Differential expression and processing of two cell associated forms of ...

Molecular Biology of the Cell Vol. 3, 349-362, March 1992

Differential Expression and Processing of Two Cell Associated Forms of the Kit-Ligand: KL-1 and KL-2 Eric J. Huang,* Karl H. Nocka,*t Jochen Buck,* and Peter Besmer* Programs in *Molecular Biology and tImmunology, Sloan Kettering Institute, New York, New York 10021; and Comell University Graduate School of Medical Sciences, New York, New York 10021 Submitted September 27, 1991; Accepted December 30, 1991

The c-kit ligand, KL, and its receptor, the proto-oncogene c-kit are encoded, respectively, at the steel (Sl) and white spotting (W) loci of the mouse. Both SI and W mutations affect cellular targets in melanogenesis, gametogenesis, and hematopoiesis during development and in adult life. Although identified as a soluble protein, the predicted amino acid sequence of KL indicates that it is an integral transmembrane protein. We have investigated the relationship between the soluble and the cell associated forms of KL and the regulation of their expression. We show that the soluble form of KL is generated by efficient proteolytic cleavage from a transmembrane precursor, KL-1. An alternatively spliced version of KL1, KL-2, in which the major proteolytic cleavage site is removed by splicing, is shown to produce a soluble biologically active form of KL as well, although with somewhat diminished efficiency. The protein kinase C inducer phorbol 12-myristate 13-acetate and the calcium ionophore A23187 were shown to induce the cleavage of both KL-1 and KL-2 at similar rates, suggesting that this process can be regulated differentially. Furthermore, proteolytic processing of both the KL-1 and KL-2 transmembrane protein products was shown to occur on the cell surface. The relative abundance of KL-1 and KL-2 in a wide variety of different mouse tissues indicates that the expression of KL-1 and KL-2 is controlled in a tissue-specific manner. Sld, a viable steel allele, is shown to encode a biologically active secreted mutant KL protein. These results indicate an important function for both the soluble and the cell associate form of KL. The respective roles of the soluble and cell associated forms of KL in the proliferative and migratory functions of c-kit are discussed. INTRODUCTION

Conventionally, growth factors are thought of as soluble molecules that bind to cognate receptors on target cells, thus initiating signals for target cell responses. Increasing evidence indicates that, in addition, these factors may also exist as membrane and/or extracellular matrix bound forms, suggesting that some target responses may require cell-cell contact (Gordon, 1991). The soluble and cell associated forms of these factors then provide means to generate the diverse responses of these multifunctional growth factors. The elucidation of mechanisms of formation of the various forms of these factors, the determination of their functional properties, and the investigation of the regulation of their expression are important in understanding how different target cell t Present address: Cytomed Inc., Cambridge, MA. © 1992 by the American Society for Cell Biology

responses of multifunctional factors are generated and controlled. c-kit encodes a transmembrane tyrosine kinase receptor that is a member of the platelet-derived growth factor receptor subfamily and is the gene product of the murine white spotting (W) locus (Besmer et al., 1986; Yarden et al., 1987; Chabot et al., 1988; Geissler et al., 1988; Majumder et al., 1988; Qiu et al., 1988; Nocka et al., 1989). The demonstration of identity of c-kit with the W locus implied a function for the c-kit receptor system in various aspects of melanogenesis, gametogenesis, and hematopoiesis during embryogenesis and in the adult animal; it suggested a role for c-kit in facilitating cell proliferation, cell migration, cell survival, differentiation, and postmitotic functions within these cell systems (Russel, 1979; Silvers, 1979). The ligand of the c-kit receptor, KL, recently has been identified and characterized on the basis of the known 349

E.J. Huang et al.

function of c-kit/W in mast cells (Anderson et al., 1990; Flanagan and Leder, 1990; Nocka et al., 1990a,b; Williams et al., 1990; Zsebo et al., 1990a,b). KL has a molecular mass of 30 kDa and is not a disulfide linked dimer, although in a nondenaturing milieu KL forms noncovalent dimers (Nocka et al., 1990a; Pronovost and Besmer, unpublished data). In agreement with the anticipated functions of the c-kit receptor in hematopoiesis, KL stimulates the proliferation of bone marrow-derived and connective tissue mast cells, and in erythropoiesis, in combination with erythropoietin, KL promotes the formation of erythroid bursts (day 7-14 burst-forming unit-erythroid). Furthermore, recent in vitro experiments with KL have demonstrated enhancement of the proliferation and differentiation of cells within the hematopoietic stem cell compartment, suggesting that there is a role for the c-kit receptor system in progenitors of several hematopoietic cell lineages (Martin et al., 1990; Nocka et al., 1990a). Although KL was purified from conditioned medium of Balb3T3 cells, implying that is a soluble protein, the predicted amino acid sequence of KL, deduced from the nucleic acid sequence of cDNAs, indicates that KL is synthesized as a transmembrane protein. The soluble form of KL then might be generated by proteolytic cleavage of the cell membrane form of KL. Interestingly, the ligand of the colony-stimulating factor 1 receptor, the closest relative of c-kit, shares the unique structural characteristics of KL and has been shown to be cleaved proteolytically to produce the soluble growth factor (Rettenmier et al., 1987; Rettenmier and Roussel, 1988; Bazan, 1991). These findings then might indicate that both a soluble and a membrane-bound form of KL function in the pleiotropic responses of the c-kit receptor system. However, the relationship of these molecules and their precise role are not known. The ligand of the c-kit receptor, KL, was recently shown to be allelic with the murine steel locus on the basis of the observation that KL sequences were found to be deleted in several severe SI alleles (Copeland et al., 1990; Nocka et al., 1990b; Zsebo et al., 1990a). In agreement with the ligand receptor relationship between KL and c-kit, mutations at the steel locus result in phenotypic characteristics that are very similar to those seen in mice carrying W mutations, i.e., they affect hematopoiesis, gametogenesis, and melanogenesis; however, in contrast to W mutations, Sl mutations are not cell autonomous and they affect the microenvironment of the c-kit receptor (Bennett, 1956; McCulloch et al., 1965; Mayer and Green, 1968; Dexter and Moore, 1977; Russel, 1979; Silvers, 1979). The original SI allele is an example of a SI null-mutation. Sli/SI homozygotes are deficient in germ cells, are devoid of coat pigment, and they die perinatally of macrocytic anemia (Bennett, 1956; Sarvella and Russel, 1956). Mice homozygous for the Sld allele are viable, although they lack coat pigment, are sterile, and have macrocytic anemia. The c-kit re350

