Cell Differentiation and Development, 28 (1989) 105-118

Elsevier Scientific Publishers Ireland, Ltd.

105

CELDIF 00618

Reversible dedifferentiation and redifferentiation of a melanized cell line from a goldfish tumor

Shui-Chou Chou, Changfeng Yang ‘, Victoria A. Kimler, John D. Taylor and T.T. Tchen

Departments of Chemistry and Biological Sciences, Wayne State University, Detroit, MI 48202, V.S.A

(Accepted 10 July 1989)

We reported previously the isolation of a melanized cell line that can undergo reversible dedifferentiation and redifferentiation. A heavily pigmented cell line, designated as P15, originally isolated by fish serum-in- duced melanization of some GEM 81 cells, cloned and serially passaged in fish serum medium, became noticeably less pigmented after several months in fetal calf serum medium and completely unpigmented after another year in the same medium. Addition of fish serum to the medium of this dedifferentiated cell line, designated P15D, induced pigmentation within a week. This re-induced pigmented cell line, designated as PlSDI, became unpigmented when cuhured in fetal caif serum medium for one month. We report here that the dedifferentiation of P15 occurs in two stages. One week after withdrawal of fish serum, the specific activity of tyrosinase of the culture dropped by - 70% and remained at this reduced level for at least one month. After one year, the specific activity of tyrosinase had dropped to a barely detectable level and the culture became completely unpigmented (P15D). Electron microscopic studies showed that the P15D cells have no melanosomes, probably no large vesicles for melanosome formation, but some dopa-positive trans-Golgi network (TGN). Addition of fish serum to the growth medium of P15 cultures led to a steady increase in the specific activity of tyrosinase, detectable after one day. There was also an increase in the amount of dopa-positive TGN within one day. Melanosomes first appeared after three days and became numerous after one week. Upon removal of fish serum, these re-induced cells (PlSDI) underwent a rapid decrease in the specific activity of tyrosinase, reaching, after eight days, the basal level seen in P15D cells. We also report that a protein designated as ~75 (M, - 75000), previously shown to be associated with melanosomes in two melanized cell types of goldfish origin, is present in all melanized cell lines, including P15 and P15DI but absent in unmelanized cell lines, including P15D.

Melanocyte; Tyrosinase; Differentiation; Dedifferentiation

Correspondence address: T.T. Tchen, Department of Chemistry, Wayne State University, Detroit, MI 48202, U.S.A.

’ Visiting scholar, Wuhan University, Wuhan, People’s Republic of China.

Abbreviations: 2D, 2 dimensional; BSA, bovine serum albumin; dopa, dihydroxyphenylalanine; FBS, fetal bovine serum; FS, fish serum from red carp; medium A, medium 199 supplemented with 10% FBS, 2% FS and 0.05 IU/ml of ACTH; medium B, Leibovitz’s L-15 medium with 10% FBS; medium F, Leibovitz’s L-15 medium with 10% FBS and 10% FS; medium 2B, Leibovitz’s L-15 medium

with 20% FBS; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PBS, Dulbecco’s phosphate buffered saline,

pH 7.4; PPO, 2,5-diphenyloxazole; A4,. relative molecular mass; TEM, transmission electron microscopy; Designation of/abbreviation for cell lines, see Table I; TGN, tram Golgi network.

0922-3371/89/$03.50 0 1989 Elsevier Scientific Publishers Ireland, Ltd.

106

Introduction

We reported previously that treatment of a goldfish erythrophoroma cell line GEM 81 (Matsumoto et al., 1980) with fish serum (FS) led to the formation of melanized clones with differ- ent phenotypic characteristics (Matsumoto et al., 1982; 1983; 1984). Recently, we cloned additional cell lines with stable phenotypic characteristics (Chou et al., 1989). Among these are cell lines which are capable of reversible dedifferentiation and redifferentiation - becoming unpigmented when cultured in the absence of FS and regaining the pigmented phenotype when transferred to a medium containing FS. These cell lines present two interesting features and usages. First, they are, to our knowledge, the only cell lines that are capable of undergoing reversible differentiation and dedifferentiation. As such, they should be useful for studying the biochemical basis of these two processes. Second, differentiation of these cell lines involves the formation of melanosomes. Pre- viously, we have shown by electron microscopic studies that these organelles are formed in a simi- lar manner as melanosomes in other systems (Seiji et al., 1971; Jimbow and Fitzpatrick, 1973) except that the goldfish melanosomes do not have the internal protein matrix. Briefly, Golgi-derived small vesicles bearing tyrosinase fuse with large vesicles and evert into the interior of the large vesicles to form multivesicular bodies. Melanin is then synthesized around the small vesicles in the multivesicular bodies until mature melanosomes are produced (Turner et al., 1975). This process thus relates to the general problem of protein trafficking, namely, how the contents of different Go&derived vesicles are delivered to different target sites such as secretory granules, lysosomes, plasma membrane and melanosomes. It seems rea- sonable to assume that this specificity depends, at least in part, on the specific recognition between the different Go&i-derived vesicles and the target sites and that this specificity involves specific pro- tein components. It is also reasonable to assume that, in order for these organelles to undergo aggregation or dispersion under different condi- tions, there should be specific organelle proteins that interact with the translocation apparatus for

