The human hepatocarcinoma H7721 cell line was obtained from the Institute of Cell Biology, Academic Sinica of China. RPMI 1640 and Geneticin (G418) were purchased from Invitrogen. Monoclonal antibody (mAb) of mouse anti-human Nm23-H1 was from Neomarkers Company. Monoclonal antibodies against human integrin α5 and β1, β-actin, rabbit polyclonal antibodies to human FAK, and Protein G PLUS agarose were from Santa Cruz Biotechnology Inc. Phosphotyrosine antibody (PT66), FITC-conjugated second antibodies, Fn and FITC-labeled phalloidin were purchased from Sigma. PVDF membrane was from Bio-Rad. TRIzol and AMV reverse transcriptase were from Promega. Other reagents, including Taq DNA polymerase, RNAase inhibitor, dNTP, oligo (dT)-18, ECL reagent were commercially available in China.

Cells were cultured at 37°C, 5% CO₂ in RPMI-1640 medium containing 10% fetal calf serum. When Tunicamycin was used, its concentration was 2 μg/ml and the incubation time was 48 h.

The pcDNA3/Nm23-H1 plasmid was a kind gift of Prof Huili Chen in our department, constructed by Guo et al as described 15. H7721 cells were transfected with pcDNA3/Nm23-H1 using lipofectamine. Stable transfectants, designated Nm23/H7721, were established by selection in 800 μg/ml G418 and were analyzed for Nm23-H1 expression by RT-PCR and western-blotting. One single clone which expressed the highest Nm23-H1 was chosen in this study. Empty vector control cells (Mock/H7721; pcDNA3 only) were generated by the same transfection and selection processes.

Expression of nm23-H1, α5 and β1 mRNAs were determined by RT-PCR. The routine method of RT-PCR in our department was described previously 16. Briefly, Total cell RNA was extracted with TRIzol and the complementary DNAs (cDNAs) were synthesized with oligo (dT)-18 primer and AMV reverse transcriptase from 3 μg total RNA. Then cDNA was amplified by Taq polymerase in 50 μl of reaction mixture containing 5 μl cDNA, 0.2 μM of the primer pair of nm23-H1, α5, β1 or β-actin (internal standard) according to the manual. From the 26th to 32nd cycle of PCR, 10 μl products were applied to agarose gel electrophoresis. The amplified DNA bands were scanned and the photographs were analyzed with NIH Image software. The semi-quantitative data were obtained by the intensity ratios of each PCR product to the β-actin. The primers of nm23-H1, α5, β1 and β-actin were synthesized by Sangon Company according to the reported sequences 171819

nm23-H1 F: 5'-ATGGCCAACTGTGAGCGTACC-3';

R: 5'-CATG TATTTCACCAGGCCGGC-3'. The product was 204 bp

α5 F: 5'-ACCAAGGCCCCAGCTCCATTAG -3';

R: 5'-GCCTCACACTGCAGGCTAAATG -3'. The product was 375 bp

β1 F: 5'-AACTTGATCCCTAAGTCAGCAGTAG-3';

R: 5'-ATCAGCAGTAATGCAAGGCC -3'. The product was 1200 bp

β-actin F: 5'-GATATCGCCGCGCTCGTCGTCGAC-3';

R: 5'-CAGGAAGGAAGGCTGGAAGAGTGC-3'. The product was 789 bp.

Briefly, cells were homogenized in 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) buffer (pH 6.5)/150 mM NaCl/2% TritonX-100/25% glycerol/0.1 mg% leupeptin and pepstatin, and then centrifuged at 1000 μg at 4°C for 15 min. After determination of protein concentration, aliquots of 50 μg of protein samples were subjected to 10% SDS-PAGE and western blot according to the modified method of Knudsen et al. 20. The membranes were blocked with 5% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) overnight and treated with 1: 500 dilutions of different primary antibodies, followed by washing with 0.05% Tween-20/PBS for 3 times and incubation with 1: 500 dilution of HRP labeled secondary antibody for further 3 h. Then the membrane was washed again and stained with ECL reagent. β-actin was used as loading control and stained with 1: 800 dilution of primary antibody and 1: 500 dilution of HRP-labeled secondary antibody. Protein bands were quantified with densitometric analysis. Expression of each protein was calculated by the ratio of the intensity of this protein to that of β-actin.

