Laboratoire Franco-Luxembourgeois de Recherche Biomédicale (CRP-Santé/CNRS), University Center, L-1511 Luxembourg, Grand-Duchy of Luxembourg

Correspondence to: N Kieffer, Laboratoire Franco-Luxembourgeois de Recherche Biomédicale, Centre Universitaire, 162A, avenue de la Faïencerie, L-1150 Luxembourg

We have shown previously that integrin-linked kinase (ILK) is upregulated in human HT-144 melanoma cells following TGF-1 stimulation. Using mRNA from TGF-1 stimulated HT-144 cells and reverse transcriptase polymerase chain reaction, we have isolated a cDNA encoding a protein highly homologous to ILK. Sequencing of the full-length 1359 base pair cDNA and polypeptide translation revealed that this protein, designated ILK-2, differs from the known ILK (hereafter called ILK-1) by only four amino acids, while the cDNA sequence diverges by 102 nucleotides, thus excluding that ILK-2 is an allelic variant of ILK-1. Expression of ILK-2 mRNA was observed in metastatic human HT-144 melanoma and HT-1080 fibrosarcoma cell lines, but not in normal human tissues. Moreover, stimulation of HT-144 cells with TGF-1, but not with EGF, PDGF-AB or insulin, induced a selective overexpression of ILK-2 mRNA as compared to ILK-1 mRNA. Bacterially-expressed GST/ILK-2 autophosphorylated and labeled myelin basic protein as well as a recombinant GST/3 integrin cytoplasmic tail peptide. Transfection of either ILK-2 or ILK-1 cDNA into the non-metastatic melanoma cell line SK-Mel-2, expressing exclusively ILK-1, induced anchorage independent cell growth and cell proliferation, as demonstrated by growth in soft agar. Our data provide evidence that ILK-2 is a new isoform of ILK-1 that is expressed in some highly invasive tumor cell lines but not in normal adult human tissues and whose expression is regulated by TGF-1. Oncogene (2000) 19, 3069-3077

integrin-linked kinase; integrin; TGF-; cDNA sequence; melanoma phosphorylation

Introduction

Tumor cell metastasis is often associated with increased production of growth factors that act as autocrine or paracrine mediators involved in many aspects of cancer cell biology, such as malignant cell transformation, tumor cell adhesive interactions (Matsumoto et al., 1995), extracellular matrix and fibrous stroma formation (Gregoire and Lieubeau, 1995), matrix metalloproteinase secretion, as well as angiogenesis (Liotta et al., 1991). Among the numerous growth factors identified, transforming growth factor s (TGF-s) are of particular interest, since biologically active TGF-s control cell proliferation and differentiation, promote cell motility and angiogenesis, stimulate the expression of integrins and matrix proteins, and mediate epithelial to mesenchymal transition during embryogenesis and wound healing (Alevizopoulos and Mermod, 1997). However, the intracellular machinery that drives these TGF- dependent processes during tumor cell metastasis remains largely unknown.

Integrin-linked kinase (ILK) is a newly identified serine-threonine kinase that interacts with the cytoplasmic tail of 1, 2 and 3 integrin subunits (Hannigan et al., 1996), and plays a central role in many of the cellular aspects involved in malignant transformation, by acting as a receptor-proximal effector in the crossmodulation between growth factor response and integrin signaling (Delcommenne et al., 1998). ILK has been shown to regulate cell survival, proliferation and differentiation by modulating integrin-mediated cell adhesion, E-cadherin expression, fibronectin matrix assembly, as well as cell adhesion-dependent cell cycle progression (reviewed in Dedhar et al., 1999 and Huang and Wu (1999)). Overexpression of ILK in rat epithelial cells activates the lymphocyte enhancer binding factor 1 (Lef-1)/-catenin signaling pathway (Novak et al., 1998), inducing anchorage-independent cell growth (Hannigan et al., 1996) and oncogenic transformation (Wu et al., 1998). The kinase activity of ILK is regulated by cell adhesion and by insulin in a phosphoinositide-3-OH kinase-dependent manner (Delcommenne et al., 1998), and downstream targets for ILK trans-phosphorylation comprise protein kinase B (PKB/AKT), and glycogen synthase kinase 3 (GSK-3) (Delcommenne et al., 1998; Lynch et al., 1999). Evidence has also been provided that the LIM-only protein PINCH (Rearden, 1994) is an ILK binding protein, suggesting that this adapter protein might connect ILK and integrins with components of growth factor receptor kinases and small GTPase signaling pathways (Tu et al., 1999; Li et al., 1999).

We have recently shown that stimulation of the human melanoma cell line HT-144 with biologically active TGF-1 upregulates integrin and ILK expression and induces translocation of -catenin from the cytoplasm to the nucleus (Janji et al., 1999). We now report the identification and cloning of a new isoform of ILK, designated ILK-2, in TGF-1 stimulated HT-144 melanoma cells. We provide evidence that ILK-2, which has antigenic cross-reactivity and functional properties similar to the known ILK-1, is not expressed in normal adult tissues, but is expressed in two highly metastatic melanoma and fibrosarcoma tumor cell lines, suggesting that this isoform might play a role in the metastatic process of these cell lines.

