EXPERIMENTAL

CELL

RESEARCH

19 1,8-13

(1990)

Fibronectin and Collagen Gene Expression during in Vitro Ageing of Pig Skin Fibroblasts MICHELEMARTIN,*.~ *Laboratoire

de Radiobiologie tLaboratoire

RITAELNABOUT,PCHANTALLAFUMA,~FRANCOISECRECHET,*ANDJACQUESREMY* Appliqub-Commissariat de Biochimie du Tissu

ci 1’Energie Atomique, IPSN-DPS, Conjonctif, UA CNRS 1174, Faculte’

INTRODUCTION The degradation of the connective tissue during ageing results in pathologies in almost all types of organs. In the skin, dermis ageing is directly related to altered interactions between the fibroblast and the extracellular matrix [l, 21. Under tissue culture conditions, normal diploid fibroblasts have a finite division potential and exhibit morphological and metabolic alterations at the end of their lifespan. Therefore, fibroblast in vitro ageing has been extensively studied as a model of in viva ageing [3,4,5]. Cells undergo numerous changes during in vitro ageing, which implies alterations in the expression of numerous genes. During the lifespan of human diploid fibroblasts, their total RNA and mRNA content reprint

requests

should

0014.4827/90 $3.00 Copyright 0 1990 hy Academic Press, All rights of reproduction in any form

MATERIALS

and

AND METHODS

Cell Culture Skin biopsies were removed from 10 Large White pigs. They comprised females and castrated males 6 months old, and weighed 50 kg. The epidermis and hypodermis were removed from the biopsy. The tissue was minced finely and then incubated serially at 37°C with 0.25% collagenase (Boehringer) and 0.5% trypsin (Difco, 1:250) in calcium-free phosphate-buffered saline (PBS). Fetal calf serum (FCS) was added to stop all enzyme activity and the cells were either frozen or immediately plated into culture flasks (Nunc, 25 cm’) at a

be addressed. 8

Inc. reserved.

France;

has been reported by most authors to exhibit a gradual increase followed by a sharp increase, as cells reached the senescent phase of their lifespan [6, 7,8]. It is, however, not known if the accumulation of messengerRNAs accounts for all types of transcripts or if it is specific to certain classes. Particularly, very little is known about the alterations of the genes coding for the extracellular matrix. The present study was therefore designed to evaluate the rate of expression of three major matrix genes during the in vitro ageing of fibroblasts. We began by studying the mRNA contents using cDNA-specific probes for the collagen type I and III and for the fibronectin. We then attempted to establish whether changes in fibroblastic mRNA resulted in changes in proteins. There are many biochemical studies of changes occurring in protein synthesis during cell ageing in the literature, but their results are very often contradictory. Most authors reported that the synthesis of total proteins increases during in vitro ageing [g-13]. The synthesis of collagen has been claimed by some to decrease during ageing [14-161 and by others to increase [17], and fibronectin synthesis has also been controversial [18, 191. In the present work, we found that gene expression in mRNA altered differently, depending on the genes studied. Thus the amount of transcripts clearly increased for type III collagen and fibronectin, but decreased for type I collagen. As regards proteins, we found that alterations in the amounts secreted were similar to those observed for mRNA. Inside the cells, complex processes of protein accumulation and degradation were observed during the senescent phase of growth.

The fihronectin, collagen type I, and collagen type III genes code for three major proteins of the cell matrix. The age-related alterations in their expression were measured during the in vitro lifespan of pig skin fibroblasts. We observed changes in the transcription rate of these specific genes during ageing. The levels of fibronectin and type III collagen mRNA rose markedly during the senescence phase. The level of collagen type I mRNA decreased during cell ageing, while that of @-actin did not change. As regards proteins, we observed a sharp increase in the secreted noncollagenous proteins and in the total proteins of the cell layer during senescence. On the contrary, the secretion of the collagenous proteins decreased during senescence. Moreover, most of the newly synthesized molecules of collagen were immediately degraded in the cells, before their extracellular secretion. The terminal phenotype of pig senescent cells was therefore characterized by overexpression of fibronectin and type III collagen genes and reduced expression of the type I collagen gene. Surprisingly, for fibronectin and type III collagen, that terminal phenotype resembled the one normally found in the fibroblasts during the processes of tissue repair, cicatrization, and development. o isgo Academic PWS, IIIC.