dispersion and aggregation (anterograde vs. retro- grade transport). With the reversible differentia- tion/dedifferentiation cell lines, it is now possible to approach the above two aspects biochemically. Since the morphological endpoints for differentia- tion and dedifferentiation (formation and disap- pearance of melanosomes) are reached only after days to weeks, and since the lack of melanosome formation could have several possible causes (no tyrosinase, no large vesicles or no fusion between Go&i-derived small vesicles with large vesicles), we began with studies on biochemical and ultra- structural changes accompanying differentiation and dedifferentiation. The results presented in this paper show the following: (1) the dedifferentiation of the pigmented cell line P15 occurs in two stages - a 70% decline in the specific activity of tyrosinase in 8 days and a subsequent decline after one year to a barely detectable level (P15D); (2) rediffer- entiation of P15D to a pigmented state (PlSDI) is accompanied by a rise in the specific activity of tyrosinase in the culture while the dedifferentia- tion of P15DI is accompanied by a decline in the specific activity of tyrosinase in the culture. These changes occur within days and before gross mor- phological changes are evident; (3) the melanized phenotype is associated with the presence of another protein designated as p75 (M, 75 000) and previously shown to be associated with melano- somes (Clark et al., 1985); (4) fully dedifferenti- ated cells still have dopa-positive trans Golgi net- works (TGN), although less pronounced than those in differentiated cells; (5) treatment of dedifferen- tiated cells with fish serum supplemented medium led to more pronounced dopa-positive TGN within one day, followed by the appearance of a few melanosomes after three days and an abundance of melanosomes within 1 month.

Materials and Methods

The materials and their sources (in parenthesis) are listed below: Ampholines (LKB, Cambridge, England); Araldite, Epon, Formvar, glutaralde- hyde, osmium tetroxide (E.M. Sciences, Fort Washington, PA); ATP, BSA, chloramphenicol,

cycloheximide, L-dopa, glycine, 2-mercaptoetha- nol, neuraminidase, Penicillin G, PPO, sodium cacodylate, TX-100, Tris, L-tyrosine (Sigma Chemical Co., St. Louis, MO); [y-‘*P]ATP, [3H]leucine, [3,5-3H]tyrosine (ICN, Irvine, CA); Celite 545, Norite A (Fisher, Detroit, MI); FBS, Leibovitz’s L-15 medium, medium 199 (GIBCO, Grand Island, NY); FS (Effort Co., Troy, MI); NP-40 (Particle Data Lab, Ltd., Elmhurst, IL); SDS, ultra-pure urea (Schwarz/ Mann, Cleveland, OH); Bio-Bead SM-2, Coomassie blue R-250 and other electrophoretic reagents (Bio-Rad, Rich- mond, CA).

Polyacrylamide gel electrophoresis and miscella- neous techniques

One-dimensional SDS-PAGE and 2D-PAGE were performed as described by Laemmli (1970) and by O’Farrell (1975). The gels were either stained with 0.05% Coomassie blue R-250 in 10% acetic acid, 25% methanol, 65% water for proteins or processed for fluorography for the radio-labeled proteins (Bonner, 1984). For localization and quantification of tyrosinase, 2-mercaptoethanol was omitted for electrophoresis and the gels were stained with L-dopa according to Jimbow et al. (1981). Gel electrophoresis performed in the pres- ence or absence of 2-mercaptoethanol did not change the protein pattern (data not shown). Pro-

TABLE I

Cell lines, growth media and phenotypes

107

tein concentration was determined by Lowry’s method (1951) as modified by Peterson (1977) using bovine serum albumin (BSA) as standard.

Cell lines and their maintenance The cell lines used and their maintenance have

been reported previously. These, together with ref- erences, are shown in Table I. Briefly, we used both tumor cell lines and normal cell lines. In the case of tumor cell lines, a GEM 81 culture was incubated with FS, leading to the formation of a few colonies of pigmented cells. Cells from indi- vidual colonies were collected and grown in FS medium, leading to cultures with a high per- centage of pigmented cells. Single cell cloning then gave different pigmented cell lines designated as P6, Pll etc. (P for pigmented). Unpigmented cells in the FS-treated GEM 81 culture were also collected and passaged (but not cloned). These cell lines, designated as Nl, N2 etc. (N for nonpig- mented) can no longer respond to FS by the formation of pigmented colonies, presumably be- cause all the cells capable of FS-induced melani- zation had already been induced in the original FS-treated GEM 81 culture. In the case of normal cell lines, scales of xanthic goldfish were subjected first to dissociation of all/most of the epithelial cells and then a second dissociation to obtain a mixture of dermal nonpigmented cells and

Cell line Origin Phenotypes Growth *

medium References

GEM 81 Goldfish erythrophoroma No pigmentation B Matsumoto et al., (1980)

N4 FS-treated GEM 81 No pigmentation F Chou et al., (1989)

P6, Pll, P15 FS-induced GEM 81 Pigmented F Chou et al., (1989) (melanized)