ceptor system in these mice, therefore, appears to display some residual activity (Bernstein, 1960). Recent analysis of the Sld allele indicated an intragenic deletion that includes the transmembrane domain and the Cterminus of the KL coding sequence (Branan et al., 1991; Flanagan et al., 1991). To further our understanding of the role of the various forms of KL encoded by alternatively spliced KL RNAs and the Sld allele, we have carried out a detailed molecular analysis of these molecules and investigated mechanisms of the regulation of their expression. The respective roles of the soluble and cell associated forms of KL in the proliferative and migratory functions of ckit are discussed. MATERIALS AND METHODS Mice and Tissue Culture WBB6 +/+, C57BI/6J, and 129/Sv-Sld/+ mice were obtained from the Jackson Laboratory (Bar Harbor, ME). 129/Sv_Sld/+ male and female mice were mated, and day 14 fetuses were obtained and used for the derivation of embryonic fibroblasts according to the method of Todaro and Green (1963). Mast cells were grown from bone marrow of adult +/+ mice in RPMI-1640 medium supplemented with 10% fetal calf serum (FCS), conditioned medium from WEHI-3B cells, nonessential amino acids, sodium pyruvate, and 2-mercapto-ethanol (RPMI-Complete [C]) (Nocka et al., 1990c). Balb/3T3 cells (Aaronson and Todaro, 1968) were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% calf serum, penicillin, and streptomycin. COS-1 cells (Gluzman, 1981) were obtained from Dr. Jerrard Hurwitz (Memorial Sloan Kettering Cancer Center, New York) and were grown in DMEM supplemented with 10% fetal bovine serum, glutamine, penicillin, and streptomycin.

Production of Anti-KL Antibodies Murine KL was purified from conditioned medium of Balb3T3 cells by using a mast cell proliferation assay as described elsewhere (Nocka et al., 1990a). To obtain anti-KL antibodies, one rabbit was immunized subcutaneously with 1 gg of KL in complete Freunds adjuvent. Three weeks later the rabbit was boosted intradermally with 1 ,ug in incomplete Freunds adjuvent. Serum was collected 1 wk later and then biweekly thereafter. Immunoprecipitation of "25I-labeled KL and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used to monitor the production of anti-KL antibody. The 125Ilabeled KL used for this purpose was iodinated with chloramine T with modifications of the method of Stanley and Guilbert as described previously (Nocka et al., 1990b).

cDNA Library Screening Poly(A) RNA was prepared by oligo(dT)-cellulose chromatography from total RNA of Balb/c 3T3 fibroblast. A custom-made plasmid cDNA library was then prepared by Invitrogen (San Diego, CA). Essentially, double-stranded cDNA was synthesized by oligo dT and random priming. Nonpalindromic BstXI linkers were ligated to bluntended cDNA, and on digestion with BstXI, the cDNA was subcloned into the expression plasmid pcDNAI (Invitrogen). The ligation reaction mixture then was used to transform Escherichia coli MC1061/P3 by electroporation method to generate the plasmid library. The initial size of the library was -10 independent colonies. For screening of the plasmid library, an end-labeled oligonucleotide probe described previously was used (Nocka et al., 1990b). Hybridization was done in 6X SSC at 63°C, and the final wash of the filters was in 2X SSC and 0.2% SDS at 63°C. The inserts of recombinant plasmids were Molecular Biology of the Cell

Cell Associated Form of KL is Absent in Sld Allele released by digestion with HindIII and Xba I and then subcloned into the phage M13mpl8 for sequence analysis.

twice. Transfected COS-1 cells were grown in DMEM plus 10% FCS, 100 mg/ml L-glutamine, and antibiotics.