Results

Nucleotide and predicted amino acid sequence of a new isoform of ILK

We have recently shown an increased ILK immunoreactivity in the metastatic melanoma cell line HT-144 following TGF-1 stimulation (Janji et al., 1999). In order to determine whether this upregulated HT-144 ILK antigen corresponded to the published ILK protein, we amplified the full-length ILK cDNA from TGF-1 stimulated HT-144 cell mRNA, using ILK specific primers corresponding to the coding 5'- and 3'-end sequence and reverse transcriptase-PCR (RT-PCR). The identity of the cloned 1359 bp PCR product with human ILK primers was verified by restriction fragment analysis. Interestingly, analysis of several subcloned full-length ILK cDNAs revealed for some the absence of a BamHI restriction site, present at position 814 of the published ILK cDNA sequence (Hannigan et al., 1996). Sequencing of the BamHI-negative full-length cDNA revealed only 93% homology with the published ILK sequence, indicating that the isolated HT-144-ILK cDNA codes for an isoform, rather than an allelic variant of ILK. In contrast, sequencing of the BamHI-positive full-length ILK cDNA revealed complete identity with the published sequence, except for three additional nucleotides (position 1231-1233) coding for a proline residue, not present in the initially published human ILK sequence (Hannigan et al., 1996), but also identified in mouse ILK (Li et al., 1997). Comparison of the BamHI-negative HT-144 ILK cDNA sequence, designated hereafter ILK-2, with human ILK (called ILK-1) and mouse ILK further revealed a closer homology of ILK-2 with mouse ILK (96%) than human ILK-1, suggesting that mouse ILK is the counterpart of human ILK-2 (Figure 1). At the amino acid level, the deduced sequence of ILK-2 diverged from ILK-1 by only four amino acids, at positions 197 (T-A), 214 (G-S), 259 (S-A) and 436 (P-S), and by three amino acids from mouse ILK, at positions 214 (G-S), 412 (V-I) and 436 (P-S). Interestingly, all of these substitutions are located in the kinase subdomains of ILK-2 (Figure 1).

Human tumor cell and normal tissue distribution of ILK-2 as compared to ILK-1

Since Northern blot analysis of TGF-1 stimulated HT-144 cell mRNA with an ILK-2 probe revealed a single 1.8 kb transcript identical to that observed with the ILK-1 probe (data not shown), we used an alternative approach to investigate ILK-2 versus ILK-1 expression in various cell types. Comparative sequence analysis of the ILK-1 and ILK-2 cDNA revealed the presence of a unique HincII restriction site at position 409 of the ILK-2 cDNA (see Figure 1), giving rise to two fragments of 411 and 948 bp, as well as a unique BamHI restriction site in the ILK-1 cDNA, generating two fragments of 536 and 823 bp, allowing for RFLP (Restriction Fragment Length Polymorphism) analysis of RT-PCR amplified cDNA. As shown in Figure 2, HT-144 melanoma cells expressed ILK-1 and ILK-2, while HT-1080 fibrosarcoma cells expressed only ILK-2. SK-Mel-1 and SK-Mel-2 melanoma, PC3 prostate carcinoma cell lines and normal primary fibroblasts expressed only ILK-1. We next analysed ILK-1 and ILK-2 mRNA expression in a series of normal human tissues using the RT-PCR/RFLP approach. Interestingly, all normal tissues tested were positive for ILK-1, but were negative for ILK-2 mRNA (Figure 3). Based on these results we conclude that ILK-2 is not expressed in normal adult tissues, but is expressed in some highly invasive tumor cell lines. Western blot analysis performed with a polyclonal anti-ILK antibody revealed a single band of 59 kDa in human tumor cell lines shown to express either ILK-1, ILK-2 or both, thus demonstrating the high homology between ILK-1 and ILK-2 protein and their antigenic cross-reactivity (Figure 4).

ILK-2- but not ILK-1- mRNA is upregulated in HT-144 melanoma cells following TGF-1 stimulation

To investigate the effect of TGF-1 stimulation on ILK-1 and ILK-2 expression in HT-144 melanoma cells, the cells were grown in the absence of serum and treated for 24 h with 10 ng of recombinant TGF-1. A typical result is shown in Figure 5a. TGF-1 induced overexpression of the ILK antigen, that correlated with complete down-regulation of E-cadherin expression. To determine which ILK isoform was upregulated following TGF-1 stimulation, we used semi-quantitative RT-PCR followed by RFLP. As shown in Figure 5b, when RT-PCR was performed for only 25 cycles using mRNA from non-stimulated HT-144 cells, no PCR products corresponding to ILK-1 or ILK-2 could be detected. However, in HT-144 cells stimulated with 10 ng of TGF-1 for 24 h, a 1359 bp band corresponding exclusively to ILK-2 was detected. Under the same experimental conditions no change in the level of -actin mRNA expression was observed. These results demonstrate a selective upregulation of ILK-2 mRNA following TGF-1 stimulation, suggesting that the ILK-1 and ILK-2 promoters are functionally distinct. Finally, to determine whether other growth factors, hormones or mitogens were able to induce ILK-2 expression, the same experiment was performed by stimulating the cells for 24 and 48 h with EGF (2 ng/ml), PDGF-AB (100 ng/ml) as well as insulin (5 g/ml). As shown in Figure 6, none of these agents were able to upregulate ILK expression.