P15D Dedifferentiated line from P15 Unpigmented 2B Chou et al., (1989)

PISDI Redifferentiated from P15D Pigmented F Chou et al., (1989)

XIM ** FS-induced from goldfish dermal Pigmented A Grabowski et al., (1983)

nonpigment cell Clark et al., (1985)

NPC-F ** FS-treated goldfish dermal Unpigmented A Grabowki et al., (1983)

non-pigment cell Clark et al., (1985)

* Growth Media B, F, 2B, A (see abbreviations, p. 105). * * XIM: a mixed culture of pigmented and nonpigmented cells was obtained by treatment of a dermal nonpigment cell culture with

FS and ACTH and propagated in the same medium (medium A). Pigmented cells (XIM) and nonpigmented cells (NPC-F) were separated by density gradient centrifugation.

108

xanthophores. These two cell groups were sep- arated by density gradient centrifugation and the nonpigmented cells maintained by serial passages in medium supplemented with fetal calf serum

and/or frozen in liquid nitrogen. When cultures of these nonpigmented cells were treated with FS,

the formation of a small number of pigmented colonies was again observed. Cells were collected from these colonies and passaged in FS medium and melanocytes isolated by density gradient centrifugation. These melanocytes were prop- agated as a mixed culture with nonpigmented cells (essential for growth of normal melanocytes) and are designated as XIM (Xanthic goldfish induced

melanocytes). From these mixed cultures, unpig-

mented cells were again collected by density gradi- ent centrifugation and are designated as NPC-F

(Non pigmented cells from fish serum medium).

Tyrosinase assay Rates of 3H-labeled tyrosine hydroxylation to

dopa with the formation of tritiated water (Pomerantz, 1964) were determined according to Hearing and Eke1 (1976). Tyrosinase activity, cor-

rected for blanks (without enzyme), were calcu- lated according to Townsend et al. (1984). Incuba- tions were at 37°C for 1 h with 50 PM L-dopa as cofactor. All tyrosinase data are shown as specific activity: units of tyrosinase/mg protein, with each unit of tyrosinase representing the oxidation of 1

pm01 of tyrosine per min at 37 o C.

Labeling of cells and analysis of cellular proteins Labeling of XIM and NPC-F cells with

[3H]leucine and two dimensional electrophoretic

analysis of proteins (C)‘Farrell, 1975) were as de- scribed by Lynch et al. (1986). GEM-81 and de- rived cell lines were labeled with [3H]leucine (200 pCi/ml with specific activity of 110 Ci/mmol)

for 24 h in the culture medium based on L-15 following the procedure of Lynch et al. (1986). Endogenous protein phosphorylation was per- formed according to de Graan et al. (1985).

Electron microscopic studies of melanogenesis Cells in T25 flasks were rinsed with ice-cold

PBS, fixed with 4% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 60 min, washed by

8 E

3 s

2000 1 +3-a

0 0 2 4 6 8 10

Days in Medium F

Fig. 1. Changes of tyrosinase (spec. act.) during redifferentia-

tion of PlSD and GEM 81 cells. P15D and GEM 81 cells were

cultured in medium F for different durations of time and

assayed for tyrosinase activity. 4 x lo6 cells were used for each

tyrosinase measurement. (a) P15D; (b) GEM 81. There was a

steady increase of tyrosinase activity for P15D cells, coincident

with melanization (Fig. 5). Although some GEM 81 cells were

induced to differentiation in medium F, the percentage was too

small to produce a measurable contribution to tyrosinase. Note

that there is low basal tyrosinase activity at day 0, seemingly

slightly higher with P15D than with GEM 81 cells.

the same buffer and incubated with or without 0.1% (w/v) L-dopa at 30 o C for 5 h (Turner et al., 1975). They were post-fixed with 1% osmium tetroxide for 1 h, dehydrated with an ethanol series and embedded in an Epon/Araldite mix- ture according to Mollenhauer (1964). Sections were cut with a diamond knife, stained with uranyl

acetate and lead citrate, and viewed in a Philips 201 or 301 electron microscope (60 or 80 Kv).

Results

Tyrosinase activity during redifferentiation Addition of FS to cultures of two un-melanized

cell lines P15D and GEM 81 led to an increase in tyrosinase activity in P15D cells but not in GEM 81 cells (Fig. 1). It should be noted that tyrosinase activity is expressed as specific activity and not total activity per culture. Thus, the increase of specific activity of tyrosinase observed with P15D cells is not due to increase of cell number in the cultures but reflects increase in enzyme activity per cell. In the case of GEM 81 cells, there was no significant change in tyrosinase activity. This is as

109

IA1 Effect of FS Concentration on Tyrosinase Activity

3000 A

+-a . +b

0 2 4 6 8 10

Days of Culture

I81 Effect of FS Concentration on Cell Proliferation

0 10 20 Days of Culture

Fig. 2. Effects of fish serum concentration on P15D cells.