’ To whom

Centre de Saclay, 91191 Gif Sur Yvette, de M&decine, 94010 Creteil, France

FIBRONECTIN

AND

COLLAGEN

density of 105/cm2. All cells were grown in DMEM supplemented with 10% FCS, 0.4% antibiotics, 2% glutamine and 10 mM Hepes. Cells were serially passaged until senescence. When the cultures reached confluency, cells were dissociated with a 0.5% trypsin solution in PBS and counted in a hemocytometer. A constant inoculation density of lo6 cells per 25 cm2 flask was maintained throughout their lifespans. We checked that each cell strain was free of mycoplasma contamination. The replicative lifespan of the cells was determined using the formula: PD = Log(N/N” *Att Eff)/Log 2, where PD = population doublings, N the cell density per confluent flask, N” the density of the inoculated cells, and Att Eff = attachment efficiency, i.e., the fraction of inoculated cells that remained attached to the culture flask 16 hr after seeding. The lifespans of the cell lines were considered to be terminated when they failed to reach confluency after a period of 4 weeks, with the usual three changes of medium per week. For RNA and proteins studies, we used only the senescent cells still able to reach confluency. Culture density was then very low, as the culture was mainly composed of giant senescent cells. RNA

Analysis

RNA isolation. Skin fibrobl as t s were cultured as described above. The culture medium was supplemented with 50 pg/ml of ascorbic acid 24 hr before the end of the culture. Total ribonucleic acid (RNA) was extracted from confluent cultures by 5 M guanidinium isothiocyanate solubilization and ultracentrifugation of the extract through a 5.7 M cesium chloride density gradient, according to Chirwing et al. [20]. For qualitative assessment, a fraction of total RNA was enriched in poly(A)+ RNA on an oligo(dT) cellulose column [21] and was submitted to Northern transfer analysis according to Maniatis et al. [22]: briefly, after denaturation for 1 hr at 50°C with glyoxal and DMSO, 1 Kg of the poly(A)+ RNA fraction underwent electrophoresis on 0.8% agarose gel in 10 mMphosphate buffer, pH 6.7. The RNAs thus separated were transferred by blotting onto a Nytran membrane (Schleicher and Schuell). For quantitative assessment, aliquots of 0.03 to 1 fig of total RNA were denatured and dotted onto Nytran membrane, using a commercial Bio-Dot microfiltration apparatus (Bio-Rad). Pig lymphocyte RNA was prepared in the same way and used as a negative control. DNA probes. The following cDNA probes were used: pFH1 specific for human fibronectin gene, described by A. R. Kornblihtt [23]. l pHCAL 1U specific for human pro-cu 1 (I) collagen and pHFS3 specific for human pro-a 1 (III) collagen, both described by M. Sandberg and E. Vuorio [24]. l PAL 41, a mouse P-actin probe described by M. Buckingham [25], was used to evaluate cellular activity. l

All cDNA inserts used as hybridization probes were separated from plasmid DNA on low melting agarose and radio-labeled by nicktranslation according to Maniatis et al. [22], so that their specific activities ranged from 1 to 4.10’ cpmlfig. Northern analysis. Prehybridization and hybridization were carried out at 42°C for 24 and 36 hr, respectively, in a solution containing 50% formamide, 5X SSPE, 0.1% sodium dodecyl sulphate (SDS), 0.1% each of Ficoll, polyvinylpyrrolidone, and bovine milk albumin and 200 pg/ml of denatured sonicated salmon DNA. The filters were washed at 37°C with a solution containing concentrations of up to 900 mM NaCl, 90 mM sodium citrate, and 0.1% SDS. The dried filters were exposed to X-ray films at -70°C using intensifying screens. Dot-blot analysis. RNA dot-blots were quantified by direct radioactive determination on the filters in a Beckman spectrometer. Protein