Polymerase Chain Reaction Amplification (RT-PCR) and Sequence Determination

Pulse-Chase and Immunoprecipitation Analysis of KL Proteins

Total RNA from tissues and cell lines was prepared by the guanidium isothiocyanate/CsCl centrifugation method of Chirgwin (1979). The RT-PCR amplification was carried out essentially as described previously (Nocka et al., 1990b; Tan et al., 1990). The following primers were used for RT-PCR: primer 1: 5'-GCCCAAGCTTCGGTGCCTTTCCTTATG-3' (nt. 94-107); primer 2: 5'-AGTATCTCTAGAATTTTACACCTCTTGAAATTCTCT-3' (nt. 907-929); primer 3: 5'-CATTTATCTAGAAAACATGAACTGTTACCAGCC-3' (nt. 963-978); and primer 4: 5'-ACCCTCGAGGCTGAAATCTACTTG-3'(nt. 1317-1333). For cDNA synthesis, 10 Mg of total RNA from cell lines or tissues in 50 ,l of 0.05 mM tris(hydroxymethyl)aminomethane (Tris)-HCl (pH 8.3), 0.75 M KCl, 3 mM MgCl2, 10 mM dithiothreitol, 200 ,uM dNTPs, and 25 U of RNasin (BRL, Gaithersburg, MD) were incubated with 50 pmol of antisense primer and 400 U of Moloney murine leukemia virus reverse transcriptase (BRL) at 37°C for 1 h. The cDNA was precipitated by adding 1/10 volume of 3 M NaOAc (pH 7.0) and 2.5 volumes of absolute ethanol and resuspended in 50 Ml of ddH20. PCR was carried out for 30 cycles in 100 Ml of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.01% (wt/vol) gelatin, 200 MM dNTPs, 500 pmol of both sense and antisense primers, and 2.5 U of Taq polymerase (Perkin Elmer-Cetus, Norwalk, CT). HindlIl sites and Xba I sites were placed within the sense and antisense primers, respectively. The amplified DNA fragments were purified by agarose gel electrophoresis, digested with the appropriate restriction enzymes, and subcloned into M13mpl8 and M13mpl9 for sequence analysis (Sanger et al., 1977). The KL-1, KL-2, KL-2, and KL-Sld PCR products were digested with HindIll and Xba I and subcloned into the expression plasmids pCDM8 or pcDNAI (Invitrogen). Miniprep plasmid DNA was prepared by the alkaline-lysis method (Sambrook et al., 1989) followed by phenol-chloroform extraction and ethanol precipitation. Maxiprep plasmid DNA used for the transfection of COS-1 cells was prepared by using the "Qiagen" chromatography column procedure.

Transfected COS-1 cells were used for pulse-chase experiments 72 h after the transfection. Cells were incubated with methionine-free DMEM containing 10% dialyzed FCS for 30 min and labeled with 35S-methionine (New England Nuclear, Boston, MA) at 0.5 mCi/ml. At the end of the labeling period, the labeling medium was replaced with regular medium containing an excess amount of methionine. To determine the effect of phorbol 12-myristate 13-acetate (PMA) and A23187 on the proteolytic cleavage of KL, 1 MM PMA or 1 ,M A23187 was added to the transfected cells at the end of the labeling period after replacement of the labeling medium with regular medium. To assess protein kinase C involvement in KL cleavage, labeled transfected cells were incubated for 1 h with 1 MM of the protein kinase C inhibitor calphostin C (Kamiya Biochemical, Thousand Oaks, CA) and then treated with 1 MM PMA in the presence of calphostin C. The cells and supematants were collected individually at the indicated times for immunoprecipitation analysis. Cell lysates were prepared as described previously (Nocka et al., 1989) in 1% Triton X-100, 20 mM Tris (pH 7.5), 150 mM NaCl, 20 mM EDTA, 10% glycerol, and protease inhibitors phenylmethyl sulfonyl chloride (1 mM) and leupeptin (20 Mg/ml). For the immunoprecipitation analysis of KL protein products, the anti-mouse KL rabbit antiserum was used. The anti-KL serum was conjugated to protein-A Sepharose (Pharmacia, Piscataway, NJ) and washed three times with Wash A (0.1% Triton X-100, 20 mM Tris [pH 7.51, 150 mM NaCl, 10% glycerol). Anti-KL serum-protein A Sepharose conjugate was incubated with supematant and cell lysate at 4°C for .2 h. The immunoprecipitates then were washed once in Wash B (50 mM Tris, 500 mM NaCl, 5 mM EDTA, 0.2% Triton X100), three times in Wash C (50 mM Tris, 150 mM NaCl, 0.1% Triton X-100, 0.1% SDS, 5 mM EDTA), and once in Wash D (10 mM Tris, 0.1% Triton X-100). For gel analysis, immunoprecipitates were solubilized in SDS sample buffer by boiling for 5 min and analyzed by SDS-PAGE (12%) and autoradiography.

RNase Protection Assay

Determination of Cell Surface Expression of KL Protein Products

A ribo probe for RNase protection assays was prepared by linearizing the KL-1 containing pcDNAI plasmid with Spe I. The antisense riboprobe was then synthesized by using SP6 polymerase according to the Promega Gemini kit (Promega, Madison, WI). Riboprobe labeled to high specific activity was then hybridized to 10 or 20 Mg of total RNA in the presence of 80% formamide at 45°C overnight. The hybridization mixture was digested with RNase A and Ti (Boehringer Mannheim, Indianapolis, IN) and treated with proteinase K (Sambrook et al., 1989), and the protected labeled RNA fragments were analyzed on a 4% urea/polyacrylamide gel. Autoradiograms of RNase protection assay were analyzed by densitometry, and parts of the films were reconstructed on a Phospholmage analyzer (Molecular Dynamics, Sunnyvale, CA) for better resolution.