Kinase activity of ILK-2

Although ILK-2 differs from ILK-1 by only four amino acids, the location of the four substitutions in the kinase subdomains prompted us to investigate the kinase activity of ILK-2. For this purpose, ILK-1 and ILK-2 were expressed in bacteria as gluthatione-S-transferase (GST) fusion proteins. Western blot analysis performed with a polyclonal anti-GST antibody identified a specific band of 88 kDa in the bacterial lysates corresponding to GST/ILK-1 and GST/ILK-2, as well as a band of 29 kDa corresponding to free GST (Figure 7). For in vitro kinase activity studies, ILK-1 and ILK-2 fusion proteins were purified by glutathione affinity chromatography and incubated with [-³²P]ATP in the presence or absence of the exogenous substrate, myelin basic protein (MBP), or the integrin 3 cytoplasmic tail fusion protein (GST/3_cyto) (Vallar et al., 1999). As shown in Figure 8a, GST/ILK-2 was autophosphorylated and was also able to phosphorylate MBP. Furthermore, both GST/ILK-1 and GST/ILK-2 phosphorylated the cytoplasmic tail of the 3 integrin subunit. When the kinase reaction was performed in the presence of [-³²P]ATP, used as a negative control, no GST/ILK-2 dependent auto- or trans-phosphorylation could be observed (Figure 8b).

Transfection of ILK-2 or ILK-1 into SK-Mel-2 melanoma cells induces anchorage-independent cell growth

In order to compare the in vivo functional activity of ILK-2 versus ILK-1, we overexpressed either ILK-2 or ILK-1 in SK-Mel-2 cells shown to express only ILK-1. As previously reported, these SK-Mel-2 cells are tumorigenic, but not metastatic when injected into nude mice, and exhibit anchorage dependent cell growth in vitro (Gouon et al., 1996). Two cell clones were selected, SK-Mel-2/9 overexpressing ILK-2 and SK-Mel-2/8 overexpressing ILK-1. RT-PCR/RFLP analysis of mRNA isolated from the SK-Mel-2/9(ILK-2) cell clone revealed the presence of mRNA encoding the transfected ILK-2 protein, not found in parental SK-Mel-2 cells (Figure 9a). Successful transfection of recombinant ILK-1 cDNA in the SK-Mel-2/8(ILK-1) cell clone was demonstrated by PCR amplification of genomic ILK-1 DNA using vector specific primers (Figure 9a).

To compare the effect of ILK-1 or ILK-2 expression on SK-Mel-2 cell growth, the proliferative capacity of parental SK-Mel-2 cells as well as the SK-Mel-2/9 (ILK-2) and SK-Mel-2/8(ILK-1) clones was tested in a soft agar assay. As shown in Figure 9b, 3 weeks following cell seeding, both SK-Mel-2/9(ILK-2) and SK-Mel-2/8(ILK-1) cells formed large colonies, in contrast to SK-Mel-2 parental cells that underwent only one or two divisions prior to cell death, while mock-transfected SK-Mel-2 cells exhibited a phenotype identical to parental SK-Mel-2 cells (data not shown).

Discussion

In this study we have identified a novel human ILK protein kinase, called ILK-2, which shares 99% sequence homology at the amino acid level with the published human ILK (Hannigan et al., 1996), called here ILK-1, but only 93% sequence homology at the cDNA level, providing evidence that this new ILK-2 protein is an isoform, rather than an allelic variant of the known ILK-1. Interestingly, the cDNA sequence of human ILK-2 revealed a closer homology to mouse ILK (Li et al., 1997) than human ILK-1, suggesting that mouse ILK is in fact the human ILK-2 counterpart.

Fluorescence in situ hybridization (FISH) analysis of metaphase spreads from karyotypically normal human lymphocytes as well as cell lines harboring well characterized reciprocal translocations allowed the identification of a single locus for the human ILK-1 gene on chromosome 11, between the HBBC and CALC loci in the 11p15.5-p15.4 band interval, a part of chromosome 11 strongly associated with tumorigenesis (Hannigan et al., 1997). In contrast, two mouse chromosome regions, region E on chromosome 7, sharing contiguous sequences with human chromosome 11, and region E1-E3 on chromosome 9, were identified with the mouse homolog of human ILK, supporting the hypothesis of the existence in the mouse genome of a second ILK homolog gene or of an ILK pseudogene (Li et al., 1997). Our preliminary data on the chromosomal localization of the human ILK-2 gene by FISH analysis have identified a single locus overlapping with the 11p15.5-p15.4 region. Since this position has previously also been determined for the ILK-1 gene (Hannigan et al., 1997) we conclude that the ILK-2 gene is located in close proximity to the ILK-1 gene.