Studies

At confluency, fibroblasts were incubated of culture medium containing 0.75 MBq/ml

for 24 hr at 37°C in 5 ml of (4.5) H proline (1.5

GENE

9

EXPRESSION

TBQ/ml mol, Oris, CEA, Saclay, France) and ascorbic acid (Sigma) at a final concentration of 50 pg/ml. At the end of the incubation, the medium was collected and the cell layer removed in a solution of 0.2% Triton X-100 in PBS. Noncollagenous proteins and collagen content and synthesis were estimated in the culture medium as previously described [26]. Briefly, the fractions were dialyzed against distilled water for 42 hr at 4°C. Aliquots of dialysate were hydrolyzed and derivatized with phenylisocyanate, analyzed by reverse phase high performance liquid chromatography (HPLC, Waters Picotag System) and finally evaluated for [3H]proline ([3H]pro) and [3H]hydroxyproline ( [3H] hypro) radioactivity and concentration. Aliquots of the dialyzed fractions were submitted to limited pepsin digestion (Sigma, ref. 6887) according to Kern [27]. For the typing of the different collagen chains, pepsin digests were resolved by SDSpolyacrylamide gel electrophoresis (SDS-PAGE) according to Laemli [28] both under reducing and nonreducing conditions. Hydroxylation of collagen was estimated as the ratio of [3H]Hyp to [3H]Pro in the pepsin digest. To convert the radioactivity incorporated into hydroxyproline into the percentage of collagen, we used Wiestner’s formula [29];

collagen

% =

2([3WHypro) 5([3H]Pro

- [3H]Hypro)

+ 2([3H]Hypro)

In the cell layer, the total [3H](Pro + Hypro) radioactivity was measured by HPLC both before and after dialysis against distilled water, for 24 hr at 4”C, which permitted us to evaluate the intracellular protein degradation. Intracellular degradation of collagen was estimated by separating radioactive hydroxyproline-containing molecules of low molecular weight (~10 KDa) by dialysis [30] and expressed as the ratio of [3H]Hypro to [3H]Pro before and after dialysis.

RESULTS

Cell Proliferation Freshly dissociated cells were inoculated into culture flasks at about lo5 cells per cm’. The initial attachment of the cells to the culture substrate constituted a selection process, since only 3% adhered. They took about 16 days to reach confluency, during which five population doublings were completed. The initial kinetics of growth in primary culture (phase I of cell growth) have been described in detail elsewhere [31, 321. The pig skin cells had a morphologically homogeneous appearance, and contained small spindle-shaped fibroblasts with a low cytoplasm to nucleus ratio. The mean growth curve for 10 cell lines (Fig. 1) shows that from passages 1 to 9, there was a phase of active proliferation (phase II of growth) during which all the growth parameters remained stable. From passage 9, i.e., after 15 cumulative population doublings, a third phase began during which proliferation declined (phase III of growth, i.e. senescence). Saturation density progressively decreased as the number of large flat cells with a high cytoplasm to nucleus ratio increased. Karyotype anomalies appeared concomitantly, and up to 49% polyploidy was observed in senescent fibroblasts. The mean maximum number of doublings was 17.

10

MARTIN

5

ET

AL.

wRNA 03

t

Fibronectin

Collagen

I

Collaaen

III

0,25 0,125 0,06 I

a

bc

a

de

I

bcde

FIG. 3. Dot-blot analysis of fibronectin, pro-a 1 (I) and pro-a 1 (III) collagen messenger-RNA levels in pig skin fibroblasts at different phases of in vitro cell ageing. Total RNA was prepared either from fibroblasts in primary cultures (b), at passages 1 (c), 6 (d) and 14 (e), or from pig lymphocytes (a) used as a negative control. Aliquots of 0.5 to 0.06 pg of RNA were applied onto a nylon membrane and hybridized with the 32P-labeled cDNA probes. I PO

I

I PI0

I

I P20

w

Passages

FIG. 1. Growth curve of pig skin fibroblasts. Cells passaged from the primary culture (PO) to the eventual cell strains. At each subculture, the 25-cm2 flasks were and the total number of cells at confluency was given as the passage number. Cell senescence was observed from tive population doublings, i.e., at passage 9. The data are for 10 cell lines.

Transcriptional

were serially death of the trypsinized a function of 15 cumulamean values

Control

Northern blot analysis permitted evaluation of the molecular weights of the pig mRNAs. There was one major band for the fibronectin corresponding to 8.7 Kb. There were two bands for the pro-cu 1 (I) chain, with the major band corresponding to 5.6 Kb, and two bands for the pro-cr 1 (III), with the major band corresponding to 5.9 Kb (Fig. 2). Age-related changes in mRNA levels were estimated by dot-blot hybridization (Fig. 3 and Table 1). Fibronectin. The amount of fibronectin mRNA increased during senescence, so that at passage 14 the