Transient Expression of KL cDNAs in COS-1 Cells For transient expression of KL cDNAs, COS-1 cells were transfected with the DEAE-dextran method described previously (Kriegler, 1990) with minor modifications. Briefly, COS-1 cells were grown to subconfluence 1 d before use and were trypsinized and reseeded on 150mm petri dishes at a density of 6 X 106 cells per dish. After 24 h, the cells had reached -70% confluence and were transfected with 5 Ag of plasmid DNA in the presence of 10% DEAE-dextran (Sigma, St. Louis, MO) for 6-12 h. Medium containing plasmid DNA was removed, and the cells were chemically shocked with 10% dimethyl sulfoxide/phosphate-buffered saline++ (DMSO/PBS++) for exactly 1 min. Residual DMSO was removed by washing the cells with PBS++ Vol. 3, March 1992

For flow-cytometry, transfected cells were harvested from monolayers with PBS containing 5 mM EDTA. The cells were labeled with antiKL serum in PBS containing 5% bovine serum at 4°C for 30 min, washed, and labeled with fluorescein isothiocyanate conjugated goat anti-rabbit IgG serum (Becton-Dickinson, Rutherford, NJ). Cells were then washed and fixed in PBS containing 1% paraformaldehyde and analyzed with a flow cytometer (FACSCAN, BD). Results are expressed as cell number (linear scale) versus fluorescence (log scale). Cell surface iodination was done with the lactoperoxidase-glucose oxidase method as previously described (Harlow and Lane, 1988). Cells were harvested from monolayers by treatment with PBS containing 5 mM EDTA and resuspended in 0.5 ml PBS at 8 X 106 cells/ ml. lodination was carried out in the presence of 1 mCi Na'25I (New England Nuclear). Cell lysates and supematants were analyzed by immunoprecipitation with anti-KL serum and SDS-PAGE.

Determination of Biological Activity of Soluble KL Mast cells were grown from bone marrow of adult WBB6 +/+ mice in RPMI-1640 medium supplemented with 10% FCS, conditioned medium from WEHI-3B cells, nonessential amino acids, sodium pyruvate, and 2-mercaptoethanol (RPMI-Complete) as described previously (Nocka et al., 1990a). Nonadherent cells were harvested by centrifugation and refed weekly and maintained at a cell density of 95% mast cells

351

E.J. Huang et al. and were used for proliferation assay. Supematants from transfected COS-1 cells were collected from 48 to 72 h after transfection. The biological activity of soluble KL in the supematants was assessed by culturing bone marrow-derived mast cells (BMMCs) with different dilutions of COS-1 cell supematants in the absence of interleuken-3 (IL-3). BMMCs were washed three times with complete RPMI and grown in 0.2% IL-3. The following day cells were harvested and suspended in complete RPMI (minus IL-3), and 104 BMMCs in 100 MAI/ well were seeded in a 96-well plate. Equal volume of diluted supernatant was added to each well, and cultures were incubated for 24 h at 37°C, 2.5 uCi of [3H]thymidine/well was then added, and incubation was continued for another 6 h. Cells were harvested on glass fiber filters (GF/C Whatman, Hillsboro, Oregon), and thymidine incorporation was determined in a scintillation counter. Assays were performed in triplicate, and the mean value is shown. Standard deviations of measurements typically did not exceed 10% of the mean values.