P15D cells were cultured in different culture media (indicated

below), and their tyrosinase activity (A), cell number (B) were

measured on different days. Curves a-d: (a) medium 2B; (b)

medium B+5% FS; (c) medium F (medium B+ 10% FS); (d)

medium B+ 20% FS. Increase of FS concentration in the

culture medium resulted in more rapid increase of the specific

activity of tyrosinase and the rate of cellular proliferation.

expected as only a small percentage of GEM 81 cells are capable of undergoing melanization (Matsumoto et al., 1982).

Effects of FS concentration on tyrosinase activity and growth of P15D cells

The effects of different concentrations of FS on Pl5D cells are shown in Figs. 1 and 2. It is clear

that FS promoted not only tyrosinase activity

(Figs. 1 and 2A) but also the growth of P15D cells (Fig. 2B) in a concentration-dependent manner.

Tyrosinase activity during dedifferentiation

Upon the removal of FS, both PlSDI and P15

cells underwent a decrease in the specific activity of tyrosinase (Fig. 3). Demelanization of the

heavily pigmented P15 cells occurred in two stages. There was a rapid decline in the specific activity

of tyrosinase for the first 4 days, followed by a

IA1

-m- P15DI

A

0 2 4 6 8 10 Deprivation of FS (days)

IBI

-G E 6000 5 3

.: 5000

2 5 4000

w

3000 -I

0 2 4 6 8 10

Deprivation of FS (days)

Fig. 3. Tyrosinase activity during dedifferentiation of P15 and

P15DI. Removal of fish serum (FS) by changing medium F to

medium 2B caused decline in the specific activity of tyrosinase

of: (A) PlSDI cells: obtained by culturing P15D cells in

medium F for 1 month. The specific activity of tyrosinase

declined to a very low but detectable level within 8 days; (B)

heavily melanized P15 cells. The decline of tyrosinase activity

leveled out after 6 days and remained at this level for at least

one month (data not shown).

110

slower decline for another 4 days (Fig. 3B). Thereafter, tyrosinase activity was maintained at a reduced level for at least 45 days (data not shown). After 1 year, however, the culture became com- pletely unpigmented and the resulting dedifferen- tiated cell line (P15D) had very low tyrosinase activity (see Figs. 1, 2A and 4). Redifferentiated cells (P15DI), obtained after 1 month of culture of P15D in FS medium, were not as heavily melanized as P15 cells and had a lower specific activity of tyrosinase (compare initial values in Fig. 3A and B). Upon FS removal, the specific activity of tyrosinase declined to a barely detectable level within 6-8 days (Fig. 3A).

One-dimensional SDS-PAGE analysis of tyrosinase and proteins during redifferentiation of PISD

During induction of melanization of P15D by FS, there was only one noticeable change of stain- ing pattern after SDS-PAGE, namely, an increase in staining of a band with M, of 68000 (Fig. 4A). This band was black instead of blue in the gel, suggesting dopa staining of tyrosinase. This is shown more clearly in a gel which was stained only with dopa and not with Coomassie blue (Fig. 4B). These results are in general agreement with the results shown in Fig. 1 except that the small increases during the first 2 days seen in Fig. 1 were not visible by the less sensitive method used for Fig. 4.

Fig. 4. SDS-PAGE analysis of tyrosinase during redifferentia- tion of PlSDI. Pellets of 1 X lo6 cells were dissolved in 100 pl

of SDS sample buffer (Laemmli, 1970) without 2-

mercaptoethanol, as tyrosinase was inhibited by mercaptoetha-

nol. 100 /.tg of SDS-soluble proteins from individual samples

were loaded onto different lanes of the gel. (A) stained with

dopa and then with Coomassie blue; (B) stained with dopa

only. Lanes l-8 are from cells in medium without FS (1) and after 2, 3, 4, 5, 6, 7, and 8 days (2-8 respectively) after transferring from medium 2B into medium F. The overall protein profile (Coomassie staining, blue color in original gel) of the different samples was essentially the same. The only difference, indicated by arrow in panel A, was due to dopa staining of tyrosinase (black color in original gel) which is shown more clearly in panel B where the gels were stained with dopa only. The IU, of tyrosinase was estimated to be ap-

proximately 68000 by comparison to M, standards (data not shown).

Dopa-positive TGN and melanosomes during redif- ferentiation

The melanized P15 cells have large numbers of melanosomes (Fig. 5a) and well developed TGN. In contrast, melanosomes are absent in dedifferen- tiated unpigmented P15D cells (Fig. 5b). In some sections of P15D cells, however, dopa-positive

Fig. 5. TEM of melanized, demelanized and re+melanizing cells. (a) Fully melanized P15 cells showed an abundance of melanosomes; (b) fully demelanized unpigmented P15D cells showed absence of melanosomes but presence of some dopa-positive TGN; (c and e) one and three days after transfetring P15D cells into medium F. Both are dopa-stained. Note increase in dopa-positive TGN after 1 day and the presence of some melanosomes after 3 days in medium F; (d) one day after transfer into medium F but without dopa

staining. The electron dense TGN is absent; (f) after 1 month in medium F, there is again an abundance of melanosomes.

