Biodumica et Biophydca Acta, ! ! 18 (1992) 239-248 ~5 1992 Elsevier Scie~;ce Publishers B.V. All rights reserved 0167-4838/92/$05.(~1

239

BBAPRO 344J75

Purification and partial characterization of two types of growth-inhibitory protein latently present in rabbit serum T a k a s h i K i m u r a l, K a o r u M i ya z a ki 2, J i n p e i Y a m a s h i t a 3, T a k e k a z u H o r i o 4 and Tomisaburo Kakuno t I Diz'ision of En~'n;otog); Insrilute fi~r Protein Research, Osaka Unicersity, Osaka (JapanL " Dirision of Cell Biofogq. Kihara Insz&ae fiJr Biological.Research. Yokohama Ciff Unh't,a#l. Yokohama (Japan), ~ Radioisotol~, Research Cemer, O~ka Urdrersiff. Osaka (Japan) and 4 Nagahama ln.~titute]'or Bk~'hemical St'h,llt'e, Oriental Yeast Co. Ltd.. Nagahama (Japan) (Received 24 August 1991)

Key words: Growth inhibitor; Serum; Cell culture; (Rabbit)

Normal rabbit serum contained two kinds of growth-inhibitory protein, GI-I and GI-II, in latent forms. These latent inhibitors were activated by incubation at 37°C for 12 h, and their activation was lowered by inhibitors for serine, cysteine and metaUoproteinases. Both growth inhibitors were highly purified in active forms by successive column chromatographies. GI-! showed a major protein band with an M r of 18000 on sodium dodecyl sulfate.polyacrylamide gel electrophoresis, while GI-I! showed a major protein band with an M r of 36000. GI-I and GI-ll half-inhibited the growth of rat tumorigenic cell line (RSV-BRL) at concentrations of 0.5 n g / m l anti 10 n g / m l , respectively. Their effects were cytostatic and reversible at low concentrations, but cytotoxic and irreversible at excess concentrations. Of the 15 cell lines tested, GI-! specificall~ inhibited the growth of rodent and lagomorph cells, whereas GI-ll nonspecifmally inhibited the growth of all cell lines tested. Specificities for cell type and malignancy were not observed with either inhibitor. These growth inhibitors were stable to a reducing reagent and proteinase inhibitors, but labile to urea, acid, organic solvents, trypsin, plasmin and heating at 9S°C for 5 rain. These properties suggested that both growth inhibitors might be distinct from known growth-inhibitory factors.

Introduction There is increasing evidence demonstrating that cell growth is regulated by both signals from growth-stimulatory and inhibitory factors [i]. As the negative growth regulators, type /3 transforming growth factors (TGF/3s), intefferons (a, ft, y) and tumor necrosis factor have been well characterized by their structures and biological activities. These factors exhibit diverse effects on various cellular functions in vitro, de0ending on ceil types and experimental conditions. Recentiy,

Abbreviations: DIP. dii~propylfluorophosphate" FCS, fetal calf serum: IFN-a. -/J and -v, intefferon-a. -/3 and -y; 2-ME. 2-mcrcaptoethanol; PBS, Dulbeccn's phosphate-buffered .saline (Ca"÷- and Mg:%free): PCMB. p-ch|oromercuribenzoic acid; SDS, sodium dodecyi sulfate: TGF-B, t~3~e /3 transforming grov~lh factor, "INF,

tumor necrosisfactor. Correspondence: 1-. Kakuno. Division of Enzymolo~. Institute for Protein Research. Osaka Unix'emily. Suila. Osaka 565, Japan.