Cell adhesion experiment was carried out according to the methods described by Busk et al 21. In brief, the wells of culture plate were coated with 0.1 ml of different concentrations of Fn. In addition, 1 mg/ml poly-L-lysine and 1% BSA were coated for 2 wells each as maximal and minimal adhesion controls respectively. The plate was incubated at 37°C for 1 h, and blocked by 1% BSA at 37°C for 0.5 h after washing. Cells (1 × 10⁵) were added to each coated well and incubated for 2 h at 37°C, followed by staining with crystal violet after two washing, then the absorbance (Abs) at 595 nm was measured. Cell adhesion to the coated wells was calculated following a formula described in previous study 15. The data were expressed as the mean of triplicate wells.

Glass coverslips were coated with fibronectin as described above. Cells were plated onto the coverslips in 35-mm dishes and cultured for 24 h. Then they were fixed with 3.7% paraformaldehyde in PBS for 10 min and permeabilized with 0.5% Triton X-100 and 4% paraformaldehyde in PBS for 5 min. Actin filaments were stained with FITC-labeled phalloidin.

Wound-induced migration assay was performed as described elsewhere 22. Cells (2 × 10⁵ cells/well) were plated onto 12-well plastic plates coated with Fn (10 μg/ml) and cultured for 24 h. Then, subconfluent monolayers of the cells were scraped with a plastic pipette tip and washed with Hanks' solution twice, and the medium was replaced with serum-free RPMI-1640. The distance between migrating cell fronts was measured at 0 and 6 h after scraping.

Detection of cell surface integrin subunits was performed according to the method reported by Zhou et al 23. Cells were dispersed in 2 mM EDTA in PBS and washed twice in PBS. Then 1 μ10⁶ cells were incubated with monoclonal antibodies against α5 or β1 integrin subunits at a dilution of 1:100 in blocking buffer (1% BSA in PBS) for 45 min at 4°C. The cells were washed twice in blocking buffer, mixed with a 1:100 dilution of FITC-labeled goat anti-mouse IgG in blocking buffer, incubated for 30 min at 4°C, then the cells were washed again with PBS, suspended in 0.5 ml PBS and subjected to flow cytometry for fluorescence analysis. Integrin expression was determined to be the percentage of FITC-positive cells. The gate setting was determined by fluorescence intensity of the same cells stained with FITC-conjugated secondary antibody only.

Cells were plated onto culture dishes coated with 10 μg/ml fibronectin. Three hours after plating, the cells were washed twice with ice cold PBS, and the monolayer cells were lysed in 200 μl lysis buffer(50 mM pH7.4 HEPES/150 mM NaCl/100 mM NaF/1 mM MgCl2/1.5 mM EGTA/1% Nonidet P-40/10 μg/ml leupeptin and pepstatin, 1 mM PMSF). Cell lysate containing 500 μg protein (determined by Lowry's method) was incubated with 2 μg monoclonal antibody specific for FAK at 4°C for 1 h. Then 20 μl Protein G PLUS agarose suspension was added, and the sample was further incubated at 4°C for 3 h to immuno-precipitate FAK. Immuno-precipitated FAK was divided into two parts and subjected to 8% SDS-PAGE and western blot as described above. The membranes were probed with 1:1000 dilution of mouse monoclonal phosphotyrosine antibody (PT66) or 1: 500 dilution of FAK antibody, followed by incubation with 1: 500 dilution of HRP labeled second antibody. The color was developed with ECL reagent. The tyrosine phosphorylation (Tyr p) of FAK was calculated from the ratio of staining intensity of Tyr p to that of FAK.

Values were expressed as mean ± SD. Statistical significance was determined with SPSS 10.0. Results were evaluated by Student's t tests. P < 0.05 and p < 0.01 were considered statistically significant and very significant respectively.