112

TGN were present (arrows in Fig. 5b). There were

also some multivesicular bodies in this section. Without serial sectioning of a large number of cells, it is not possible to say whether all P15D

cells have these structures. After l-3 days of culture in the presence of FS, dopa-positive TGN became progressively more pronounced (Fig. 5c

and e). As a control for the dopa-positive TGN, a section which was not dopa-stained is shown in

Fig. 5d. Clearly, the TGN (arrows) did not con- tain electron dense segments. Melanosomes were

first observed after 3 days (Fig. 5e’) and were abundant within 1 month (Fig. 5f).

~7.5: A protein specifically associated with the melanized state

When proteins of different cell lines were ex-

amined by 2D-PAGE, a protein of M, 75000 (p75), previously suggested to be specific for goldfish melanosomes (Clark et al., 1985) was

shown to be present in pigmented cell lines (Figs. 6a-c and 7a and c) and absent in unpigmented cell lines (Figs. 6d-f and 7b). Three features in

these figures are noteworthy. First, during dedif- ferentiation and redifferentiation (P15, P15D and PlSDI in Fig. 7), this protein disappeared after dedifferentiation and reappeared after rediffer-

entiation. Second, in the case of the cell line Pll, which requires FS for growth, ~75 is present in

cells maintained either in the presence of FS (growing cells) or in the absence of FS (non-grow- ing cells) (Fig. 6b and c). Third, cell lines which

were cultured in the presence of FS but remained unpigmented did not have this protein (Fig. 6d-f). It might also be mentioned that our previous work showed the absence of this protein in goldfish xanthophores (Lynch et al., 1986). The identity of this protein to the phosphosialoprotein ~75 previ- ously reported was confirmed by its position in

the 2-D gels (Figs. 6 and 7) its sensitivity to

113

neuraminidase (Fig. 8a), and its in vitro phos- phorylation (Fig. 8b).

Discussion

A few years ago, we came upon the observation that FS can induce the formation of melanized

goldfish cell lines with various phenotypic char-

acteristics and proposed that these cell lines would

be valuable as model systems to study several main problems in cell biology, including regu- lation of cell shape, growth, differentiation (organelle formation) and organelle translocation

(reviewed by Matsumoto et al., 1983). Subse- quently, it was found that the cell lines are often phenotypically unstable. In an effort to isolate phenotypically stable cell lines, we obtained ad- ditional clones with different characteristics with regard to growth, differentiation and response to epinephrine (pigment aggregation) (Chou et al., 1989). One of these cell lines seemed to us to be

particularly suited for studies on the formation,

and regulation of formation, of melanosomes and perhaps useful ultimately in studies on melano- some translocations (aggregation and dispersion). We report here some biochemical and ultrastruct-

ural properties of this cell line. Since these studies are preludes to more detailed studies on these

processes, we shall first discuss the current status of the role of organelle membrane proteins in these processes, followed by the significance of the results presented.

Electron microscopic studies have shown that the formation of melanosomes involves two vesic- ular components; large vesicles which form the melanosome membranes and small Golgi-derived vesicles that deliver tyrosinase into the large

vesicles (Seiji et al, 1971; Jimbow and Fitzpatrick, 1973; Turner et al., 1975). The process also re-

Fig. 6. ~75 as a marker for the melanized state. The protein profiles seen after ZD-PAGE are shown for two pigmented cell lines and

three unpigmented cell lines. (a) pigmented cell lines XIM; (b) Pll in medium F; and (c) Pll in medium 2B for one month; (d)

unpigmented cell lines GEM 81 in medium F for 1 week; (e) N4 in medium F for 2 years; (f) dermal NPC-F in medium A for 6

months. (For designation of cell lines and growth media, see Table I). Arrowheads: three internal standards, clockwise from the top;

~60, p45a (two intermediate filament proteins, Walker et al., 1985) and actin.

quires maturation of tyrosinase, probably involv- ing processing of carbohydrate sidechains (for re- view, see Mishima and Imokawa, 1985). Overall, the molecular aspects of melanosome formation

are still poorly understood although several melanized cell lines are now available, (see Anders et al., 1985; Fisher et al., 1985; Hu, 1985; Eisinger

et al., 1985; Ide, 1985; Matsumoto et al., 1983; Burnett, 1985). Specifically, the mechanism for the proper recognition between these two types of vesicles for their fusion to produce melanosomes is unknown. For this purpose, it would be im- portant to identify the melanosome membrane

protein that specifically recognizes the Golgi-de- rived vesicles bearing tyrosinase.

A better understanding of the melanosome membrane proteins is also important for another major cellular process, namely, organelle translo-

cation. Among pigment cells, many from lower

vertebrates can undergo induced pigment organelle translocations (aggregation and dispersion). Al- though this property of hormonal or neural regu- lated pigment translocations has been known for decades (for reviews, see Fujii and Oshima, 1986; Obika, 1986), it is only recently that the mecha- nisms for these processes have been studied. The most recent studies, using permeabilized cells, showed clearly that there are several different

mechanisms. The translocation of carotenoid droplets in swordfish erythrophores are Ca*+-reg-

4’ ulated and microtubule-dependent (see MacNiven

b and Ward, 1988) while the dispersion of carotenoid _. __ “, ‘.

