INTRODUCTION

3Y1, a 3T3-type Fischer rat cell line, has widely been utilized as a model system to study whether or not viral or cellular gene(s) can transform cultured cells (1). When the cells reach confluence in cultured dishes, they stop growing and change their shape from a spindle-like form to a cobblestone-like form, so that the cell sheet becomes a typical monolayer of low cell density. In contrast, 3Y1 cells bearing E6 genes of oncogenic human papillomaviruses such as type 16 (hereafter called 3Y1-16E6) continue to grow without changing their morphological characteristics after reaching confluence, so that the cell sheet becomes a higher cell density than that of the original 3Y1 cells (2). A derivative line of 3Y1 (named 3Y1 BU^r-1 by others and hereafter called 3Y1 BU) and F2408 cell line (established independently from the 3Y1 line) also keep growing without changing their shape after reaching confluence. This difference in growth patterns between 3Y1 and the latter three cell lines has led us to speculate that 3Y1 cells carry a gene(s) whose mRNA is up-regulated after confluence so as to bring about the growth arrest. The speculation was supported by findings with hybrids between 3Y1 and 3Y1 BU or F2408. In both cases, the growth pattern of the 3Y1 cells appeared dominant,¹ prompting us to isolate the candidate gene(s).

As an approach to isolation of cDNAs whose corresponding mRNAs are detectable in growth-arrested confluent but not in growing subconfluent 3Y1 cells, we chose the mRNA differential display method (3, 4). Our screening allowed us to obtain four cDNA fragments that satisfied this criterion: one of them was considered to be a rat homologue of gas1 (growth arrest-specific reported by others (5), but the other three appeared novel and were tentatively named cca1-cca3 (confluent 3Y1 cell-associated) (6). RNase protection assay with a cca1 cDNA fragment as a probe revealed that although 3Y1 cells contained an appreciable amount of the mRNA after confluence as expected, the other line cells referred in the above did not. In the present paper, we describe (i) the construction of cca1 cDNA covering a large open reading frame (ORF),² (ii) the association of the corresponding mRNA levels with cell growth patterns, and (iii) the biological activity of this cDNA in restoring the 3Y1 cell-type growth pattern after introduction into 3Y1 BU or its derivative cells.

MATERIALS AND METHODS

The cells were grown in Eagle's minimal essential medium supplemented with 10% fetal calf serum except where specified. Fisher rat-derived embryonic 3Y1 cells (1) and F2408 cells (7) were generously donated by Dr. Kimura and Dr. Hakura, respectively. The thymidine kinase-minus 3Y1 BU^r-1 cells (called 3Y1 BU cells for convenience) were also kindly provided by Dr. Kimura. 3Y1 cells containing E6 gene of human papillomavirus type 16, called 3Y1-16E6 cells, were prepared as described earlier (2). Neomycin-resistant cells were selected in the medium supplemented with 200 mg/ml G418. PA317 and COS cells were cultured in Dulbecco's modified Eagle's minimal essential medium with 10% fetal calf serum for packaging of retrovirus (8) and for transient expression of cca1 cDNA, respectively. Cells harvested 2 days before reaching confluence were used as subconfluent cells and 3 days after as confluent cells.

To monitor the growth activity of 3Y1 BU cells by bromodeoxyuridine (BrdUrd) incorporation, the cells were transfected with pTK4 containing the herpesvirus thymidine kinase gene (9), followed by selection in the medium supplemented with 100 mM hypoxantine, 0.5 mM aminopterin, and 20 mM thymidine. 300 colonies of drug-resistant cells were pooled (named 3Y1 BU/pTK) and used as 3Y1 BU equivalents in the BrdUrd incorporation assay.

The cca1 cDNA fragment (372 nucleotides (nt)) used as probe DNA for cca1 cDNA cloning was obtained as one of four cDNA fragments isolated by the mRNA differential display method on the basis of the preferential accumulation of the corresponding mRNAs in confluent but not in subconfluent 3Y1 cells (6).