GGGACTATCTGCAGCCGCTGCTGGTGCAATATGCTGGAGCTCCAGAACAGCTAAACGGAG

60

TCGCCACACCGCTGCCTGGGCTGGATCCAGCGCTGCCTTTCCTTATGAAGAAGACACAA

120 5

X

X

X

T

Q

SP

ACTTGGATTASCACTTGQAT T W

I

I

T

I

C

_TTACTCC mATTACCSCTTGTCAAAACC Y

L

L

Q

L

L

F

N

P

L

V R

T

AAGGAGATCTGCGGGAATCCTGTGACTGATAATGTAAAAGACATGACAAAACTGGTGGCA R E

I

C

G

N

P

V T D

N

V

K

D

I

T

R L

V

4S

A

AATCTTCCAAATGACTATATGATAACCCTCAACTATGTCGCCGGGATGGATGTTTTGCCT N

P

L

N

Y R

D

I

T

N

L

Y

V

A

G M

D

V

L

300

6S

P

AGTCATTGTTGGCTACGAGATATGGTAATACAATTATCACTCAGCTTGACTACTCTTCTG S

C

H

W

L

D N

R

Q

I

V

L

S

L

S

L

T

T

L

360 8S 420

L

GACAAGTTCTCAAATATTTCTGAAGGCTTGAGTAATTACTCCATCATAGACAAACTTGGG D

R F

S

N

I

S

Z

G

L

S

N

Y

S

I

I

D

R L G

l05 480

AAAATAGTGGATGACCTCGTGTTATGCATGGAAGAAAACGCACCGAAGAATATAAAAGAA K

I

V

D

D

L

V

L

C N Z

Z

N

A

P R

N

R E

I

125

TCTCCGAAGAGGCCAGAAACTAGATCCTTTACTCCTGAAGAATTCTTTAGTATTTTCAAT T

P EZ

Y V

A

FF S I F N AGATCCATTGATGCCTTTAAGGACTTTATGGTGGCATCTGACACTAGTGACTGTGTGCTC S

P

X

R

P

Z T

R

R

S

I

D

A F X

D

S

F

F

S

D T

S

D

C

V

180 25 240

540 14S 600

16S

L

V

TCTTCAACATTAGGTCCCGAGAAAGATTCCAGAGTCAGTGTCACAAAACCATTTATGTTA

RESULTS Alternatively Spliced KL Transcript Encodes a Truncated Transmembrane Form of KL and the Steel Dickie Allele a C-Terminally Truncated Secreted Form of KL A cDNA clone, which had been isolated from a mouse 3T3 fibroblast library and contained most of the KL coding sequences (267 amino acids), has been described previously (Nocka et al., 1990b). To obtain the complete cDNA sequences corresponding to the 6.5 kb KL mRNA, a plasmid cDNA library was constructed by using polyA+ RNA from Balb/c 3T3 fibroblasts. The nucleotide sequence of a clone that contains the complete KL coding sequences shown in Figure 1 is in agreement with the previously published sequences, except for a single base change at position 664 that results in the substitution of serine 205 to alanine (Anderson et al., 1990; Nocka et al., 1990b). To identify altematively spliced KL RNA transcripts in RNA from tissues and cell lines, we used the RTPCR method employing primers that corresponded to the 5' and 3' ends of the coding region of the KL cDNA. RT-PCR reaction products from spleen, testis, and lung contained two fragments of -870 and 750 bp in size. DNA sequence analysis indicated that the larger PCR product corresponds to the known KL cDNA sequence, subsequently referred to as KL-1. In the smaller PCR product, a segment of 84 nucleotides of the KL coding sequences was lacking, generating an inframe deletion. The deletion end points corresponded to exon boundaries in the rat and the human KL genes (Martin et al., 1990). Therefore, the smaller PCR product appeared to correspond to an alternatively spliced KL RNA transcripts, designated KL-2. The exon missing in KL-2 precedes the transmembrane domain; it contains one of the four N-linked glycosylation sites and includes the known C-terminus (Ala-166 and Ala-167) of the soluble form of KL (55) (Figure 1). To investigate the molecular basis of the 51d allele, similar to Flanagan et al. (1991), we characterized the KL coding sequences in this allele by using PCR cloning technology. We derived primary embryo fibroblasts 352

S

T

S

L

Z3

P

G

D

R

S

V

S

V

T

K

P

660 185

M L

F

v V CCCCCTGTTGCAGCCAGCTCCCTTAGGAATGACAGCAGTAGCAGTAATAGGAAAGCCGCA 720 P

V

P

AIAT5

S

L

N

R

D

S

S

S

S

N

R

A

K

20S

A

31

-CTC_C_GAC_---

AAGGCCCCTGAAGACTCGGGCCTACAATTGACAGCCATGGCATTGCCGGCTCTCCATTTCG 780 E D

P

K

A

a

* L

S

G

L

Q W T A N A

L

P

A

L

I

22S

S

51a

-

_ TXS

AGGGCAGTTGAAAATATACAGATTAATGAAGAGGATAATGAGATAAGTATGTTGCAACAG

840 245 900

E Z

265

CTTGTAATTGGCTTTGCTTTTGGAGCCTTATACTGGAAGAAGAAACAGTCAAGTCTTACA L R

V A

I

G

V E

F

F

A

Q

I

N

G

I

A

L

N

Y

----_

W

D

K N

K

K

Z

Q

I 'S

S

L

N L _

S

C

S

T

Q Q

aTTOACA Sid

AAAGAGAGAGAATTTCAAGAGGTGTAATTGTGGACGTATCAACATTGTTACCTITCGCACA 960 X

Z

R

Z

F

Z

Q

V

273

-

GA

TCCCATT'r

GTGGCTGGTAACAGTTCATGTTTGCTTCATAAATGAAGCAGCCTTAAACAAATI

Si' 1020

CTGTCTCAAGTGACAGACCTCATCCTTACC GSTCTTGCTACCCGTGACCTTGITGTGGAT 1080 TGAACCCTGCACTGAATTATCA TCTGTAA 1140 1200 1260 CACCCAGCTTTGTTCTTCAGTCACAACCTGCAGCCTGTCGTTAATTATGGTCTI CTGCAAG 1320 1344 TAGATTTCAGCCTGGATGGTGGGG

GATTCAGTTGTTGGAGCAGAGSSCI

AGAAAAATCTGCACGGAGCAGGACTCTGGAGGTGCAAGTGATGATAGGGA4 CAAGAAC ATGTGTCCAGTCTACTTGCACCGTTTGCATGGCGAAA GTCTGAGTGCTCGAAAACC

I

m proteolytic cleavage sites

-

.. 1743

,

u

1 89N ::. 2 02-~

%-*#

> - -205

- -41 I

I

I

I

cell membrane

KL-1

KL-2

KL-Sl

KL-S2

Sld

Figure 1. Nucleotide and predicted amino acid sequence of KL-1, KL-2, and KL-Sld cDNAs. (Top) The nucleotide sequence of the KL cDNA obtained from the Balb3T3 cell plasmid cDNA library is shown. The RT-PCR products from different tissues and Sld/+ total RNA, KL-1, KL-2, and KL-Sld, were subcloned and subjected to sequence analysis. Open triangles indicate the 5' and 3' boundaries of the exon that is spliced out in KL-2; the closed triangles indicate the deletion end points in the Sld cDNA. The 67 nucleotide insert sequence of the Sld cDNA is shown above the KL cDNA sequence. Arrows indicate the putative proteolytic cleavage sites in the extracellular region of KL-1. The signal peptide (SP) and transmembrane segment (TMS) are indicated with overlying lines. (Bottom) Schematic representation of the topological characteristics of various KL protein products. Molecular Biology of the Cell