$8 9 droplets in goldfish xanthophores is actin-depen- dent and regulated by the phosphorylation/ dephosphorylation of an organelle protein (re-

viewed by Tchen et al., 1988). In the case of tilapia melanophores, melanosome translocations

Fig. 7. Correlation of ~75 with the pigmented state during dedifferentiation of P15 and redifferentiation of P15D. This figure shows the protein profiles of pigmented cell line P15 (a); unpigmented cell line P15D (b); and pigmented cell line P15DI

bl SC (c). From Figs. 6 and 7 it is clear that ~75 is present in all

pigmented cell lines and absent in all unpigmented cell lines. Its absence or presence is not related to growth versus non- growth and is also not the direct result of FS in the medium as unpigmented cell lines maintained in medium F did not have

c

115

b Fig. 8. p75 is a phosphosialoprotein. Panel a: Effect of neur-

aminidase treatment on p75 expression. When proteins of

pigmented cell lines were incubated with neuraminidase before

ZD-PAGE, the characteristic ~75 streak (see Figs. 6 and 7)

disappeared. This is illustrated in this panel with P15 but

similar observations were made with other pigmented cell lines.

The exact location of the de-sialated protein has not been

determined. Panel b: in vitro phosphorylation of ~75. When

the proteins of pigmented cell lines (illustrated here with P15)

were incubated with [ y- 32 P]ATP, fluorography after ZD-PAGE

showed that p75 was labeled.

are regulated by protein phoshorylation/ dephos- phorylation although the identity of the protein(s) and the cytoskeletal component involved are un- clear (Rozdzial and Haimo, 1986 a, b). Unfor-

tunately, unlike the goldfish xanthophores, the other cells are not available in sufficient quantity and purity for detailed biochemical studies and the identity of the organelle membrane protein(s) that interacts with the translocation apparatus is unknown.

The cell lines described are unusual in that they can undergo reversible dedifferentiation and redif- ferentiation in response to the withdrawal and

addition of FS to the growth medium, thus provid-

ing an ideal system for studying melanosome for- mation. Since the melanosomes are stable struc-

tures, they are not readily destroyed, and morpho- logical dedifferentiation depends on the growth of the cells and the dilution of pre-existing melano-

somes. Even so, the duration (over 1 year) needed for the dedifferentiation of P15 was much too long and there was evidence of selection/growth of nonpigmented cells. In order to obtain informa- tion on the onset of dedifferentiation and differ- entiation, we investigated the changes in tyrosinase activity. The results show that the dedifferentia-

tion of P15 indeed occurs in at least two stages (Fig. 3). Within 8 days after the withdrawal of FS,

the specific activity of tyrosinase of the culture declined by - 70%. There was then a period of at least one month when the specific activity of

tyrosinase remained at this reduced, but still sub-

stantial, level. It took another year for the culture to become completely unpigmented and for tyrosinase activity to decline to a very low basal

level. This low, but detectable, tyrosinase activity is in agreement with our earlier in vivo studies

which showed that skin homogenates of com-

pletely xanthic goldfish have a detectable level of tyrosinase (Rim et al., 1962). This dedifferentiated cell line P15D was shown to respond to FS by a rapid increase in the specific activity of tyrosinase (Figs. 1, 2 and 4), and the resulting redifferenti-

ated cell line P15DI responded to withdrawal of FS by a rapid decline of the specific activity of tyrosinase (8 days to reach the low basal level)

(Fig. 3). These results, together with electron mi- croscopic results (Fig. 5), show that the rediffer- entiation of P15D and the dedifferentiation of P15DI are not due to cell selection but reflect the inductive effect(s) of FS. The results also con- firmed and extended our previous finding that a

116

phosphosialoprotein is present only in melanized cells (Figs. 6-8) and is probably a melanosome protein.

These results establish the ready reversibility of differentiation/ dedifferentiation between P15D and P15DI cells both morphologically and bio- chemically. To our knowledge, this phenomenon is unique. Although it is known that transdifferentia- tion can occur both in vivo and in vitro and that dedifferentiation followed by redifferentiation probably occur in vivo during organ regeneration, we do not know any system where reversible dif- ferentiation-dedifferentiation occurs in vitro. More importantly, however, the results presented here provide the foundation for several lines of current and future research. First, FS has two different activities: induction of differentiation (melanosome formation) and as a growth factor for some melanized cells of goldfish origin (Grabowski et al., 1983; Chou et al., 1989). Until recently, there were no practical assays for either of these two activities. The results presented here show that tyrosinase activity can be used as a convenient assay for this purpose. Results pre- sented elsewhere on a cell line that requires FS for growth (Chou et al., 1989) provides a convenient assay for the growth factor(s). We are currently in the process of purifying the FS protein(s) respon- sible for these two activities in order to determine whether there is a single factor with two activities or two factors with different activities. Needless to say, without a purified factor, it would be difficult to study its mechanism of action.