cells synthesized 2.5 more mRNA than the primary cells. Collagen. The level of pro-a! 1 (I) mRNA exhibited an overall slight decrease with age. The level of pro-a! 1 (III) mRNA was 2.7 higher during passage 14 than in primary culture. Consequently, the ratio of collagen type I to type III mRNA decreased with age. Actin. The level of P-actin mRNA remained relatively constant throughout the whole lifespan of the cells (mean level: 261 cpm/pg RNA, SD = 29). Hybridization of lymphocyte mRNA. As peripheral lymphocytes do not synthesize the matrix proteins, the expression of matrix genes in these cells was used as a negative control to evaluate nonspecific hybridization. The level of RNA hybridization in dot-blot studies was low, similar for the different probes, and represented 5% (mean value +/- 2) of the lowest values of RNA hybridization observed for the skin fibroblasts.

TABLE

mRNA Passages POn=4 Pln=4 P6n=2 P15 n = 2

1

2

3

FIG. 2. Northern blot analysis of RNA from pig skin fibroblasts. Poly(A)+ RNA (1 pg per lane) was prepared from young cultures (passage 1) and fractionated by electrophoresis on a agarose gel and transferred onto a nylon membrane. The filters were hybridized with cDNA radio-labeled inserts used as probes for fibronectin (lane l), type III collagen (pro-u 1 (III), lane 2), and type I collagen (pro-a 1 (I), lane 3).

1

Assay of mRNA Levels for Procollagen Type I and III, Fibronectin and (3 Actin in Pig Skin Fibroblasts at Different Phases of the Cell Lifespan

WI(I) 2015 f 110 2281 + 109 1303 (1413,1194) 1434 (1393,1475)

Values

(cpmlrg

ol-l(II1) 285 rt 23 319 f 53 386 (318,455) 781 (782, 780)

of total Fibronectin 210 f 70 272 + 32 295 (245, 345) 525 (505,545)

RNA) Actin 273 f 24 214 +- 43 292 (265, 319) 267 (264, 269)

Note. Aliquots of 0.03 to 1 Fg of total RNA were dot-blotted and hybridized to nick-translated cDNA probes. Relative abundance of the specific mRNAs were quantified by radioactive determination on the filters in a spectrometer. The results are expressed as mean (+-SD, n = the number of independent experiments) for the RNAs prepared on different cell lines at early passages (PO = primary culture and Pl = first subculture). At later passages, RNA prepared from one cell line was hybridized in two independent experiments, the results are given as mean and as individual values in parentheses.

FIBRONECTIN

AND

COLLAGEN

GENE

11

EXPRESSION

conclusion, aged fibroblasts synthesized intracellularly a large amount of collagen, half of which was immediately degraded, before secretion.

CUL TURE MEDIUM 3H-Pro

P DISCUSSION

PO

) P6

PI0

PI2

Passages

FIG. 4. Synthesis of collagen ([3H]Hypro) and noncollagenous proteins ([3H]Pro) in the culture medium of skin fibroblasts as a function of the number of passages. Cultured fibroblasts at confluency were incubated for 24 hr with 0.75 MBq/ml tritiated proline and 50 pg/ml ascorbic acid. The different radiolabeled amino acids were separated by HPLC, and their radioactivity was evaluated by liquid scintillation.

Protein Synthesis and Degradation Age-related changes in protein synthesis were studied for collagen types I and III and for the noncollagenous proteins in both the culture medium and cell layer. Culture medium. Protein synthesis was studied after 24-hr incorporation of tritiated proline. In the culture medium, the incorporation of [3H]proline remained relatively stable from the primary culture to the 10th passage, and then rose drastically during senescence (Fig. 4). The collagen hydroxylation rate remained constant throughout the whole lifespan of the cells, with a mean level of [3H]Hypro to [3H]Pro ratio of 0.52 +/- 0.02 (n = 8). The amounts of type I and III collagen were quantified in the culture medium. In primary cultures, fibroblasts synthesized 80% of type I and 20% of type III. In later subcultures, the secretion of both collagen types decreased. Collagen typing was not performed at passage 14, because very little material was available. Cell layer. In the cell layer, which included both the cells and their associated matrix, total protein synthesis remained stable during phase II of growth, and then rose drastically during phase III (Fig. 5). Intracellular radiolabeled collagen was particularly increased during senescence (Table 2). Measurement of tritiated proline incorporation both before and after dialysis permitted evaluation of basal protein degradation (Table 2). Total protein degradation was important and stable in pig fibroblasts, as 54% of total proteins were degraded during phase II of growth and 61% during phase III. Collagen degradation was very limited during phase II and then rose drastically during phase III, thus reaching 52%. In