Integrins, in contrast to many signaling receptors, are devoid of any endogenous enzymatic kinase activity, and rely on intracellular kinases such as ILK to activate signaling pathways involving phosphorylation processes. A mutation of the single invariant amino acid E359K in the kinase subdomain of ILK results in an inactive form of ILK, exhibiting a dominant negative function (Delcommenne et al., 1998). Since the four amino acid substitutions that differentiate ILK-2 from ILK-1 are all located in the kinase subdomains, we investigated the kinase activity of GST/ILK-2. Interestingly, these substitutions had no inhibitory effect on the kinase activity. Furthermore, we provide new evidence that both ILK-1 and ILK-2 are able to phosphorylate the 3 integrin cytoplasmic tail, as demonstrated by phosphorylation of the GST/3cyto fusion protein. Although our results show that ILK-2 is undistinguishable from ILK-1 with respect to molecular weight, immunoreactivity, autophosphorylation, and its capacity to phosphorylate the exogenous substrate MBP or the integrin 3 cytoplasmic tail, further studies will be necessary to determine whether the functional properties of ILK-2 are identical to ILK-1. Indeed, we cannot exclude that subtle differences, such as increased protein stability or enhanced enzymatic activity of ILK-2 could provide some explanation for the close correlation between ILK-2 expression and the malignant phenotype of tumor cells.

ILK has been shown to be expressed in a wide spectrum of human (Hannigan et al., 1996) and mouse tissues (Li et al., 1997) when studied at the mRNA level. In contrast, when exploring the ILK immunoreactivity in normal human cells and tissues, as well as various tumor cell types, Chung et al. (1998) found that ILK was expressed mainly in myocardiac cells and skeletal muscle fibers. Among the tumors tested, Ewing's sarcoma (ES), primitive neuroectodermal tumors (PNET) and medulloblastomas revealed strong ILK immunoreactivity, while other small round cell sarcomas were negative, suggesting a value for ILK as a specific diagnostic marker of tumors with primitive neural differentiation. With the selective expression of ILK-2 in some tumor cells, as shown here, and the cross-reactivity of anti-ILK antibodies with ILK-1 and ILK-2, further studies will be of interest to test which isoform of ILK is expressed in ES and PNET tumors. Also the discrepancy between ILK mRNA expression and ILK antigen expression in normal versus malignant tissues (Chung et al., 1998) could potentially be explained by a higher stability of ILK-2, or a higher affinity of the anti-ILK antibody for ILK-2.

Recent data have demonstrated that the kinase activity of ILK-1 is tightly regulated by stimuli derived from integrin dependent cell adhesion or growth factor stimulation (Delcommenne et al., 1998). In addition, ILK expression has been shown to be upregulated by the erbB-2 receptor tyrosine kinase in hyperplastic epidermis (Xie et al., 1998). Our data provide evidence for a distinct regulation of ILK-1 and ILK-2 expression by TGF-1. Indeed, stimulation of HT-144 cells with TGF-1, but not with other agents such as growth factors (EGF, PDGF-AB) or hormones (insulin), induced selective upregulation of ILK-2 expression. Our data suggest that the promoter activities for the ILK-1 and ILK-2 genes are different and that TGF-1 responsive elements must be present in the ILK-2 promoter. It will be of interest to test whether Smads, which transfer information from the TGF- receptor to the nucleus, interact with the ILK-2 promoter and regulate ILK-2 transcription (Derynck et al., 1998). TGF-1 could also be involved in regulating ILK-2 function through an indirect mechanism by downregulating the expression of the protein and lipid phosphatase PTEN (Tamura et al., 1998; Maehama and Dixon, 1998) a tumor suppressor gene (Stambolic et al., 1998; Suzuki et al., 1998) that is mutated in various tumors (Steck et al., 1997) including melanoma (Robertson et al., 1999). Indeed, recent data have provided evidence that PTEN inhibits ILK activity in human glioma cells (Morimoto et al., 2000) and that TGF- downregulates PTEN mRNA expression in keratinocytes (Li and Sun, 1997). Based on these findings, it will be of interest to determine the role of PTEN in regulating the activity and expression of ILK-2 in HT-144 melanoma cells.

Materials and methods

Materials

Sodium orthovanadate (Na₃VO₄), leupeptin, pepstatin, aprotinin, N-(trans-epoxysuccinyl)-L-leucine 4-guanidinobutylamide (E64), 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride (AEBSF), and benzamidine were obtained from Sigma (Bornem, Belgium); goat anti-rabbit IgG conjugated to horseradish peroxydase, polyclonal goat anti-glutathione S transferase (GST) IgG, [³²P]ATP and [³²P]ATP were from Amersham Pharmacia Biotech (Roosendaal, The Netherlands); the rabbit polyclonal anti-ILK (integrin-linked kinase) IgG (UB 06-550) and myelin basic protein (MBP) were purchased from Upstate Biotechnology (Veenendaal, The Netherlands); rabbit anti-goat IgG conjugated to horseradish peroxydase was from Jackson Immunoresearch (West Grove, PA, USA); total RNA isolated from various human tissues from Clontech (Leusden, The Netherlands); recombinant human TGF-1 from R&D Systems (Abingdon, UK); recombinant human insulin, platelet derived growth factor-AB (PDGF-AB) and epidermal growth factor (EGF) from Sigma. The mouse anti-human E-cadherin antibody was a generous gift of Dr M Mareel (Gent, Belgium).