-n_.~

Cell Associated Form of KL is Absent in Sld Allele

from an Sld/+ embryo, prepared RNA from the Sld/+ embryo fibroblasts, and used different primers to amplify the Sld KL coding region, paying attention to the possibility that Sld is a deletion mutation. By using primers 1 + 3 and 1 + 4, faster migrating DNA fragments were obtained that presumably corresponded to cDNA products of the Sld allele (unpublished observations). DNA sequence analysis of the KL_Sjd cDNA segments indicated that nucleotides 660-902 of the wild-type sequence is deleted; instead, a sequence of 67 bp was found to be inserted (Figure 1). The predicted amino acid sequence of KL-Sld cDNA consists of amino acids 1-205 of the known KL sequence plus three additional amino acids and thus includes all of the sequences in the soluble form of KL, whereas the transmembrane and the cytoplasmic domains of wild-type KL-1 are deleted. KL-2 is Expressed in a Tissue-Specific Manner The alternatively spliced transcript KL-2 had been detected in spleen, testis, and lung RNA but not in fibroblasts and brain RNA, suggesting that the expression of KL-2 may be controlled in a tissue-specific manner. To address this question in more detail, we determined the steady-state levels of KL-1 and KL-2 RNA transcripts in RNA from a wide variety of tissues by using an RNase protection assay. pcDNAI plasmid containing the KL1 cDNA was linearized with Spe I to generate an RNA hybridization probe of 625 nucleotides by using SP6 RNA polymerase. The probe was hybridized with 20 ,ug of total RNA from Balb/c 3T3 fibroblasts, brain, spleen, and testis of a 40 d old mouse, as well as from brain, bone marrow, cerebellum, heart, lung, liver, spleen, and kidney of an adult mouse and placenta (14 d p.c.). The samples then were digested with RNase and the reaction products were analyzed by electrophoresis in a 4% urea/polyacrylamide gel. In these experiments, KL-1 mRNA protected a single fragment of 575 bases, whereas KL-2 mRNA protected fragments of 449 and 42 nucleotides. As shown in Figure 2 in Balb/c 3T3 fibroblasts, KL-1 is the predominant transcript, whereas the KL-2 is barely detectable. In brain and thymus, KL-1 is the predominant transcript, but in spleen, testis, placenta, heart, and cerebellum both KL1 and KL-2 transcripts are seen in variable ratios. The ratio of the KL-1 to KL-2 in tissues determined by densitometry in brain is 26:1, in bone marrow 3:1, in spleen 1.5:1, and in testis (40 d p.n.) 1:2.6. These results suggest that the expression of KL-1 and KL-2 is regulated in a tissue-specific manner.

Biosynthetic Characteristics of KL Protein Products in COS Cells Although KL was purified from conditioned medium of Balb/c 3T3 cells and is a soluble protein, the predicted amino acid sequences for KL-1 and KL-2 suggest that these proteins are membrane associated. To investigate Vol. 3, March 1992

- 'A

.'A,*--;.-.,

I U

Testis Liver Kidney Placenta

I

.I

-

3T3 fibroblasts Spleen Brain

I

Thymus BM Lung

I

Heart Cerebellum Control

vjw

,

4~k

(0) :3

4

01 01

Figure 2. Identification of KL-1 and KL-2 transcripts in different tissues by RNase protection assays. 32P-labeled antisense riboprobe (625 nt.) was hybridized with 20 ,ug total cell RNA from tissues and fibroblasts except for lung and heart where 10 /g was used. On RNase digestion, reaction mixtures were analyzed by electrophoresis in a 4% polyacrylamide/urea gel. For KL-1 and KL-2, protected fragments of 575 and 449 nt. are obtained, respectively. Autoradiographic exposures were for 48 or 72 h, except for the 3T3 fibroblast RNA, which was for 6 h.

the relationship of KL-S with the KL-1 and KL-2 protein products, we determined their biosynthetic characteristics. The KL-1 and KL-2 cDNAs, prepared by RT-PCR, were subcloned into the HindIll and Xba I sites of the expression vectors pcDNAI or pCDM8 for transient expression in COS-1 cells. To facilitate transient expression of the KL-1 and KL-2 protein products, COS-1 cells were transfected with the KL-1 and KL-2 plasmids by using the DEAE-dextran/DMSO protocol. KL protein synthesis in the COS-1 cells was shown to be maximal 353

E.J. Huang et al.

