an opposite way by environmental conditions. A rise in temperature increases kA and decreases kB- On the other hand, ionic interactions, including those changed by pH, affect kA greatly but ks slightly, if at all. Quantitative values for kA and A:B at 29°C cannot be obtained at present because the mechanism of the reaction from monomer molecules to unstable fibrils has not been elucidated. A possible mechanism for the reaction has been proposed by the author (5), the details of which will be published elsewhere. The author gratefully acknowledges useful discussions with Prof. H. Ishikura (Jichi Medical School), Prof. Y. Nagai (Tokyo Medical and Dental University) and Prof. H. Noda (Tokyo University). He would also like to express appreciation to Dr. E.F. Eikenberry (Rutgers Medical School, NJ., U.S.A.) for correcting English of the manuscript. REFERENCES 1. Hayashi, T. & Nagai, Y. (1974) / . Biochem. 75, 651654 2. Tanzer, M.L. (1976) in Biochemistry of Collagen (Ramachandran, G.N. & Reddi, A.H., eds.) pp. 137-162, Plenum Press, New York and London 3. Hayashi, T. & Nagai, Y. (1972) / . Biochem. 72, 749758 4. Hayashi, T. & Nagai, Y. (1973) / . Biochem. 74, 253262 5. Gross, J. (1958) / . Exp. Med. 108, 215 6. Hayashi, T. (1977) in Keitaikeisei (in Japanese) (Go, N. & Wada, A., eds.) pp. 83-124, Univ. Tokyo Press, Tokyo
Time-Dependent Increase in the Stability of Collagen Fibrils Formed In Vitro I.
Effects of pH and Salt Concentration on the Dissolution of the Fibrils 1 Toshihiko HAYASHI Laboratory of Chemistry, Jichi Medical School, Minamikawachimachi, Tochigi 329-04, and Department of Tissue Physiology, Medical Research Institute, Tokyo Medical and Dental University, Kandasurugadai, Chiyoda-ku, Tokyo 101 Received for publication, September 17, 1977
The time-dependent increase in stability, as measured in terms of the rate of dissolution, of collagen fibrils formed in vitro from pepsin-treated collagen was significantly affected only by temperature, and not by either ionic strength or pH. This is in contrast with collagen fibril formation, a process which is greatly affected by ionic strength and pH. Within the range of temperature 29-37°C, lower temperature caused slower fibril formation and faster fibril stabilization. These results suggest that the intermolecular interactions involved in stabilizing collagen fibrils are entirely different from those involved in fibril formation. Based on kinetic analysis of the dissolution and stabilization of the fibrils, it is proposed that collagen molecules first form unstable fibrils which become gradually stabilized on prolonged incubation, without necessarily introducing covalent cross-links.
Collagen fibrils formed in vitro by warming solutions of either pepsin-treated or acid soluble collagen will redissolve upon cooling. In the previous study (7) it was reported that these fibrils become less soluble on cooling with increasing time of incubation at the temperature used for fibril formation. In that study it was concluded that not only covalent cross-links but also non-covalent types of interactions between collagen molecules were involved in the time-dependent increase in stability. 1
This study was supported in part by a Scientific Research Grant from the Ministry of Education, Science and Culture of Japan (No. 837008). Vol. 84, No. 2, 1978
245
These conclusions were based on three principal grounds. First, the covalent cross-links in collagen so far reported (2) are located in the nonhelical regions of the molecules which are split off on treatment with a protease such as pepsin. Thus, fibrils formed in vitro from pepsin-treated collagen contain no pre-existing covalent cross-links and it is unlikely that such cross-links are formed during the incubation. However, these collagen fibrils did increase in stability, as measured in terms of the rate of redissolution, when the incubation time was prolonged. Second, although the fibrils formed from pepsin-treated collagen gradually stabilized as mentioned above, they always redissolved completely at 0°C, even after prolonged
T. HAYASHI
246
incubation, e.g. for several weeks, at the fibrilforming temperature. That is, stabilization of collagen fibrils as measured in terms of the reduced rate of redissolution is not accompanied by irreversible insolubilization as it is in the case of stabilization by covalent cross-links. Third, if only a cross-linking reaction is responsible for stabilization of the collagen fibrils, the fibrils should be stabilized faster at 37°C than at 29°C, since collagen fibril formation is completed much earlier at 37°C than at 29°C. However, the experimental results (1) showed that fibrils were stabilized faster at 29°C than at 37°C. The previous study (1) thus suggested that it was necessary to take into account non-covalent interactions between the molecules to explain the stabilization of collagen fibrils as measured in terms of the rate of redissolution. Non-covalent interactions include, for instance, electrostatic interactions, hydrogen bonds and hydrophobic bonds. In order to investigate the role of electrostatic interactions in the stabilization