The alterations in RNA and protein metabolism during in vitro ageing are still controversial. Fibronectin and collagen are among the main proteins synthesized by fibroblasts, but very little is known about the alterations in their turn-over during cell ageing, both as regards their messenger RNA and proteins. Hence, in this work, we studied the variations in the expression of the genes specific for fibronectin and collagen type I and type III throughout the lifespan of pig skin fibroblasts. The fibroblasts were isolated from pig skin. Although the pig is increasingly used as an experimental model, very little is known about the ageing processes in that species. We first set up the growth curve of different skin cell lines. Phase III of cell growth, during which karyotype anomalies appeared concomitantly with diminished cell growth, occurred after 15 population doublings. Like for the other species previously studied, the lifespan of pig fibroblasts is linearly related to the species longevity [33]. Consequently, the present data lend support to the theory that in vitro ageing reflects in uivo ageing. Once the growth curve had been constructed, cell activities were studied by both molecular and biochemical techniques. As regards fibronectin, the present data show that the amount of messenger RNA per cell increased during cell senescence. The proteins labeled with [3H]proline, which reflect the noncollagenous proteins, exhibited a marked increase during phase III, both in the cell layer and culture medium, while their

tI

I

PO

/

I

I

I

CELL

LAYER

,

I

1

P6

I

PI0

I

I

1

I

PI2

Passages

FIG. 5. Synthesis of total proteins in the cell layer as a function of the number of passages. The cell layer included both the cells and their associated matrix. Fibroblasts were incubated for 24 hr with [3H]proline, removed by Triton X-100, and then submitted to HPLC as described under Methods.

12

MARTIN

ET AL.

TABLE Collagenous

and Noncollagenous

Protein

Degradation [3HlW~~

2

in the Cell Layer at Different

+ Pro)

Dialysis

total proteins 106 cpm/106 cells

Phase II n=8

Before After

2.4 f 0.2 1.1 f 0.1

0.024 f 0.001 0.039 -t 0.004

Phase III n=6

Before After

4.7 f 0.1 1.8 f 0.3

0.22 0.10

Growth

phase

Phases of the Cell Lifespan

Ratio [3H]Hyp/[3H]Pro

* 0.03 -+ 0.01

Percentage synthesized collagen 1.02 f 0.09 1.6 k 0.2 10.5 5

IL 1.8 -t 0.05

Note. The [3H](Hypro + Pro) radioactivity associated with the total proteins in the cell layer was evaluated by HPLC before and after dialysis. The modifications of the ratio between [3H]hydroxyproline and [3H]proline were used to estimate the variations of collagen degradation in senescent fibroblasts (phase III) as compared to proliferating fibroblasts (phase II). The data are mean f SE, n = the number of independent experiments, phase II = passages 0 to 10, phase III = passages 11 to 14.

intracellular degradation decreased. Marked changes at the protein level have been described in the literature concerning fibronectin in senescent cells. An increase in fibronectin and total cellular protein synthesis was described in late passage human diploid fibroblasts [12, 191. Fibronectin deposition at the cell surface has been reported to increase in chick old fibroblasts [181, and to decrease in human old fibroblasts [19, 341. Fibronectin isolated from late passage fibroblasts was described as unable to support the normal cell substrate interactions [35], and not identical to that isolated from young fibroblasts 1361.As it is known that different fibronectin messenger RNAs can arise from a single gene by altered splicing [37], it is possible that old fibroblasts synthesize transcripts different from those of young fibroblasts, and consequently different forms of fibronectin. As regards collagen, the present data show that the amount of messenger RNA specific for type III increased during senescence while that for type I slightly decreased. At the protein level, the collagen content and synthesis in the culture medium declined throughout subculture, reaching a low level at the end of the cell lifespan. This decline mainly accounted for that of type I collagen, as it constituted 80% of the collagen secreted during primary culture. In the cell layer, the amount of total synthesized proteins and particularly the collagen fraction increased during senescence. While the overall protein degradation remained relatively constant, intracellular collagen degradation highly increased during senescence. Half of the newly synthesized molecules of collagen were thus degraded before their extracellular secretion. The modifications in the amount of collagen observed in the culture medium during ageing may therefore result from regulations at the level of both the RNA transcription and the intracellular degradation of proteins. Very little has been published about collagen and in vitro cell ageing. Decreases in collagen synthesis similar to those observed here have been found in the protein synthesis of diploid human fibroblasts. Houck reported