different kinds of growth inhibitors have been purified from mouse 3T3 cells [2,3], rat liver [4], bovine mammary gland [5], mouse skin [6], mouse macrophages [7], neonatal mouse brain [8], bovine pituitary gland [9], human small intestine [10], human melanoma [ll], and other sources. In addition, there are a considerable number of reports indicating that sera from various animals contain growth-inhlbitory substances [12-16]. However, it is generally very difficult to isolate and identify growth-regulatory factors present in sera or plasma. Thus only TGF-/] [1] and TNF [17] have been identified as growth-inhibitory factors present in sera. We have studied growth inhibitors present in animal sera [18,19], human piatelets [20] and conditioned media of cultured cells [21,/..], using a nonmalignant epithelial cell line from rat liver (BRL) and its tumorigenic counterpart transformed with Rous sarcoma virus (RSV-BRL) as indicator cells [23]. During the course of the studies, we found that normal rabbit serum contained a factor that was growth-inhibitory for RSVBRL cells [18,24~.. Recently, a growth-inhibitory protein

240 with a molecular weight of 5611111)was highly purified from rabbit serum [25]. The purification of this growth inhibitor has also revealed other growth-inhibitor)' factors that exist in latent forms in the serum and are activated by incubation at 37°C. in the present study. we attempted to purlS" and characterize the growth inhibitors latently premnt in rabbit serum. Materials and Methods

Cell lines used and culutre conditions. The cell lines used were as follows: BRL (epithelial cell line from Buffalo rat liver} [26]. RSV-BRL {BRL transformed by Rous sarcoma virus, clone I) [~,]. NRK-49F (fibroblast cell line from normal rat kidney}. NIH3T3 (non-transtormed fibroblast cell line from Swiss mouse embryo). rasNIH3T3 (v-Ki-ras transformant of NIH3T3). !_.929 (mouse non-transformed fibroblast cell line). RK-13 (epithelial cell line from normal rabbit kidney}, YH-I (diploid fibroblast cell line from human embryo skin} [27], B-32 (SV--4|) transformant of YH-I) [27L HSC-4 (human tongue squamous carcinoma}. A-431 (human vulva epidermoid carcinoma}, HeLa-S3 (human uterine carcinoma), MDCK (epithelial cell line from normal Madin-Darby canine kidney). BSC-I (epithelial cell line from African green monkey kidney), SL-29 {fibroblast cell line from chicken embryo). NIH3T3 was a kind gift from the late Dr. T. Kakunaga, Research Institute for Microbial Diseases. Osaka University, Osaka; YH-I and B-32 from Dr. S. Gotoh, School of Medicine, University of Occupational and Environmental Health, Kita-Kyushu: BRL and A-431 from Dr. K. Nishikawa, Kanazawa Medical University, Kanazawa: and NRK-49F from Dr. K. Matumoto, the Faculty of Science, Kyushu University, Fukuoka. MDCK. HSC4. BSC-I and HeLa-S3 were provided by the Japanese Cancer Research Resource Bank. L929, RK-13 and SL-29 were purchased from Dainippon Seiyaku, Osaka. Japan. Cells were cultured at 37°C in a humidified atmosphere of 5% CO, and 95% air. The basal medium, DME/FI2. consisted of a 1 to 1 mixture of Dulbecco's modified Eagle's medium (GIBCO, Grand Island, NY, U.S.A.) and Ham's F12 medium (GIBCO) supplemented with 15 mM Hepes !.2 mg/ml NaHCO.;, 100 U / m l penicillin G and 100 # g / m l streptomycin sullate. in most cases, cultures were supplemented w~d, 10% fetal calf serum (GIBCO). The standard culture medium was designated as 11)% FCS + DME/F12. Falcon plastic culture dishes were purchased from Bectort, Dickinson Labware (Oxnard, CA, U.S.A.). Assay of growth-inhibitor" acthity. For the standard assay of growth-inhibitory activity, 2500 cells of RSVBRL were seeded in each well of 24-well plates containing I).5 ml of 11)% FCS + D M E / F I 2 and incubated for 2 to 4 h. The cultures were then supplemented with