Cell lines and culture

The melanoma HT-144, SK-MEL-1, and SK-Mel-2, the fibrosarcoma HT-1080 and the prostate carcinoma PC3 cell lines were purchased from the American Type Culture Collection (ATCC, Rockville, MD, USA) and grown in monolayer culture in Iscove's modified Dulbecco's medium (IMDM, BioWhittaker, Verviers, Belgium), supplemented with 10% (v/v) heat-inactivated fetal calf serum (FCS), glutamin (2 mM) and penicillin-streptomycin (100 U/ml). For human fibroblast cultures, surgical specimens of human foreskin tissue were minced finely and explanted in a 10-cm tissue culture dish. When the primary fibroblast cultures were confluent, they were passaged two to three times and either used directly, or frozen for later use. Adherent cells were routinely passaged with EDTA buffer, pH 7.4 (126 mM NaCl, 5 mM KCl, 1 mM EDTA and 50 mM HEPES). For growth factor stimulation, adherent HT-144 cells were washed twice with phosphate-buffered saline (PBS), pH 7.4 (136 mM NaCl, 2.7 mM KCl, 8 mM Na₂HPO₄ and 1.8 mM KH₂PO₄) and then cultured in serum-free IMDM for 24 or 48 h in the presence or absence of 10 ng/ml of TGF-1, 2 ng/ml of EGF, 100 ng/ml of PDGF-AB and 5 g/ml of insulin.

Reverse-Transcriptase/Polymerase Chain Reaction (RT-PCR)

Total RNA was extracted from either TGF-1 stimulated HT-144 cells or various tumor cell lines and primary cell cultures by the acid guanidinium thiocyanate-phenol-chloroform extraction method (Chomczynski and Sacchi, 1987). For RT-PCR, 2 g of total RNA was transcribed to cDNA with the RNA PCR Kit Ver.2.1 (TaKaRa Biomedicals, Verviers, Belgium) using an anti-sense primer (30 mer) corresponding to the published ILK nucleotide sequence 1359-1341 (Hannigan et al., 1996) with an additional XhoI restriction site (CTCGAG): 5'-CTCGAGCTACTTGTCCTGCATCTTCTCAAG-3'. The upstream sense primer was a 32 mer corresponding to the ILK nucleotide sequence 1-21 with an EcoRI restriction site (GAATTC) and two extra nucleotides (GT) before the start codon ATG: 5'-GAATTCGTATGGACGACATTTTCACTCAGTGC-3'. The cDNA was then amplified for 40 cycles on a Perkin Elmer GeneAmp PCR System 2400. The cycling program included a 1 min denaturation step at 95°C and a 2 min annealing step at 50°C, followed by a 3 min elongation step at 72°C. For semi-quantitative RT-PCR analysis, ILK cDNA was coamplified with the internal control -actin cDNA in a single tube for 25 cycles as previously described (Janji et al., 1999).

Restriction Fragment Length Polymorphism (RFLP) analysis of amplified ILK cDNA

For RFLP analysis, RT-PCR amplified ILK cDNA was purified using the PCR Preps DNA purification system (Promega, Madison, WI, USA), digested with BamHI or HincII, and submitted to electrophoresis in a 1% agarose gel. The bands were visualized after ethidium-bromide staining of the gel.

PCR amplification of the pcDNA neo vector containing the ILK-1 cDNA insert

Genomic DNA from cells transfected with pcDNA 3.1(+)neo vector containing the ILK-1 insert was purified using the GFX Genomic Blood DNA purification kit (Amersham Pharmacia Biotech). The ILK-1 cDNA insert was then amplified using pcDNA 3.1(+)neo sense 5'-CTAGAGAACCCACTGCTTAC-3' and anti-sense 5'-GCAAACAACAGATGGCTG-3' primers.

Cloning and sequencing of PCR-amplified ILK-2 cDNA

PCR-amplified ILK-2 cDNA was subcloned into the EcoRI-XhoI restriction site of the pcDNA3.1(+)neo vector (Invitrogen, Groningen, The Netherlands), and the construct verified by dideoxy sequencing using the T7 sequencing kit (Amersham Pharmacia Biotech).

Western blot analysis

Cells were lysed in ice-cold RIPA buffer (50 m Tris-HCl pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.1% SDS, 1% deoxycholate) containing 1 mM Na₃VO₄, 10 g/ml leupeptin, 10 g/ml pepstatin, 40 M AEBSF, 10 g/ml E64, 25 g/ml aprotinin, and 0.5 mM benzamidin. The lysates were pre-cleared and the protein concentration determined. Cell lysate samples (70 g protein) or bacterial lysate samples (100 g protein) were submitted to 10% SDS-polyacrylamide gel electrophoresis and transferred to a Hybond C nitrocellulose membrane (Amersham Pharmacia Biotech). The membrane was probed with specific primary antibodies followed by horseradish peroxydase conjugated secondary anti-rabbit or anti-goat IgG, and finally developed using a chemiluminescence kit (Pierce, Antwerpen, Belgium).