[15] that after 39 doublings, WI-38 fibroblasts apparently lost much of their ability to synthesize collagen. Collagen fibers, which are numerous in young mouse cell cultures, can no longer be seen under the electron microscope in old cell cultures [38]. Different species may exhibit different alterations in collagen metabolism during ageing, as an increase in collagen synthesis with cell age has been reported in the rat [17, 391. Various theories concerning in vitro cell ageing and senescence are described in the literature. For example, the genetically programmed theory states that ageing is the last stage of development and is controlled by specific genes. The present experiments, in which the same activations were found in the gene expression of different cell lines, lend support to that theory. It may be that a set of genes such as one that includes the cyclin gene [9] is activated at the end of the cell lifespan, leading to the arrest of cell growth and to changes in overall gene expression. Other genes have also been reported to be activated or repressed during cell ageing. For instance, the proto-oncogene Ha-Ras-1 undergoes up to fourfold amplification in human senescent cultured fibroblasts, which results in correspondingly high levels of mRNA and its p21 protein [40]. On the contrary, the expression of the erbB, myc, ras-K and src proto-oncogenes was reduced by 80% at the end of the in vitro lifespan of primate fibroblasts [41]. If the alterations in the fibronectin and collagen genes that we observed indeed result from a switch to a new differentiation program, it is surprising to observe that, for the extracellular matrix, the terminal phenotype of pig senescent cells resembles the phenotype normally found during the processes of repair, cicatrization, and embryonic development. In humans and in several animal species, it was shown that the synthesis of fibronectin and collagen type III proteins are both activated during the early phases of wound healing [42, 431. In embryonic development, a similar link between the two proteins has been described: fibronectin can be detected very early in mouse embryo [44], the propor-

FIBRONECTIN

AND

COLLAGEN

tion of type III to type I collagen is much higher in fetal than adult tissue [45], and the distribution of fibronectin and collagen type III are often similar [46]. If the theory of programmed senescence reflects the reality, the activation of two genes characteristic of immature scars and embryonic development probably has a precise function. The activation of a gene linked to enhanced replication like the Ha-Ras-1 gene is also a paradox in human nonreplicating senescent cells. More data on the expression of specific genes are clearly necessary to improve understanding of the fundamental processes of cell ageing and death.

GENE

21.

Aviv, 1408.

22. 23.

Maniatis, Cloning.

1. 2. 3.

Daly, Tsuji,

C. H., and Odland, T., and Hamada,

Hayflick,

G. F. (1979) J. Znuest. Dermatol. 73,84. T. (1981) Brit. J. Dernatol. 105, 57.

L., and Moorhead, Adu.

T., Fritsch, A Laboratory

24.

Sandberg,

M.,

25.

Alonso, S., Minty, A., Bourlet, Mol. Evol. 23, 11.

26. 27.

Proc.

Natl.

Acad.

E. F., and Sambrook, Manual. Cold Spring

A., Vibe-Pedersen, Acad. Sci. USA 80, and Vuorio,

K., and 3219.

E. (1987)

J. (1982) Harbor,

J. Cell. Biol.

Bienkowsky,

31.

Martin, 49,821.

M., Remy,

32.

Martin, to1. 93, Rohme,

M., Remy, J., and Daburon, 497. D. (1981) Proc. Nutl. Acad.

R. S. (1984)

G. M.,

6.

Schneider,

Cell 6, 179.

7. 8.

Whatley, S. A., and Hill, B. T. (1979) Cell. Biol. Znt. Rep. 3, 671. Adam, G., Simm, A., and Braun, F. (1987) Exp. Cell. Res. 169, 345.

9.

Wang, K. M., Rose, N. R., Bartholomew, E. A., Balzer, M., Berde, K., and Foldvary, M. (1970) Exp. Cell. Res. 61, 357. Kaftory, A., Hershko, A., and Frey, M. (1978) J. Cell. Physiol. 94, 147.

36. 37.

Kornblitt, A. R., Umezawa, K., Vibe-Pedersen, F. E. (1985) EMBO J 4, 1755.