Expression of Nm23-H1 was monitored by RT-PCR and western blot. In Nm23-H1 transfected cells, mRNA level of nm23-H1 was increased significantly when compared with that in mock-transfected cells. The ratio of nm23-H1 mRNA in Mock/H7721 to that in Nm23/H7721 was 1:2.94 ± 0.58 (p < 0.01). Meanwhile, the expression level of nm23-H1 between mock and wild H7721 cells showed no significant difference (Fig 1A). The western blot result was similar to that of RT-PCR with a ratio of Nm23/H7721 over Mock/H7721 Nm23-H1 level of 2.16 ± 0.37 (p < 0.01) (Fig 1B). These data indicates a successful transfection of H7721 cells with Nm23-H1.

Adhesion of mock and Nm23-H1 transfected cells to Fn was Fn concentration-dependent. However, the adhesion of the Nm23-H1 transfected cells to Fn was decreased in all concentrations tested as compared with the mocked cells tranfected with pcDNA3 vector (p < 0.05) (Fig. 2A).

Actin filaments were visualized with FITC-labeled phalloidin staining 24 hrs after cells being plated onto dishes coated with fibronectin. Fig. 2B showed mock-transfected cells formed well-developed actin stress fibers in ordered, compact and clear-cut structure with undisturbed edges. In contrast, Nm23-H1 transfected cells was disturbed and failed to form a complete cytoskeleton on fibronectin-coated dish.

As shown in Fig. 2C, cell migration was also decreased in Nm23-H1 transfected cells when compared with the mock-transfected cells (p < 0.01).

Taken together, these results are consistent with the conclusion that increased Nm23-H1 expression changed cell adhesion and migration to Fn.

Given overexpression of Nm23-H1 impaired cell binding to Fn, it was important to determine if cell surface α5β1 integrin levels were altered. Fig 3A,B showed that the expression of β1 integrin subunit was down regulated to 39.6 ± 5.1% of the "Mock" level in Nm23/H7721 cells (p < 0.01). However, the expression of α5 subunit was unaltered on Nm23/H7721 cells compared with the Mock/H7721 cells.

Surface expression of integrin subunits was mainly regulated at transcriptional and post-transcriptional levels. In order to elucidate the mechanism of how Nm23-H1 regulates the expression of cell surface integrin subunits, we determined the mRNA levels of integrin subunits by RT-PCR. We found that mRNA levels of α5 and β1 subunit were not changed in Nm23/H7721 cells (Fig. 4). This data suggested that the decrease of cell surface integrin β1 subunit was not affected by transcriptional regulation.

To further study whether the decrease of integrin β1 subunits on cell surface was due to post-transcriptional regulation, we compared the total expression level of cellular β1 subunit by western blotting. As previously reported, two bands are typically observed in western blots of β1 integrin 24, namely a 115 kD partially glycosylated precursor and a 130 kD fully glycosylated mature form. It was very interesting to find that the total amount of β1 subunit was also unaltered in Nm23/H7721 cells, but the ratio of mature to precursor integrin isoforms was decreased significantly, being 1:1.21 ± 0.39 in Nm23/H7721 cells compared with 1:0.33 ± 0.12 in Mock cells (Fig 5A). This result suggested that overexpression of Nm23-H1 did not change total expression levels of β1 integrin. Instead, Nm23-H1 modulated the posttranslational processing of β1 integrin.

To further demonstrate that the alterated expression of mature β1 subunit was due to aberrant glycosylation, rather than other post-transcriptional regulation, we treated the cells with tunicamycin, an N-glycosylation inhibitor, and observed the deglycosylated form of β1 subunit. As shown in Fig. 5B, both Nm23/H7721 and Mock/H7721 cells only showed one band of about 90 kD crossed with intergrin β1 subunit antibody. Their size corresponded to the completely deglycosylated core peptide of the β1 subunit and their levels were almost equal. So these results indicated that the reduction of cell surface integrin β1 subunits in cells transfected with Nm23-H1 might be due to the changes of glycosylation.

FAK is associated with the intracellular domain of integrin β subunit and involved in signaling transduction for cell adhesion and migration 25. We tested whether Nm23-H1 overexpression affected phosphorylation of FAK on cells stimulated with fibronectin. As shown in Fig. 6, tyrosine autophosphorylation of FAK in Nm23-H1 transfected cells was decreased to 32.2 ± 6.4% (p < 0.01) compared with Mock cells.