A cDNA library was constructed with 1 µg of poly(A)⁺ RNA extracted from confluent 3Y1 cells (SuperScript Choice system, Life Technologies, Inc.). First strand cDNA synthesis was performed using an oligo(dT) primer. After second strand cDNA was synthesized, the obtained double stranded cDNA was ligated to EcoRI (NotI) adaptors. After size fractionation, 50 ng of EcoRI-adapted double stranded cDNA was cloned into the Lambda ZAPII vector arm (Stratagene). The recombinant  DNA was packaged with packaging extract (Gigapackll Gold, Stratagene), and then the resulting phage was mixed with strain XL1-Blue MRF host cells and plated on NZY plates. The library contained 2 × 10⁶ recombinants. For screening, 1 × 10⁵ plaques of the amplified library were plated at a density of 5000 plaque forming units/100 × 150 mm² dish. Duplicate lifts were made from each of 20 dishes using nylon membranes (Gene Screen, NEN Life Science Products). The membranes were hybridized with cca1 probe labeled with [³²P]dCTP, and those plaques that specifically hybridized with the probe were isolated. After a second plaque hybridization, six clones were isolated, the longest of which was found to contain a 3.8-kb cDNA region. When the library was rescreened, no cDNA clone with a length of more than 3.8 kb could be isolated. We therefore generated the remaining 5 DNA region (1.2 kb) by the rapid amplification of cDNA ends method (5-AmpliFINDER RACE, CLONTECH). Three independent clones were isolated by independent PCR and sequenced (Sequencing PRO, Toyobo). Although sequence heterogeneity was observed for the GC-rich portion of the 5-untranslated region, a probable sequence was determined. The DNA fragment obtained was ligated at the EcoO109I site (nt 1173) to the above cDNA clone. To amplify the entire 1905-nt ORF (nt 363-2267) of the cca1 cDNA, the 11AFNOT primer containing the nt 347-377 region of the cDNA and the 11AP5 primer corresponding to the nt 3436-3450 region of the cDNA were used.

The retroviral vector derived from the moloney murine leukemia virus, named pLRNL, contains the neomycin resistance gene under control of Rous sarcoma virus promoter (10). To obtain cca1 cDNA expression vector (pLRNL-cca1), the 3.6-kb DNA fragment containing the cDNA region of 3571 nt (nt 345-3915) was introduced into the BamHI site of the pLRNL vector. To construct pLRNL-cca1(-atg), the six nucleotides (nt 360-365) containing the first ATG site of the cDNA were replaced with a BamHI site, and the BamHI fragment containing 3550 nt (nt 366-3915) of the cDNA was cloned into the BamHI site of the pLRNL vector. By using this first ATG-minus construct, the N-terminal region-deleted protein (222 amino acids (aa); aa 414-635) would be expected to be produced from the ATG (nt 1602) for the second methionine. The upstream in-frame stop codon was located in the vector DNA sequence.

To express cca1 cDNA in COS cells at a high level, the pEF-BOS vector containing the promoter region of human elongation factor 1 and the SV40 replication origin was used (11). After removing the stuffer DNA by digesting the vector with BstXI and NotI, the DNA sequence for the Myc epitope (10 aa; EQKLISEEDL) located just before the new cloning site (BamHI) was introduced as a BstXI-NotI fragment so as to produce the expression vector for the Myc epitope-tagged protein (pEF-Myc). The cca1 cDNA expression vector, pEF-Myc cca1, was constructed by cloning the BamHI fragment of pLRNL-cca1(-atg) into pEF-Myc vector.

The rat gapdh sequence was amplified with specific primers (CLONTECH) by PCR, using oligo(dT) primed 3Y1 cDNAs as template DNAs. The PCR product was cloned into pMOSBlue-T vector (Amersham Corp.), and then the resulting plasmid, pMOS-gapdh, was used for the preparation of the control probe for the RNase protection assay.

10-µg aliquots of pLRNL-cca1, pLRNL-cca1(-atg), or pLRNL were introduced into the amphotropic helper line PA317 (8) by electroporation (Gene Pulser, Bio-Rad) at 350 V and 500 microfarad. Two days post-transfection, the medium was harvested from the cells, filtered, and then used for infection in the presence of 8 µg/ml polybrene (Sigma). 200-300 colonies of neomycin-resistant cells were pooled and used as the plasmid DNA-introduced cells.

Total RNA samples were extracted from the cultured line cell by the acid guanidium thiocyanate-phenol/chloroform method as described previously (12) or by using an RNeasy kit (Qiagen). Total RNAs were also prepared from several organs of 10 week-Fischer rats (male) using ISOGEN (Nippon Gene). Poly(A)⁺ RNA fractions were enriched using oligo(dT)-Latex (OLIGOTEX-dT30, Roche).

RNase protection assays were performed with antisense strand RNA probes synthesized with T7 polymerase (MAXIscript, Ambion). Total RNA was hybridized with RNA probes at 42 °C overnight. After RNase treatment with 0.5 units of RNaseA and 20 units of RNase T1 (RPAII, Ambion), the protected RNA was electrophoresed on 5% urea-polyacrylamide gels.