11.

Tollefsbol, 53.

38.

Brachet, J. (1985) Jovanovich, Ed.),

12.

Shevitz, J., Jenkins, ing Dev. 35, 221.

39.

Mollenhauer, 165.

Dev. 35,47.

40.

Srivastava, A., Norris, stein, S. (1985) J. Biol.

Proc.

41.

Nakamura, 177.

42.

Kurkinen, M., Vaheri, Lab. Invest. 43, 47.

T. O., and Cohen,

H. J. (1985)

C. S., and Hatcher,

V. B. (1986)

Age-

Sojar,

16.

Stidworthy, M. L. (1972)

17.

Kontermann, K., and Bayreuther, K. (1979) Gerontology 25, 261. Courtois, Y., and Hugues, R. C. (1976) Gerontology 22, 371.

20.

P. J., Kurtz,

Ageing

Mech.

Green, H. (1966) Nature (London) 5, 631. Houck, J. C., Sharma, V. K., and Hayflick, Exp. Biol. Med. 137, 331.

19.

Mech.

Dev. 30,

14. 15.

G. H., Ellis-Gill, In vitro 7, 275.

M. (1986)

Ageing

13.

18.

H. T., and Rothstein,

Mech.

L. (1971) M.

J., and

Sot.

Burrans,

Vogel, K. G., Kendall, V. F., and Sapien, R. E. (1981) J. Cell. Physiol. 107, 271. Chirgwin, J. M., Przybyla, A. E., Macdonald, R. J., and Rutter, W. J. (1979) Biochemistry 18, 5294.

Received Revised

October 25, 1989 version received November

13, 1989

34.

Aizawa, S., Mitsui, Exp. Cell. Rex 127,

35.

Chandrasekar, Natl. Acad. Sorrentino, 83.

F. (1989)

J. 182,

R., and

Sci.

Y., Kurimoto, 143.

J. Invest.

USA

F., and

Biol.

Derma-

78,5009. Nomura,

K. (1980)

S., Sorrentino, J. A., and Millis, A. J. (1983) Proc. Sci. USA 80, 4747. J. A., and Millis, A. J. (1984) Mech. Ageing Deu. 28,

Molecular cytology. Cell Vol. 2, pp. 421. Academic

J., and Bayreuther,

interactions. Press, New

R. W. (1987)

A., Roberts,

Mech.

Gold-

AgeingDev.

39,

P., and Stenman,

Grinnel,

R. E., and Johnson,

26,

32,

R. J., and

Wartiovaara, J., and Vaheri, A. (1980) Development mals (M. H. Johnson, Ed.), Vol. 4, pp. 233, Elsevier/North land Biochemical Press. Epstein, E. H. (1974) J. Biol. Chem. 249,3225. Borg, T. K., Gay, 2,211.

J. Cell. Biochem.

(H. B. York.

Differentiation

J. S., Shmookler Reis, Chem. 260, 6404.

K. D., and Hart,

F. (1984)

K. (1986)

K., and Baralle,

44.

46.

J.

L., and

Int. J. Radiat.

43.

45.

1077.

M. (1986)

T., Timpl,

F. (1986)

Cristofalo,

33.

104,

Coil. Rel. Res. 4, 399.

J., and Daburon,

Holliday, R., Huschtscha, L. I., Tarrant, T. B. (1977) Science 198,366. S. S. (1975)

Molecular New York.

El Nabout, R., Martin, M., Remy, J., Kern, P., Robert, Lafuma, C. (1989) Matrix 9, 411. Kern, P., Moczar, M., and Robert, L. (1987) Biochem. 337. Laemli, V. K. (1981) Nature (Londonl227,680.

30.

69,

F. E. (1983)

Y., and Buckingham,

Wiestner, M., Rohde, H., Hell-, O., Krieg, Muller, P. K. (1982) EMBOJ. 513.

and Kirkwood,

Sci. USA

Baralle,

29.

Cell. Res. 25,585.

Res. 4, 45.

P. (1972)

5.

E. L., and Shorr,

Cerontol.

Exp.

Leder,

4.

10.

V. J. (1972)

P. S. (1961)

H., and

Kornblihtt, Proc. Natl.

28. REFERENCES

13

EXPRESSION

S. (1980)

107.

L. D. (1982)

in MamHol-

Coil. Rel. Res.