Purification of fusion proteins produced in Escherichia coli

The full-length cDNA encoding human ILK-1 and ILK-2 were purified, and inserted into the EcoRI-XhoI site of the pGEX-4T-2 vector (Amersham Pharmacia Biotech), in frame with the GST-coding sequence and containing a thrombin cleavage site. The sequence was verified using the T7 sequencing kit. Native GST and in-frame GST-fusion proteins were expressed in E. coli BL21 following IPTG induction and affinity-purified as previously described (Vallar et al., 1999). Briefly, bacterial pellets were suspended in PBS containing 10 mM EDTA and 50 M AEBSF, and incubated for 30 min at room temperature in the presence of 200 g/ml lysozyme. The bacterial suspension was submitted to short bursts of sonication, and further treated with 1% Triton X-100 for 30 min at 4°C. The insoluble material was pelleted by a 20 min centrifugation at 10 400 g at 4°C. Supernatants were filtered on a 0.8 m membrane, and submitted to affinity chromatography on a glutathione-Sepharose 4B (Amersham Pharmacia Biotech) column (100´20 mm), previously equilibrated with PBS. Bound fusion proteins were eluted with 10 mM reduced glutathione in 50 mM Tris-HCl, pH 8.0.

In vitro kinase activity of bacterially expressed GST-ILK

In vitro kinase assays were performed using standard experimental conditions as described (Hannigan et al., 1996). Briefly, bacterially-expressed and glutathione-purified GST/ILK-1 or GST/ILK-2 (6 g) was incubated for 20 min at 30°C in the presence or absence of 10 g of the exogenous substrate myelin basic protein (MBP) or the integrin 3 cytoplasmic tail fusion protein (GST/3_cyto) in a total volume of 50 l kinase reaction buffer (50 mM HEPES pH 7.0, 10 mM MnCl₂, 10 mM MgCl₂, 2 mM NaF, 1 mM Na₃VO₄), containing 10 Ci of [-³²P]ATP. The reaction was stopped by the addition of an equal volume of 2´SDS-PAGE sample buffer. The kinase reaction products were resolved in a SDS-PAGE gel, transferred to nitrocellulose and the membrane exposed to autoradiography for 48 h at -80°C in the presence of amplifying screens. The same kinase reaction in the presence of [-³²P]ATP was also performed as a negative control.

Transfection of ILK-2 or ILK-1 cDNA into SK-Mel-2 cells

The pcDNA3.1(+)neo vector (5 g) (Invitrogen, Leek, The Netherlands) containing the full-length ILK-2 or ILK-1 cDNA, or the empty vector were mixed with 40 g of lipofectamine (Life Technologies) in a final volume of 200 l of IMDM and added to SK-Mel-2 melanoma cells grown to 60% confluence in 100-mm tissue culture plates. After 24 h, FCS was added to the culture medium, and 48 h after transfection, the cells were grown in the presence of 1 mM neomycin. Neomycin-resistant cell clones were analysed by RT-PCR/RLFP for ILK-2 expression and by PCR amplification of genomic DNA for the presence of the ILK-1 cDNA vector, using vector specific primers.

Soft agar assay

To compare the clonogenic potential of parental as well as ILK-2 or ILK-1 transfected SK-Mel-2 cells in a semi-solid medium, soft agar assays were performed as described (Chen et al., 1997). Briefly, 6´10³ cells were suspended in 1 ml of 0.4% low melting point agarose (Life Technologies) dissolved in IMDM containing 10% FCS and antibiotics, and plated on top of a 1 ml underlayer of 0.8% agarose in the same culture medium in 6-well culture plates. After 3 weeks of incubation at 37°C in a humidified incubator with 5% CO₂, the wells were microphotographed and the cell colonies stained for viability analysis using the tetrazolium salt MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) (Sigma) as described (Mosmann, 1983).

Acknowledgements

This work was supported by grants from Centre de Recherche Public-Santé (CRP-Santé, Luxembourg), CNRS (France), Fondation Luxembourgeoise contre le Cancer (Luxembourg) and EC Biomed-Project BMH4-CT98-3517. B Janji is a recipient of a fellowship from the Ministère de la Culture, de l'Enseignement Supérieur et de la Recherche, Luxembourg. The data presented were obtained by B Janji as part of his doctoral thesis submitted to the University Pierre et Marie Curie, Paris, France.

The nucleotide sequence reported in this paper has been submitted to the EMBL/GenBank/DDBJ database with the Accession number AJ277481.

Alevizoploulos A and Mermod N. (1997). Bioessays 19, 581-591. MEDLINE

Cheng C, Wang XF and Sun L. (1997). J. Biol. Chem. 272, 12862-12867. MEDLINE

Chomczynski P and Sacchi N. (1987). Anal. Biochem. 162, 156-159. Article MEDLINE

Chung DH, Lee JI, Kook MC, Kim JR, Kim SH, Choi EY, Park SH and Song HG. (1998). Virchows Arch. 433, 113-117. MEDLINE

Dedhar S, Williams B and Hannigan G. (1999). Trends Cell Biol. 9, 319-323. MEDLINE

Delcommenne M, Tan C, Gray V, Rue L, Woodgett J and Dedhar S. (1998). Proc. Natl. Acad. Sci. USA 95, 11211-11216. Article MEDLINE

Derynck R, Zhang Y and Feng XH. (1998). Cell 95, 737-740. MEDLINE

Gouon V, Tucker GC, Kraus-Berthier L, Atassi G and Kieffer N. (1996). Int. J. Cancer 68, 650-662. MEDLINE

Gregoire M and Lieubeau B. (1995). Cancer Metastasis Rev. 14, 339-350. MEDLINE

Hannigan GE, Bayani J, Weksberg R, Beatty B, Pandita A, Dedhar S and Square J. (1997). Genomics 42, 177-179. MEDLINE