The cca1 RNA probe corresponds to the 372-nt cca1 cDNA sequence (nt 4651-5022). Because the pLRNL-cca1 plasmid DNA lacks the 3 DNA region of cca1 cDNA (nt 3916-5022), the cca1 mRNA transcribed from the exogenous template derived from this plasmid DNA would not be expected to be protected by the cca1 probe. Therefore, the cca1* probe corresponding to the 384-nt cDNA sequence (nt 345-728) was used to detect the mRNA transcribed from the plasmid DNA template, as well as the endogenous template. The size of the band protected by the cca1* probe in the mRNA transcribed from pLRNL-cca1(-atg) DNA was slightly smaller (363 nt), because the plasmid DNA lacked the 21-nt DNA sequence (nt 345-365) of the cDNA region corresponding to the probe.

An oligopeptide (IPTTYEKQRADDPC) of the predicted CCA1 protein (aa 38-53) was synthesized, coupled with keyhole limpet hemocyanin, and then used for immunization of rabbit. For purification of the antibodies against CCA1 protein by affinity chromatography, rabbit serum was applied to the 2-fluoro-1-methylpyridinium toluelene-4-sulfonate-activated cellulofine column coupled with the 15-aa oligopeptide. The bound antibodies were eluted with 0.1 M glycine-HCl (pH 2.5) and immediately neutralized with 1 M Tris. After NaCl and sodium azide were added to 0.15 M and 0.03%, respectively, the resulting fraction was used as the anti-CCA1 antibodies (0.1 mg/ml protein).

For Western blotting, cell lysates were prepared by adding sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 5% -mercaptoethanol, 10% glycerol, 0.01% bromphenol blue), followed by vigorous mixing and boiling for 5 min. After SDS-polyacrylamide gel electrophoresis, proteins were blotted electrophoretically onto polyvinylidene difluoride membranes (Millipore) and blocked by soaking in PBS-T buffer (0.1% Tween 20 in PBS) supplemented with 10% nonfat dry milk and 0.1% sodium azide at 4 °C overnight. After washing with PBS-T buffer, the membranes were incubated with rabbit anti-Myc epitope antibodies (Medical & Biological Laboratories) at 3 µg/ml or rabbit anti-CCA1 antibodies at 0.3 µg/ml and then were probed with horseradish peroxidase-labeled goat anti-rabbit IgG antibodies (Zymed), both steps for 1 h at room temperature, and finally proteins were visualized using the ECL system (Amersham Corp.).

BrdUrd incorporation assays were performed using a cell proliferation kit (Amersham Corp.) with slight modifications. Briefly, the cells were labeled by incubation for 24 h in fresh medium supplemented with BrdUrd, fixed in acid-ethanol (5% acetic acid and 5% distilled water in ethanol) for 30 min at room temperature, and incubated with mouse anti-BrdUrd monoclonal antibody for 1 h at room temperature. After incubation with horseradish peroxidase-labeled anti-mouse IgG at room temperature for 30 min. BrdUrd incorporation into nuclei was visualized by staining with 3,3-diaminobenzidine tetrahydrochloride. The percentage of BrdUrd-incorporating cells was determined by counting the mean numbers of BrdUrd-incorporating cell in three different areas in two independent experiments.

After culturing in chamber slides (Nunc), cells were fixed with 4% freshly prepared paraformaldehyde in PBS for 15 min and permeabilized for 10 min with 0.5% Triton X-100 in PBS as described (13). Cells were then incubated with tetramethylrhodamine B isothiocyanate-conjugated phalloidin (Sigma) in a moist chamber for 20 min. The organization of the cytoskeletal actin filaments of the cells was then examined under a fluorescent microscope.