a small volume of test samples ( < 0.1 ml), which had previously been dialyzed against PBS (8.0 mM Na:HPO~. 1.5 mM KH,PO~. 137 mM NaCI and 2.7 mM KCI). ~:nd further cultured for 4 days. Control cultures ~'~cre supplemented with the same volume of PBS. The gro~n cells were counted with a mierocelicounter CC-I'D8 (Sysmex, Kakogawa, Japan). In the control cultures, the number of RSV-BRL cells increased li111-150-fold during the incubation. The percentage of the number of cells in a test culture to the averaged number of cells in the control cultures was expres.~d as "relative cell number'. Ion-e.~change chronlatograpl~v. Fresh serum from 8week-old rabbits, which was a gift from Oriental Yeast Co. Ltd.. T o ~ o . Japan, was used as the starting material for purification of growth inhibitors. DEAE-Sepharose CL-6B and CM-Sepharose CL-6B (Pharmacia LKB, Uppsala, Sweden) were used for the initial fractionation of rabbit serum. The serum (300 ml) was diluted 10-fold with 20 mM potassium phosphate buffer (pH 6.8) (KP-I buffer) and applied to a DEAE-Sepharose CL-6B column (5 × 15 cm) preequilibrated with KP-I buffer. The charged column was washed with 1000 ml of the same buffer. The resultant nonadsorbed fraction was adjusted to pH 6.0 with acetic acid and then applied to a CM-Sepharose CL-6B column (4 × t0 cm) preequilibrated with 10 mM potassium acetate buffer (pH 6.0). The charged column was washed with 500 ml of the same buffer, and the nonadsorbed fractions were collected and pooled for further purification of growth inhibitors. Hydrogen-bond chromatography Hydrogen-bond chromatography was carried out on a Sepharose CL-6B column (2.5 × 28 cm) (Pharmacia LKB), preequilibrated with KP-I buffer containing 2 M ammonium sulfate, according to the method of Fujita et al. [28]. Samples were supplemented with mlid ammonium sulfate to 2 M and applied to the column. The charged column was eluted with a linear gradient of ammonium sulfate concentration from 2.0 to 0.5 M in 1500 ml of KP-! buffer at a flow rate of 100 ml/h.

Molecular-sieve high-performance liquid chromatography (molecular-sieve HPLC). Molecular-sieve HPLCs were carried out on a preparative TSKgel G3000SWG column (2.5 x 60 cm) and on an analytical TSKgel G3000SW column (0.75 ×60 cm) (TOSOH, Tokyo, Japan). preequilibrated with PBS containing 0.1% (w/v) CHAPS) and with 0.2 M sodium phosphate buffer containing !% (w/v) CHAPS, respectively. Samples were applied to the large and small columns and developed with the above buffer at flow rates of 2.5 ml/min and 0.5 ml/min, respectively.

Anion-exchange high-per[ommnce liquid chromatographies (anion-exchange HPLCs). Anion-exchange HPLCs were carried out on a Shodex IEC QA-824 column (8 x 75 mm) (Showadenko, Tokyo,

241 Japan) and Shodex IEC DEAE-420N column (4.6 x 35 mm) (ShowadenkoL preequilibrated with 10 mM potassium phosphate buffer (pH.8.0)containing !% (w/v) CHAPS (KP-ll buffer). Samples were supplemented with CHAPS to 1% (w/v) followed by dialysis against 10 mM potassium phosphate buffer (pH 8.0) and applied to the column. SDS-polyaco'lamide gel electrophoreds. SDS-polyacrylamide gel electrophoresis was carried out on 12.5% (w/v) and 15% (w/v) polyacrylamide gels (60 x 80 x 1 ram) by the method of Laemmli [29] with a Bio-Rad MINI 2-D electrophoresis apparatus (BIO-RAD Laboratories, Richmond, CA, U.S.A.). Samples were dissolved in 62.5 mM Tris-HC1 (pH 6.8) containing 2% (w/v) of SDS, 8 M urea and 7% (v/v) glycerol without 2-mercaptoethanol. After electrophoresis at 20 mA for 1 h, the gel slabs were stained with a Wako silver staining kit (Wako Chemicals, Osaka, Japan). The molecular weight markers used were phosphorylase b (M, 97400), bovine serum albumin (M, 66200), ovalbumin (M~ 42700), carbonic anhydrase (M, 29000), soybean trypsin inhibitor ( M~ 20100) and lysozyme ( M~ 14300). Determblation of protein concentration. Protein concentrations of purified growth inhibitors were determined with a micro BCA protein assay kit (PIERCE, Rockford, I L U.S.A.), using bovine serum albumin as the standard. Results