between 72 and 96 h subsequent to the transfection. To determine the biosynthetic characteristics of the KL1 and KL-2 proteins, pulse-chase experiments were performed. Cells were labeled with 35S-methionine (0.5 mCi/ml) for 30' and then chased with regular medium. The cell lysate and supernatants then were collected at the indicated times and processed for immunoprecipitation with anti-KL antiserum, prepared by immunizing rabbits with purified murine KL, and analyzed by SDSPAGE (12%). In cells transfected with the KL-1 plasmid, at the end of the labeling period, KL specific protein products of 24, 35, 40, and 45 kDa are found (Figure 3). These proteins presumably represent the primary translation product and processed KL protein products that are progressively modified by glycosylation. Increasingly longer chase times reveal the 45-kDa form as the mature KL protein product, and it is quite likely that this protein represents the cell membrane form of KL. In the supernatant beginning at 30', a time-dependent increase in the 28-kDa KL protein product is seen. Two minor products of 38 and 24 kDa were also found with increasing time. A pulse-chase experiment of COS-1 cells transfected with the KL-2 plasmid is shown in Figure 3. The KL-2 protein products are processed to produce products of 32 and 28 kDa that likely include the presumed cell membrane form of KL-2. As one might have predicted, the cell membrane form of KL-2 is more stable than the corresponding KL-1 protein with a half life of >5 h. In the cell supernatant, after 3 h a soluble form of KL-2 of -20 kDa is seen. The appearance and accumulation of the soluble form of KL-2 in the cell supernatant are delayed compared with that of KL-1, in agreement with less efficient proteolytic processing of the KL-2 protein product. A 38-kDa KL-1 protein product seen in the supernatant may represent a cleavage product that involves a cleavage site near the transmembrane domain (Figure 1).

Proteolytic Processing of KL-1 and KL-2 in COS Cells is Modulated by PMA and the Calcium Ionophore A23187 The protein kinase C inducer PMA is known to facilitate proteolytic cleavage of cell membrane associated receptors and growth factors, such as the colony-stimulating factor 1 (CSF-1) and the c-kit receptors and CSF-1 and transforming growth factor-a, to produce soluble forms of the extra cellular domain of these proteins (Downing et al., 1989; Pandiella and Massague, 1991; Stein and Rettenmier, 1991; Yee and Besmer, unpublished data). We have determined the effect of PMA treatment on the biosynthetic characteristics of KL-1 and KL-2 in COS-1 cells. The pulse-chase experiments shown in Figures 4B and 5 indicate that PMA induces the rapid cleavage of both KL-1 and KL-2 with no difference in time course of processing and that the released KL-1 354

kD 47 -

33 24 16-

_.,;j.:Z cell lysate 0 0.5 1 3 5

cell Iysate

kD 47-

0 0.5 1 ...

b

33 24-

3 5

supernatant 0 0.5 1 3 5

do**

-.

KL-1

supernatant ...

0 0.5 1

3 5

..

_?; __. .F

.4

KL-2

.....

8#rY t.

16-

cell lysate 0 1 3 6

supernatant 0 1 3

66 45 31 -

COS-1

21 Figure 3. Biosynthetic characteristics of KL-1 and KL-2 protein products in COS cells. COS-1 cells were transfected with 5 Ag of the KL-1 and KL-2 expression plasmids using the DEAE-dextran method. After 72 h the cells were labeled with 3 S-Met for 30 min and then chased with complete medium for 0.5, 1, 3, and 5 h. Supernatants and cell lysates were immunoprecipitated with anti-KL rabbit serum. Immunoprecipitates were analyzed by SDS-PAGE (12%). In the lower panel, results from nontransfected COS-1 cells are shown. Migration of molecular mass markers is indicated in kilodaltons.

and KL-2 protein products are indistinguishable in size from those obtained in the absence of inducer. To establish a role for protein kinase C in PMA-induced proteolytic cleavage of the KL-1 protein product, cells were treated with the protein kinase C inhibitor calphostin C before PMA induction. Under these conditions, PMAinduced cleavage of KL-1 was inhibited completely (Figure 6). These results suggest that the proteolytic cleavage machinery for both KL-1 and KL-2 is activated similarly by PMA. On one hand this may mean that two distinct proteases, specific for KL-1 and KL-2, respectively, are activated by PMA, or alternatively, that there is one protease that is activated to a very high level that cleaves both KL-1 and KL-2 but with different rates. We have also determined the effect of the calcium ionophore A23187 on KL cleavage. Both KL-1 and KLMolecular Biology of the Cell

Cell Associated Form of KL is Absent in SId Allele

A

KL-1 cell lysate II supernatant 0 30' 45'lhr3hr 0 30' 45' lhr3hr -

_

kD 126 80 49 32 27-

KL-2 cell supernatant lysate 0 0

-

O_",

I,

-

0 30' 45'1 hr 3hr 0 30' 45' 1 hr 3hr ,-OM one! PM,. 0-m- W.".

kD 12680 49 -

1.

32-U 27-

18 -

18 -

B

KL-1 cell lysate supernatant 0 30' 45'l hr 3hr 0 30' 45' l hr 3hr kD 9766 45 -

lUS

r

an

.

;

cell lysate

31 -

21 14 -

21 14 -

C

cell lysate

KL-1

....

,

:.