of collagen fibrils, the effects of pH and ionic strength were studied. It was found that collagen fibril stabilization essentially did not depend on either pH or ionic strength. MATERIALS AND METHODS Pepsin-treated collagen from calf skin, prepared as reported previously (5), was used throughout these experiments. A collagen solution (final concentration, 2 mg/ml) was prepared by mixing a chilled buffer of appropriate pH and concentration with a chilled solution of collagen (4 mg/ml) in 0.05% acetic acid. Collagen fibril formation was followed by measuring the turbidity of the solution at 430 nm using a Shimadzu model 20 spectrophotometer. Redissolution of collagen fibrils was determined by measuring the decrease in turbidity at 430 nm upon transferring .the tubes containing the fibrils to a lower temperature. RESULTS AND DISCUSSION Collagen fibrils formed by incubation at 37°C for 6 h began to disappear upon transferring them to a lower temperature (Fig. 1), At 0°C, 5°C, or 10°C, collagen fibrils disappeared completely in a few minutes and gave a clear solution. The higher .the cooling temperature, the more slowly the fibrils
08 33>37>41°C) as described previously (1). This is in contrast with the formation of collagen fibrils, a process which is
Vol. 84, No. 2, 1978
50 BO (hr) Fig. 3. Effect of temperature on the stabilization of collagen fibrils. Collagen fibrils were formed at 29°C (O), 33°C ( • ) . 37°C (A), or 41°C (A), for the period indicated on the abscissa, and then cooled to 20°C. Values of Fc/Ft were obtained as descnbed in the legend to Fig. 2.
0-, Q5 turbidity of Fc turbidity of R
10
Fig. 4. Correlation between an arbitrary measure of stabilization and the degree of stabilization, FcjFt, defined in this report. The arbitrary measure of stabilization is the ratio of the turbidity remaining 60 min after cooling to 20°C to the turbidity before cooling. promoted by elevated temperature up to the temperature of denaturation of the collagen molecules (4). In the previous study (/), the degree of stabilization of collagen fibrils was expressed as the ratio of the turbidity remaining 0.5 h after cooling to that before cooling. However, as shown in Fig. 4, there is no essential difference between the previous and present ways of expressing the degree of stabilization. The rate of collagen fibril formation is known to depend strongly on pH (4). The effect of pH on the stabilization of collagen fibrils was therefore
T. HAYASHI 1.0
-JO • • •
•
lion (! he/Ft
$
Jtion (Fci'Ft)
248
o
05
1
65
7.0
D 4
• • •
•
g 5 A
A ^
D
°
°2
D D
01
0
60
°
B
o o °
V)
6 A
§05
o o
8.-'
n
75
PH Fig. 5. Effect of pH on the stabilization of collagen fibrils. Collagen fibrils, formed by incubation either at 29°C ( • ) or at 37°C (O) for 6 h from solutions of various pH values in 0.3 M potassium phosphate, were cooled to 20°C. Values of Fc/Ft were obtained as described in the legend to Fig. 2.
examined (Fig. 5). Although the pH of the collagen solution greatly affected the stabilization of collagen fibrils formed at 37°C, little dependency of the stabilization of collagen fibrils on pH was observed at 29°C. This again is in contrast to the case of collagen fibril formation, which is more affected by pH at 29°C than at 37°C. These experimental results indicate that the main driving force involved in fibril formation in collagen is essentially different from that involved in the stabilization of the collagen fibrils. Experimental data concerning the degree of the stabilization below pH 6.7 at 29°C were not obtained because under these conditions collagen solutions formed no fibrils within 6 h. The contribution of electrostatic interactions to the stabilization of collagen fibrils can be tested by changing the ionic concentration. The effect of varying the NaCl concentration on the stabilization of collagen fibrils was studied (Fig. 6a). Redissolution of collagen fibrils formed at 37°C occurs more readily as the concentration of NaCl is increased. The results are consistent with those of Gross (5). However, redissolution of collagen fibrils formed at 29°C was not affected by changes in NaCl concentration at any pH tested. Since stabilization of collagen fibrils proceeds faster at 29°C than at 37°C (Ref. 1 and Fig. 3), these results indicate that ionic interactions are not primarily responsible for the stabilization of collagen fibrils. Formation of collagen fibrils is more dependent on phosphate concentration than on NaCl concen-
0
0] 02 03 QA 05 cone, of Nad added (M)
10-
a
5 05
•
"1
02 03 QA 05 06 cone of phosphate (M)
Fig. 6. Effect of salt concentration on the stabilization of collagen fibrils. Collagen fibrils were formed either at 29CC (filled symbols) or at 37°C (open symbols) for 6 h in various concentrations of NaCl and of phosphate buffer, and were cooled to 20°C. Values of Fc/Ft were obtained as described in the legend to Fig. 2. (a) Stabilization as a function of NaCl concentration in 0.05 M phosphate buffer at three pH's (circles: pH 6.2; triangles: pH6.8; squares: pH7.1). (b) Stabilization as a function of phosphate concentration at the same three pH values in the absence of NaCl.