Existence of latent growth inhibitor To examine the existence of a latent form of growth inhibitor, fresh serum prepared at 4°C from whole blood of rabbits was incubated at 37°C for 12 h under sterile conditions and then subjected to molecular-sieve HPLC on a TSKgel G3000SWG column. When the fractions eluted from the column were assayed for growth-inhibitory activity on RSV-BRL cells, a strong inhibitory activity was found in fractions corresponding to M, around 10 k (elution volume from 185 to 200 ml) (Fig. 1). However, when non-incubated rabbit serum was fractionated on the same column, the growth-inhibitory activity was hardly detected. These results indicated that normal rabbit serum contained a latent growth inhibitor which could be activated by the incubation at 37°C. The growth-inhibitory activity was also detected in the same region, even if the incubation at 3"PC was carried out after the chromatography (data not shown). When the non-incubated and incubated whole sera were assayed for growth-inhibitory activity before the chromatographic separation, they showed a similar, low activity (50% growth inhibition at about 3% (v/v) serum) (data not shown). This suggested that rabbit serum also contained a factor which could block the

i

i

~

,

i

291~142k ~.7'k3~k 1~'4k

=

[

1

~.O~-

T

oL

o.o

~00 1,50 ~ Etution volume (ml)

50

250

Fig. I. Molecular-sieve HPLC of 4°C- and 37°C-treated normal rabbit sera on a TSKgelG30t~I)SWGcolumn. The rcsullanl fractions were assayed for growth-inhibitoryactivityat a dose of 2 pl/welL Other experimentalconditionsare described in the text. Relativecell number of RSV-BRL with 4°C- (o1 and 37°C- to) treated sera. Az~,J( ~ ) . activity of the activated inlff:,itor. Essentially the same results were obtained with freshly obtained rabbit plasma, indicating that the presence and activation of the latent growth inhibffor were independent of platelets. In order to test the possibility of proteolytic activation of the latent inhibitor, effects of proteinase inhibitors on the activation of the latent inhibitor, which had been obtained by the HPLC of non-incubated serum (elution volume from 180 to 2(10 ml in Fig. I), were examined ('Fable l). The activation of the latent inhibitor was completely blocked by DFP, PCMB or 1,10-phenan~hroline/EDTA but not by pepstatin. This showed that two or three types of proteinase including serine and cysteine proteinases were involved in the activation of the latent growth inhibitor, and that these proteinases and the growth inhibitor were separated in the same fractions by molecular-sieve HPLC.

Fractionation of normal rabbit serum in order to purify the growth inhibitor, 300 ml of rabbit serum was used as the starting material. First, TABLE ! Effects o.fproteinase inhibilors on actit'ation of GI Treatment Control (R.T.. 24 h) 10 mM DFP (R.T. 24 h) I mM PCMB(R.T.. 24 h)" 5 mM IA(J-phenanthrotineand 5 mM EDTA(R.T.. 24 h)" 10 p.g/ml pepslatin(R.T., 24 h ~ " 4°C for 24 h a

Activation('Tr) Ill0 5 3 6 !112 2

After the treatment, sample was dialyzed against PBS, and then assayed. R.T., room temperature.

242 B -i- A

F,

loo

i

2

!" .................... C 0

"

C

",

~

i

~

~' 1

I -

.

.

.

.

.