RESULTS

A cDNA clone (3.8 kb) was isolated from the confluent 3Y1 cell library using the cca1 probe obtained as described under "Materials and Methods." The remaining DNA region (1.2 kb) was generated with 5 rapid amplification of cDNA ends method and then ligated to the above cDNA clone at the overlapping EcoO109I site. Thus obtained cca1 cDNA consisted of 5022 nt (Fig. 1), which was compatible with the size of the corresponding mRNA estimated by Northern blot analysis (5 kb) (6). There was an ORF of 1905 nt, which started from the first ATG (nt 363) right after the in-frame TGA stop codon (nt 354) and encoded a 635-aa protein with a calculated molecular mass of 69 kDa. The DNA sequence (GCAATGG) including the first ATG resembled Kozak's consensus sequence (14). To preclude the possibility that the ORF was a chimeric artifact between two partially identical fused ORFs, a reverse transcription-PCR experiment was conducted with 11AFBAM and 11AP5 primers, designed to amplify the entire ORF as described under "Materials and Methods": a single PCR product with the size and restriction enzyme sites expected was detected from the mRNA of confluent 3Y1 cells (data not shown). The 3-untranslated DNA region contained a polyadenylation signal (AATAAA) at nt 5005 (17 nt from first A of poly(A) tail; shown as underlined in Fig. 1) and three copies of the consensus element (ATTTA; shown as boxed in Fig. 1) involved in mRNA instability (15) at nt 3884, 3948, and 4749. The DNA region (372 nt) used as the cca1 probe for the isolation of the above cDNA clone was located at nt 4651-5022. Homology searches indicated that cca1 cDNA and the predicted amino acid sequences have no similarities to genes and proteins accessible in the data bases of GenBank^TM and EMBL nucleotide sequence and SWISS-PROT, PIR, and PRF protein sequence.

RNA preparations in a number of rat organs were examined for the cca1 mRNA level. The cca1 mRNA was detected in all organs tested with relative amounts in brain and jejunum comparable with that in confluent 3Y1 cells (Fig. 2).

A single protein band was detected with both anti-Myc epitope (lane 2 in Fig. 3A) and anti-CCA1 antibodies (lane 2 in Fig. 3B) in COS cells transiently transfected with the cDNA for the Myc-tagged CCA1 protein (pEF-Myc cca1). In each case, the size of the protein detected was estimated to be 70 kDa, compatible with the size of the predicted molecular mass of CCA1 protein (69 kDa). Immunofluorescent analysis of the COS cells transfected with pEF-Myc cca1 indicated a cytoplasmic location for the protein (data not shown). In confluent 3Y1 cells, the anti-CCA1 antibodies we prepared could not detect any specific proteins (data not shown). Because the relative level of cca1 mRNA in confluent 3Y1 cells was found to be only less than one-thousandth of that in the COS cells transfected with pEF-Myc cca1 (data not shown), it seemed likely that the level of the CCA1 protein in the confluent 3Y1 cells was too low for antibody detection.

During the formation of confluent 3Y1 monolayer, 3Y1 cells changed their shape (Fig. 4A) with a different organization of the cytoskeletal actin filament (Fig. 4B), and the percentage of BrdUrd-incorporating cells was reduced to 6% (Fig. 4C). In contrast, 3Y1 BU, its derivative (3Y1 BU/pTK), 3Y1-16E6, and F2408 cells retained their typical morphology with the same actin filament organization even after confluence, and a considerably higher percentage of cells continued to incorporate BrdUrd (see Table I). The observations were essentially the same for over at least 1 week after confluence (data not shown). Using RNA preparations of the above cells, the cca1 mRNA level in 3Y1 BU, 3Y1 BU/pTK, 3Y1-16E6, and F2408 cells were compared with those in 3Y1 cells. In the former cells, cca1 mRNA was not detected irrespective of their culture conditions (Fig. 5 and Table I). The results indicate that cca1 mRNA level is up-regulated in association with the growth pattern of confluent 3Y1 cells.

Table I. Association of cca1 mRNA levels with cell growth patterns Cells Culture condition cca1 mRNA level Actin filament organization Percentage of BrdUrd incorporating cells 3Y1 Subconfluence   Stress fibers 100 Confluence + Cortical filaments 6 3Y1 BU Subconfluence   Stress fibers NDa Confluence   Stress fibers ND 3Y1 BU/pTK Subconfluence   Stress fibers 100 Confluence   Stress fibers 20 3Y1-16E6 Subconfluence   Stress fibers 100 Confluence   Stress fibers 24 F2408 Subconfluence   Stress fibers 100 Confluence   Stress fibers 59 a ND, not determined.

To determine the association of cca1 mRNA accumulation with the confluent 3Y1-type growth pattern was causative or coincidental, the pLRNL-cca1, pLRNL-cca1(-atg), or pLRNL vector was introduced into 3Y1 BU, 3Y1 BU/pTK, 3Y1-16E6, and F2408 cells, and then the cells obtained were examined for restoration of the 3Y1-type growth pattern after reaching confluence by phalloidin staining and BrdUrd incorporation. The pLRNL-cca1 was designed to produce the entire CCA1 protein (635 aa; aa 1-635), whereas the pLRNL-cca1(-atg) was expected to give rise to an N-terminal deleted protein (222 aa; aa 414-635) translated from the second methionine (shown boxed in the amino acid sequence in Fig. 1).