Second, as mentioned earlier, the membrane proteins of organelles play important roles in both the formation and translocations of the organelles. Currently, there is very little information on the identity and functions of these organelle proteins, with the xanthophore carotenoid droplet protein p57 being the lone exception (reviewed by Tchen et al., 1988). Unfortunately, there are many differ- ent integral proteins in these membranes and membrane proteins are difficult to purify. With the availability of the P15D and P15DI cell lines, we can approach this problem by cDNA technol- ogy. We reason that these cell lines should have very similar mRNAs except that P15DI should have, in addition, a group of mRNAS specifically

associated with melanosome formation. It should be possible to isolate the corresponding cDNAs by standard methods and to use these cDNAs to study the detail of the regulation of transcription of the corresponding genes. In the long term, it should also be possible to use these cDNAs to produce the corresponding proteins and their anti- bodies as well as anti-sense oligodeoxyribonucleo- tides. These can then be used to study their effects on melanosome translocations in permeabilized cells and on the formation of melanosomes. Since these are really long term objectives, we have not discussed the details of these experiments. Rather, we mention these objectives to indicate the poten- tial usage of the cell lines which we have described here. In the mean time, we are also attempting to generate antibodies against p75 in order to test whether these antibodies may inhibit melanosome translocations in permeabilized melanophores.

Acknowledgement

This work was supported by a NIH grant DK 13724.

References

Anders, F., G. Dess, S. Nishimura and H. Kersten: A molecu-

lar approach to the study of malignancy and binignancy in melanoma of Xiphophoms. In: Pigment Cell, Biological,

Molecular and Clinical Aspects of Pigmentation, eds. J. Bagnara, S.N. Klaus, E. Paul and M. Schartl (University of

Tokyo Press) pp. 315-324 (1985). Bonner, W.M.: Fluorography for the detection of radioactivity

in gels. Methods Enzymol. 104, 460-465 (1984). Burnett, J.B.: Long term culture of nonmammalian pigment

cells. In: Pigment Cell, Biological, Molecular and Clinical Aspects of Pigmentation, eds. J. Bagnara, S.N. Klaus, E. Paul and M. Schartl (University of Tokyo Press) pp. 247-

252 (1985).

Chou, S-C., J.D. Taylor and T.T. Tchen: (1989) Isolation of melanized cell lines with stable phenotypes from a goldfish

erythrophoroma cell line and cryopreservation of these cells

by the use of autologous serum. In vitro. in press. Clark, C.R., G.R. Walker, T.J. Lynch, J.D. Taylor and T.T.

Tchen: Identification of a 75 K dal phosphoprotein specific to melanophore. In: Pigment Cell, Biological, Molecular

and Clinical Aspects of Pigmentation, eds. J. Bagnara, S.N. Klaus, E. Paul and M. Schartl (University of Tokyo Press) pp. 37-43 (1985).

117

De Graan, P.N.E., W.H. Gispen and F.C.G. Van de Veerdonk:

a-Melanotropin-induced changes in protein phosphoryla-

tion in melanophores. Mol. Cell. Endocrinol. 42, 119-125

(1985).

Eisinger, M., 0. Marko, Z. Darzynkiewicz and A.N. Hough-

ton: In: Pigment Cell, Biological, Molecular and Clinical

Aspects of Pigmentation, eds. J. Bagnara, S.N. Klaus, E.

Paul and M. Schartl (University of Tokyo Press) pp.

377-390 (1985).

Fisher, P.B., H. Hermo, S. Pestka and I.B. Weistein: In:

Pigment Cell, Biological, Molecular and Clinical Aspects of

Pigmentation, eds. J. Bagnara, S.N. Klaus, E. Paul and M.

Schartl (University of Tokyo Press) pp. 325-332 (1985).

Fujii, R. and N. Oshima: Control of chromatophore move-

ments in teleost fishes. Zool. Sci. 3, 13-47 (1986).

Grabowski, S., J.D. Taylor and T.T. Tchen: Induction of

melanization in primary cultures of dermal cells from

xanthic goldfish by adrenocorticotropic hormone and

goldfish or red carp serum. Cell Differ. 12, 239-244 (1983).

Hearing, V.J. and T.M. Ekel: Mammalian tyrosinase: A com-

parison of tyrosine hydroxylation and melanin formation.

B&hem. J. 157, 549-557 (1976).

Hu, F.: Pigment cell differentiation in vitro. In: Pigment Cell,

Biological, Molecular and Clinical Aspects of Pigmenta-

tion, eds. J. Bagnara, S.N. Klaus, E. Paul and M. Schartl

(University of Tokyo Press) pp. 369-372 (1985).

lde, H.: In vitro proliferation of iridophores without conver-

sion. In: Pigment Cell, Biological, Molecular and Clinical

Aspects of Pigmentation. eds. J. Bagnara, S.N. Klaus, E.

Paul and M. Schartl (University of Tokyo Press) pp.

341-345 (1985).

Jimbow, K. and T.B. Fitzpatrick: Characterization of a new

melanosomal structural component: the vesiculoglobular

body - by conventional, high voltage and scanning electron

microscopy. J. Ultrastruct. Res. 48, 269-283 (1973).