II ,

'

I .=

1

0.5

-~0.4 ^

'~ o.~

tz- .

i,

J

=_

"._....... o

C

=

~,r:-:___

J¢0

o¢=

~

0.1t-

0.1

± T

0.0

0--

0.0

;

20

30

40

it

50

60

Fraction mmd~r (0.5 nd each)

Fig. 3. Purification of GI-I. (A) Anion-exchange HPLC on a Shodex IEC DEAE-420N column of GI-I fraction obtained by molecular-sieve HPLC. (B) Molecular-sieve HPLC on a TSKgel G3000SW column of Gl-! fraction obtained by anion-exchange HPLC. Relative cell number of RSV-BRL at a dose of 2 Ixl/wetl (e), ,4z~ ~( ), NaCI ( - - - - - - ) . Other experimental conditions are described in the text.

243

97k-

29k-

~

2¸

'

~,.

4.3

44

45

46

47

~.,,~ : : .

,

.,~..

:=.~

~.....,

48

.,..

49

Fraction number

Fig. 4. SDS-polyacrylamidegel electrophoresis under nonreducing conditions of each fraction of GI-I obtained by molecular-sieve HPLC.The arrow indicatesthe major protein hand of 18 kDa

Two peaks of growth-inhibitory activities were eluted at NaCI coucentrations of around 0.05 and 0.2 M, which were tentatively named GI-I and GI-II, respectively.

Purification of GI-I GI-I fractions obtained by the above anion-exchange HPLC (elution volume from 22.5 to 32.5 ml in Fig. 2B) were collected and rechromatographed on the same anion-exchange HPLC column, but with a more shallow NaCi gradient. GI-I was again eluted at 0.05 M NaCI, increasing its specific activity 9-fold (data not shown). The GI-I fractions were combined and concentrated to 0.2 ml with a Centricon 10 (Amicon) and then subjected to molecular-sieve HPLC on a TSKgel G3000SW column (7,5 x 600 ram). GI-I was recovered in a region corresponding to Mr around 15000 (data not shown). The GI-I fractions were collected and

further purified by anion-exchange HPLC on a Shodex IEC DEAE-420N column (Fig. 3A). The charged column was eluted with 2-step linear gradients of NaCI concentration. The growth inhibitor was eluted at approximately 50 mM NaCI (elution volume from 19 to 24 ml). These GI-I fractions were combined and then finally purified by molecular-sieve HPLC on a TSKgel G3000SW column (7.5 × 600 mm) (Fig. 31]). GI-i was eluted in a region corresponding to an Mr of 15000 (fraction No. 45-47). A major peak of A z~0 in fractions 48-49 was due to CHAPS concentrated with a Centricon 10 before HPLC (data not shown). Seven fractions (fractions 43-49) around the activity peak were individually analyzed by SDS-polyacrylamide gel electrophoresis under non-reducing conditions. The peak fractions (fractions 45-47) showed a major specific band of M r 18000 and several minor bands including one with M r 40000 (Fig. 4). Among these bands, the intensities of the protein bands with M, 18000 and M r 40000 well correlated with growth-inhibitory activity, indicating that either or both bands were responsible for the growth inhibition. It seems possible that the 40 kDa band was a dimer of the 18 kDa band. The purification of GI-I is summarized in Table !I.

Purijication of Gl-ll The partially purified GI-II fraction obtained by anion-exchange HPLC on a Shodex IEC QA-824 column (elution volume from 38 to 44 ml in Fig, 2B) was subjected to anion-exchange HPLC on a Shodex IEC DEAE-420N column (Fig. 5A). GI-ll was eluted around 0.1 M NaCI. The peak fractions of GI-ll activity (elution volume from 22 to 25 ml) were combined, concentrated to 0.1 ml and subjected to molecular-sieve HPLC on a TSKgel G3000SW column (7.5 × 600 ram). GI-il was eluted forming an activity peak of 25 kDa, which B --

F

101~ ~

. . . . . . . . . . . .

-

I

1 .

~ 5~

! 2

o.1¢-

/

... "

z

---,

i

.,