tration (4). Thus, it is possible that even though the concentration of NaCl does not have a marked effect on fibril stabilization, that of phosphate might. However, the data in Fig. 6b show that the stabilization of collagen fibrils formed at 29°C is essentially unaffected by the concentration of phosphate, though that of fibrils formed at 37°C is affected. This again suggests that the molecular interactions involved in the stabilization of collagen fibrils at 29°C are not ionic in nature. Thus, all the results indicate that a type of force primarily controlled by temperature is the predominant factor influencing the stabilization of collagen fibrils and that ionic interactions are either absent or of secondary importance to the stabilization. /. Biochem.
EFFECTS OF pH AND IONS ON COLLAGEN FIBRIL STABILIZATION 10 05 02
0
100
50
(hr) Fig. 7. Change in Fb(=Ft—Fc) as a function of time. Collagen fibrils formed at 37°C for the time indicated on the abscissa from a solution containing 0.3 M potassium phosphate, pH 6.8, were cooled to 20°C. Values of Fc were obtained as described in the legend to Fig. 2. Stabilization of collagen fibrils can be considered to proceed in the following way: Monomers
> Unstable fibrils Fb • Stable fibrils (without cross-links) Fc
The model is based on the finding that the fibrils have at least two different stabilities as measured in terms of the rate of dissolution of the fibrils at 20°C (Fig. 2). The amount of stable fibrils increased with increasing time of incubation used for fibril formation (Fig. 3). If the rate of fibril formation is much faster than that of stabilization, then virtually all of the collagen molecules may be assumed to be in the unstable form upon completion of fibril formation. Given this, the following equation should hold in the later stages of the stabilization: dFcldt=
-dFbldt=kBFb.
Thus \nFb=\n(Ft—Fc)= where Ft=Fb+Fc=constant
—kBt+const. (3)
As shown in Fig. 7, Eq. 3 holds for collagen solution incubated at 37°C. This means that at 37°C the collagen fibril formation and stabilization
VoL 84, No. 2, 1978
249
proceed as shown in the scheme and that the rate constant k± is much greater than ke- However, at 29°C, where the fibril formation was slower and the fibril stabilization was faster than at 37°C, Eq. 3 is not valid because fibril stabilization proceeded to a significant degree before all the free molecules had been incorporated into fibrils. Thus, the kinetic constants k*. and ks are influenced in an opposite way by environmental conditions. A rise in temperature increases kA and decreases kB- On the other hand, ionic interactions, including those changed by pH, affect kA greatly but ks slightly, if at all. Quantitative values for kA and A:B at 29°C cannot be obtained at present because the mechanism of the reaction from monomer molecules to unstable fibrils has not been elucidated. A possible mechanism for the reaction has been proposed by the author (5), the details of which will be published elsewhere. The author gratefully acknowledges useful discussions with Prof. H. Ishikura (Jichi Medical School), Prof. Y. Nagai (Tokyo Medical and Dental University) and Prof. H. Noda (Tokyo University). He would also like to express appreciation to Dr. E.F. Eikenberry (Rutgers Medical School, NJ., U.S.A.) for correcting English of the manuscript. REFERENCES 1. Hayashi, T. & Nagai, Y. (1974) / . Biochem. 75, 651654 2. Tanzer, M.L. (1976) in Biochemistry of Collagen (Ramachandran, G.N. & Reddi, A.H., eds.) pp. 137-162, Plenum Press, New York and London 3. Hayashi, T. & Nagai, Y. (1972) / . Biochem. 72, 749758 4. Hayashi, T. & Nagai, Y. (1973) / . Biochem. 74, 253262 5. Gross, J. (1958) / . Exp. Med. 108, 215 6. Hayashi, T. (1977) in Keitaikeisei (in Japanese) (Go, N. & Wada, A., eds.) pp. 83-124, Univ. Tokyo Press, Tokyo