Hannigan GE, Leung-Hagesteijn C, Fitz-Gibbon L, Coppolino MG, Radeva G, Filmus J, Bell JC and Dedhar S. (1996). Nature 379, 91-96. MEDLINE

Huang Y and Wu C. (1999). Int. J. Mol. Med. 3, 563-572. MEDLINE

Janji B, Melchior C, Gouon V, Vallar L and Kieffer N. (1999). Int. J. Cancer 83, 255-262. Article MEDLINE

Li DM and Sun H. (1997). Cancer Res. 57, 2124-2129. MEDLINE

Li F, Liu J, Mayne R and Wu C. (1997). Biochim. Biophys. Acta. 1358, 215-220. MEDLINE

Li F, Zhang Y and Wu C. (1999). J. Cell. Sci. 112, 4589-4599. MEDLINE

Liotta LA, Steeg PS and Stetler-Stevenson WG. (1991). Cell 64, 327-336. MEDLINE

Lynch DK, Ellis CA, Edwards PA and Hiles ID. (1999). Oncogene 18, 8024-8032. MEDLINE

Maehama T and Dixon JE. (1998). J. Biol. Chem. 273, 13375-13378. Article MEDLINE

Matsumoto K, Ziober BL, Yao CC and Kramer RH. (1995). Cancer Metastasis Rev. 14, 205-217. MEDLINE

Morimoto AM, Tomlinson MG, Nakatani K, Bolen JB, Roth RA and Herbst R. (2000). Oncogene 19, 200-209. MEDLINE

Mosmann T. (1983). J. Immunol. Methods 65, 55-63. MEDLINE

Novak A, Hsu SC, Leung-Hagesteijn C, Radeva G, Papkoff J, Montesano R, Roskelley C, Grosschedl R and Dedhar S. (1998). Proc. Natl. Acad. Sci. USA 95, 4374-4379. MEDLINE

Rearden A. (1994). Biochem. Biophys. Res. Commun. 201, 1124-1131. MEDLINE

Robertson GP, Herbst RA, Nagane M, Huang JH and Cavenee WK. (1999). Cancer Res. 59, 3596-3601. MEDLINE

Stambolic V, Suzuki A, de la Pompa JL, Brothers GM, Mirtsos C, Sasaki T, Ruland J, Penninger JM, Siderovski DP and Mak TW. (1998). Cell 95, 29-39. MEDLINE

Steck PA, Pershouse MA, Jasser SA, Yung WK, Lin H, Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis T, Frye C, Hu R, Swedlung B, Teng DH and Tavtigian SV. (1997). Nat. Genet. 15, 356-362. MEDLINE

Suzuki A, de la Pompa JL, Stambolic V, Elia AJ, Sasaki T, del Barco Barrantes I, Ho A, Wakeham A, Itie A, Khoo W, Fukumoto M and Mak TW. (1998). Curr. Biol. 8, 1169-1178. MEDLINE

Tamura M, Gu J, Matsumoto K, Aota S, Parsons R and Yamada KM. (1998). Science 280, 1614-1617. Article MEDLINE

Tu Y, Li F, Goicoechea S and Wu C. (1999). Mol. Cell. Biol. 19, 2425-2434. MEDLINE

Vallar L, Melchior C, Plançon S, Drobecq H, Lippens G, Regnault V and Kieffer N. (1999). J. Biol. Chem. 274, 17257-17266. MEDLINE

Wu C, Keightley SY, Leung-Hagesteijn C, Radeva G, Coppolino M, Goicoechea S, McDonald JA and Dedhar S. (1998). J. Biol. Chem. 273, 528-536. MEDLINE

Xie W, Li F, Kudlow JE and Wu C. (1998). Am. J. Pathol. 153, 367-372. MEDLINE

Figure 1 The cDNA and deduced amino acid sequence of ILK-2 is aligned with the published human ILK-1 (Hannigan et al., 1996) and mouse ILK sequence (Li et al., 1997). Only non-identical nucleotides are shown for the human ILK-1 and mouse ILK sequences. The black boxes represent the unique restriction sites HincII in ILK-2 and BamHI in ILK-1. The deduced ILK-2 amino acid sequence differs from ILK-1 by four substitutions at position 197 (A-T), 214 (S-G), 259 (A-S), 436 (S-P), and from mouse ILK by three substitutions at position 214 (S-G), 412 (I-V), 436 (S-P) (shaded boxes)

Figure 2 RT-PCR/RFLP analysis to monitor ILK-1 and ILK-2 mRNA expression in human tumor cells and normal human primary fibroblasts. Two micrograms of total RNA, prepared from TGF-1 stimulated HT-144 cells, as well as mRNA from SK-Mel-1 and SK-Mel-2 melanoma, PC3 prostate carcinoma, HT-1080 fibrosarcoma cell lines or normal primary fibroblasts were transcribed into cDNA and processed using primers co-amplifying ILK-1 and ILK-2. The amplified cDNA fragments were purified and digested with either BamHI (B) or HincII (H). Non-digested (N) and digested fragments (B, H) were submitted to 1% agarose gel electrophoresis and stained with ethidium bromide. Molecular weight markers (1 kb Plus DNA Ladder) are shown in the left lane of the gel