Jimbow, K., M. Jimbow and M. Chiba: Characterization of

structural properties for morphological differentiation of

melanosome: I. Purification of tyrosinase by tyrosinase

affinity chromatography and its characterization of B16

and Hardin-Passey melanomas. J. Invest. Dermatol. 77,

213-218 (1981).

Kim, K., T.T. Tchen and W. Chavin: Tyrosinase of the gold-

fish Curassim auratus L: Il. Correlation of the tyrosinase

activity with pigmentation. B&him. Biophys. Acta 59,

577-580 (1962).

Laemmli, U.K.: Cleavage of structural proteins during the

assay of the head of bacteriophage T4. Nature 227,680-685

(1970).

Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall:

Protein measurement with the Folin-phenol reagent. J. Biol.

Chem. 193, 265-275 (1951).

Lynch, T.J., J.D. Taylor and T.T. Tchen: Regulation of pig-

ment organelle translocation: I. Phosphorylation of the

organelle-associated protein ~57. J. Biol. Chem. 261,

4204-4211 (1986).

MacNiven, M.A. and J.B. Ward: Calcium regulation of pig-

ment transport in vitro. J. Cell Biol. 106, 111-125 (1988).

Matsumoto, J., T. lshikawa, Prince Masahito and S. Takayama:

Permanent cell lines from erythrophoromas in goldfish

(Carussius auratu.s). J. Natl. Cancer Inst. 64,875-890 (1980).

Matsumoto, J., T.J. Lynch, S.M. Grabowski, J.D. Taylor and

T.T. Tchen: Induction of melanized cells from a goldfish

erythrophoroma: isolation of pigment translocation vari-

ants. Science 217, 1149-1151 (1982).

Matsumoto, J., T.J. Lynch, S. Grabowski, CM. Richards, S.J.

Lo, C. Clark, D. Kern, J.D. Taylor and T.T. Tchen: Fish

tumor pigment cells: differentiation and comparison to

their normal counterparts. Am. Zool. 23, 569-580 (1983).

Matsumoto, J., T. lshikawa, Prince Masahito, S. Takayama,

J.D. Taylor and T.T. Tchen: Clonal heterogeneity in physi-

ological properties of melanized cells induced from goldfish

erythrophoroma cell lines. Differentiation 27, 36-45 (1984).

Mishima, Y. and G. lmokawa: Role of glycosylation in initial

melanogenesis: post-inhibition dynamics. In: Pigment Cell

1985, Biological, Molecular and Clinical Aspects of Pig-

mentation, eds. J. Bagnara, S.N. Klaus, E. Paul and M.

Schartl (University of Tokyo Press) pp. 17-30 (1985).

Mollenhauer, H.H.: Plastic embedding mixtures for use in

electron microscopy. Stain Technol. 39, 111-114 (1964).

Obika, M.: Intracellular transport of pigment granules in fish

chromatophores. Zool. Sci. 3, l-11 (1986).

O’Farrell, P.H.: High resolution two-dimensional electrophore-

sis of proteins. J. Biol. Chem. 250, 4007-4021 (1975).

Pomerantz, S.H.: Tyrosinase hydroxylation catalyzed by mam-

malian tyrosinase: an improved method of assay. B&hem.

Biophys. Res. Commun. 16, 188-194 (1964).

Peterson, G.L.: A simplification of the protein assay method of

Lowry et al. which is more generally applicable. Anal.

B&hem. 83, 346-356 (1977).

Rozdzial, M.M. and L.T. Haimo: Reactivated melanosome

motility: differential regulation and nucleotide require-

ments of bidirectional pigment granule transport. J. Cell

Biol. 103, 2755-2764 (1986a).

Rozdzial, M.M. and L.T. Haimo: Bidirectional pigment gran-

ule movements are regulated by protein phosphorylation

and dephosphorylation. Cell 47, 1061-1070 (1986b).

Seiji, M., H. ltakura and K. Miyazaki: The membrane system

and melanin formation in melanocytes. In: Biology of

Normal and Abnormal Melanocytes, eds. T. Kawamura,

T.B. Fitzpatrick and M. Seiji (University Press, Baltimore)

pp. 221-233 (1971).

Tchen, T.T., R.D. Allen, J. Hayden, S.J. Lo, T.J. Lynch, R.E.

Palazzo, G. Peng, G.R. Walker, B-Y. Wu, F-X. Yu and J.D.

Taylor: Regulation of the distribution of carotenoid drop-

lets in goldfish xanthophores and possible implication to

secretion. Cell Motil. Cytoskeleton 10. 143-152 (1988).

Townsend, D., P. Guillery and R.A. King: Optimized assay for

mammalian tyrosinase (polyhydroxyl phenyloxidase). Anal.

B&hem. 149, 345-352 (1984).

Turner, W.A., J.D. Taylor and T.T. Tchen: Melanosome for-

mation in the goldfish: the role of multivesicular bodies. J.

Ultrastruct. Res. 5, 16-31 (1975).

Walker, G.R., J. Matsumoto, J.D. Taylor and T.T. Tchen:

Intermediate filaments with novel protein composition from

certain goldfish cells. Biochem. Biophys.’ Res. Commun.

133, 873-877 (1985).