=

.

w1ww r__

31 -

4.,w

wt*w,.-:

supernatant

0 30' 45'1hr3hr 0 30' 45' 1hr 3hr kD PR $pq, 97 - -t,..+. 66 45 p ao

cell lysate

IETw,

v

21 14 -

21 14 -

KL-2

"i

supernatant

0 30' 45'1lhr 3hr 0 30' 45' 1 hr 3hr inw w W

kD 97 66 4531-

supernatant

0 30' 45' 1 hr 3hr 0 30' 45' 1 hr 3hr P, " rFi

kD 9766 45 -

31 -

KL-2

,,.

9

WI -

_

I,

-

im .-1

Figure 4. PMA-induced cleavage of the KL-1 and KL-2 protein products. COS-1 cells were transfected with 5 pg of the KL-1 and KL-2 expression plasmids, and after 72 h the cells were labeled with 35S-Met for 30 min and then chased with medium (A) in the absence of serum; (B) containing the phorbol ester PMA (1 pM); and (C) containing the calcium ionophore A23187 (1 pM). Supematants and cell lysates were immunoprecipitated with anti-KL rabbit serum. Immunoprecipitates were analyzed by SDS-PAGE (12%). Migration of molecular mass markers is indicated in kilodaltons. Vol. 3, March 1992

355

E.J. Huang et al.

A. KL-1

0.5

1

and cell lysates and supematants prepared at 120 min from the start of the chase. Quantitation of the immunoprecipitation analysis shown in Figure 9 indicates that the rates of PMA-induced proteolytic cleavage of the cell membrane forms of KL-1 and KL-2 are similar to those shown in Figures 4 and 5, suggesting that PMAinduced cleavage of these molecules may occur on the cell surface.

B. KL-2

2 1.5 2.5 Time (Hour)

3

3.5

0

0.5

1

1.5

2

2.5

3

3.5

Time (Hour)

Figure 5. Quantitative representation of PMA- and A23187-induced cleavage of KL-1 and KL-2 protein products. The 28-kDa KL-1 (A) and the 20-kDa KL-2 (B) supematant product bands of the SDS/ PAGE analysis shown in Figure 4 were quantitated by using a phospholmage analyzer (Molecular Dynamics). 0, serum free medium; 10% fetal bovine serum; A, 1 ,M PMA; *, 1 ,uM A23187. L,

2 cleavage is accelerated by this reagent, although not effectively as with PMA, indicating that mechanisms that do not involve or bypass the activation of protein kinase C can mediate proteolytic cleavage of both KL1 and KL-2 (Figs. 4C and 5). Furthermore, a comparison of the PMA, A23187, and serum induced KL-1 and KL2 cleavage products indicates that they are indistinguishable in size (Figure 7).

Biological Activity of the Released KL Protein Products To test the biological activity of the released KL protein products, the supematants of transfected COS-1 cells were collected 72 h after transfection and assayed for activity in the mast cell proliferation assay. BMMCs were incubated for 24 h with different dilutions of the collected supematants and assayed for 3H-thymidine incorporation as described previously (Figure 10). Super-

A.

CL

as

Determination of Cell Surface Expression of KL Protein Products To determine whether the membrane-associated forms of the KL-1 and KL-2 protein products are expressed on the cell surface and whether proteolytic processing occurs while on the cell surface, COS-1 cells expressing the KL-1 and KL-2 protein products were investigated by flow cytometry and cell surface labeling procedures. Flow cytometry shown in Figure 8A suggests that -1020% of the transfected COS cells expressing either KL1 or KL-2 protein products express KL determinants on their cell surface, and this is in agreement with the results of Flanagan et al. (1991). Furthermore, on labeling of cell surface proteins with 1251I-iodine, immunoprecipitation analysis shown in Figure 8B indicated that the 45-kDa KL-1 and the 32-kDa KL-2 protein species are the main labeled proteins, and consequently, these proteins are expressed on the cell surface. In pulse-chase experiments, within 60 min of the labeling period both the KL-1 and KL-2 protein products have been processed to the cell surface (Figure 3). To determine whether the soluble KL-1 and KL-2 protein products are being generated from the cell surface forms of these proteins, COS cells expressing the KL-1 and KL-2 cDNAs were pulse labeled for 30 min. After 60 min of chase, the medium was changed and PMA added to cultures at 60 (60), 75 (45), 90 (30), and 105 (15) min 356

Sup

O' 15' 30' 45' 60'

15' 30' 45' 60'

WI

97 68

46-

_

g t 21

_m

"..

-

B.

CL 0' 15' 30'

Sup

I4 45'

60'

1 0'

1 5'

30' 45' 60'

97 - &LA ::" 68 46

-

30-

21-

14

-

Figure 6. Inhibition of PMA-induced KL-1 cleavage by the protein kinase C inhibitor calphostin C. COS-1 cells transfected with the KL1 and KL-2 expression plasmids were labeled with 35S-Met, incubated with calphostin (1 MM), and then treated with PMA in the presence of calphostin and cell lysates (CL) and supematants (Sup) processed for immunoprecipitation analysis. (A) No calphostin; (B) 1 ,uM calphostin. Migration of molecular mass markers is indicated in kilodaltons.

Molecular Biology of the Cell

Cell Associated Form of KL is Absent in Sld Allele

A.

Sup.

CL I

LL

.

XO

LL

00

9769 46 -

31

;.

-

..

:*Lai

-4-

1

-

1

21 -

B.

CL I

Sup. CO1l ---m

LL --,o

0,

0

v-

97-

P. co CII) C14