Figure 3 ILK-1 and ILK-2 expression in normal human tissues. RT-PCR/RFLP was performed using 2 g of total RNA from a collection of normal human tissues (Clonetech) as described in the legend to Figure 4. The amplified PCR fragments were directly digested with either BamHI (B) or HincII (H). Non-digested (N) and digested fragments (B, H) were submitted to 1% agarose gel electrophoresis and stained with ethidium bromide. Molecular weight markers (1 kb Plus DNA Ladder) are shown in the right lane of the gels

Figure 4 Western blot analysis of ILK expression in tumor cells and normal primary fibroblasts. Cell extracts from various human tumor cell lines or primary human fibroblasts (70 g protein) were submitted to 8% SDS-PAGE and Western blot. The samples were probed with a polyclonal anti-ILK antibody. A single band of 59 kDa was identified in each cell extract

Figure 5 (a) Western blot analysis of ILK expression in HT-144 melanoma cells before (-TGF-1) and after (+TGF-1) stimulation with 10 ng of exogenous TGF-1. The membrane was first probed with anti-E-cadherin antibody, stripped and reprobed with a polyclonal anti-ILK antibody. (b) Semi-quantitative RT-PCR/RFLP analysis was performed to monitor ILK-1 and ILK-2 expression in TGF-1-stimulated HT-144 melanoma cells. Total RNA (2 g), prepared from HT-144 melanoma cells cultured in serum free medium in the presence (+TGF-1) or absence (-TGF-1) of 10 ng of exogenous TGF-1 for 24 h, was transcribed into cDNA and amplified for 25 cycles with primers co-amplifying ILK-1 and ILK-2, as well as primers for human -actin. The amplified cDNA fragments were digested with either BamHI (B) or HincII (H). Non-digested (N) and digested fragments (B, H) were submitted to 1% agarose gel electrophoresis and stained with ethidium bromide. (*) The amplified -actin cDNA fragment (264 bp) contains a HincII restriction site and is digested into two fragments of 198 and 66 bp. Molecular weight markers (1 kb Plus DNA Ladder) are shown in the left lane of the gel (M)

Figure 6 Western blot analysis of ILK expression in HT-144 cells stimulated with exogenous growth factors. Cells were grown for 24 or 48 h in the presence of TGF-1, EGF, PDGF AB or insulin. Cell extracts (70 g protein) prepared from untreated (control) or treated cells were submitted to 8% SDS-PAGE and Western blot and were probed with a polyclonal anti-ILK antibody

Figure 7 Expression of recombinant ILK-1 and ILK-2 proteins in Escherichia coli. Extracts (100 g protein) of transformed bacteria expressing the fusion proteins GST/ILK-1, GST-ILK-2 or GST alone, were submitted to 10% SDS-PAGE and Western blot analysis using a polyclonal anti-GST antibody. An 88 kDa band corresponding to GST/ILK and a 29 kDa band corresponding to free GST were detected in the ILK-1 or ILK-2-producing bacteria. The 29 kDa band corresponding to GST alone was detected in GST-producing bacteria, whereas no band was detected in non-transformed bacteria (control)

Figure 8 In vitro kinase activity of recombinant GST/ILK-1 and GST/ILK-2. (a) Glutathione affinity purified GST/ILK-2 was incubated with [-³²P]ATP in the presence or absence of the substrate MBP. The kinase reaction products were submitted to 11% SDS-PAGE, blotted onto a nitrocellulose membrane, and the membrane exposed to autoradiography in the presence of amplifying screens. (b) Glutathione affinity purified GST/ILK-1 or GST/ILK-2 was incubated with the GST/3_cyto in the presence of [-³²P]ATP, and the kinase reaction products submitted to 9% SDS-PAGE, Western blot and autoradiography. As a negative control experiment, GST/ILK-2 was incubated with MBP in the presence of [-³²P[ATP

Figure 9 Anchorage-independent growth of SK-Mel-2 cells transfected with ILK-2 or ILK-1 cDNA. (a) The non-metastatic SK-Mel-2 melanoma cells were transfected with the pcDNA-neo vector containing full-length ILK-2 or ILK-1 cDNA, and neomycin resistant clones selected. Expression of ILK-2 in the selected clone SK-Mel-2/9 was demonstrated by semi-quantitative RT-PCR/RFLP as described in Figure 5b. ILK-cDNA and -actin were co-amplified using mRNA from parental SK-Mel-2 cells and from the SK-Mel-2/9 cell clone. Non-digested (N) and BamHI (B)- or HincII (H)-digested fragments were submitted to 1% agarose gel electrophoresis and stained with ethidium bromide. (*) -actin cDNA fragments digested by HincII. The presence of the pcDNA neo vector containing full-length ILK-1 cDNA in the genomic DNA of the selected SK-Mel-2/8 clone was confirmed by PCR analysis using sense and anti-sense pcDNA vector primers. Molecular weight markers (1 kb Plus DNA Ladder) are shown in the left lane of the gel (M). (b) Parental SK-Mel-2 cells as well as the SK-Mel-2/9 ILK-2 and SK-Mel-2/8 ILK-1 cell clones were plated on top of low melting agarose in 6-well culture plates. After 3 weeks of incubation, the cell colonies were microphotographed (upper panel), and the cell viability visualized by the MTT staining method as described in Materials and methods (lower panels)