When the growth pattern of 3Y1 BU/pTK cells carrying pLRNL-cca1 (termed 3Y1 BU/pTK/pLRNL-cca1 cells for convenience, with the others also named similarly) was monitored by phalloidin staining, the confluent 3Y1 BU/pTK/pLRNL-cca1 cells changed their actin filament organization from predominantly actin stress fibers to mainly cortical actin filaments (a in Fig. 6A), which was indistinguishable from that of the confluent 3Y1 cells (b in Fig. 4B). Similar results were obtained in the cca1 cDNA-introduced 3Y1 BU cells (data not shown). In contrast, a consistent pattern of actin stress fibers was detected in the 3Y1 BU/pTK/pLRNL-cca1(-atg) (b in Fig. 6A) and 3Y1 BU/pTK/pLRNL cells (c in Fig. 6A), independent of their culture conditions. Similarly, no significant alternation of the actin filament organization was observed after introduction of cca1 cDNA into 3Y1-16E6 or F2408 cells (data not shown).

When the growth patterns of the 3Y1 BU/pTK/pLRNL-cca1, the 3Y1 BU/pTK/pLRNL-cca1(-atg), and the 3Y1 BU/pTK/pLRNL cells were monitored by BrdUrd incorporation, almost all these cells in subconfluent culture were found to incorporate BrdUrd into their nuclei. After confluence, only 3% of the 3Y1 BU/pTK/pLRNL-cca1 cells were found to be BrdUrd-incorporating cells. In contrast, the values were 19% for the 3Y1 BU/pTK/pLRNL-cca1(-atg) cells and 20% for the 3Y1 BU/pTK/pLRNL cells, essentially the same as for 3Y1 BU/pTK cells (20%). Similarly, no significant decrease in the percentage of BrdUrd-incorporating cells after reaching confluence was observed on introduction of cca1 cDNA into 3Y1-16E6 or F2408 cells (data not shown).

cca1 mRNA levels after reaching confluence were determined by the RNase protection assay with the cca1* probe additionally used to detect the mRNA transcribed from the exogenously introduced template, as well as the endogenous template. The amounts of cca1 mRNA in both 3Y1 BU/pTK/pLRNL-cca1 and 3Y1 BU/pTK/pLRNL-cca1(-atg) cells were estimated to be 10 times as much as in confluent 3Y1 cells (compare lanes 3 and 4 with lane 1 in Fig. 6B) and to be essentially the same in the cca1 cDNA-introduced cells derived from 3Y1 BU, 3Y1-16E6, or F2408 cells (data not shown). However, the anti-CCA1 antibodies we prepared failed to detect any specific proteins in 3Y1 BU/pTK/pLRNL-cca1 cells (data not shown).

To study whether or not the cca1 cDNA might exert a growth suppressive effect on subconfluent cells, the doubling time of cca1 cDNA-introduced 3Y1 BU/pTK cells was compared with that of the unintroduced cells. No significant differences were observed (data not shown), indicating that cca1 cDNA in itself was not sufficient to suppress cell growth in subconfluent 3Y1 BU/pTK cells.

DISCUSSION

cca1 cDNA fragment was previously isolated on the basis of preferential accumulation of the corresponding mRNA in growth-arrested confluent but not in growing subconfluent 3Y1 cells (6). Different from 3Y1 cell case, cca1 mRNA was found not to be detected in confluent 3Y1 BU, 3Y1 BU/pTK, 3Y1-16E6, or F2408 cells, whose growth patterns monitored by BrdUrd incorporation and phalloidin staining were different from that of the confluent 3Y1 cells (Table I). When cca1 cDNA was introduced into 3Y1 BU or 3Y1 BU/pTK cells, a restoration of the confluent 3Y1-type growth pattern was observed after reaching confluence. The results indicated that cca1 is required for formation of the growth-arrested monolayer of 3Y1 cells. However, the lack of any significant alternation of the growth pattern on introduction of cca1 cDNA into 3Y1-16E6 or F2408 cells and no significant growth suppressive effect of the cca1 cDNA on the subconfluent 3Y1 BU/pTK cells imply that the up-regulation of the other gene(s) may also be required for the influence to be exerted.

Introduction of a first ATG-deleted cDNA into the 3Y1 BU/pTK cells failed to restore the confluent 3Y1-type growth pattern after reaching confluence. Because the cca1 mRNA level in the cells was almost the same as the level in the undeleted cca1 cDNA-introduced 3Y1 BU/pTK cells (Fig. 6B), it was considered that not only the expression of cca1 mRNA corresponding to the ORF but also the production of the entire CCA1 protein were required for the biological activity.

The cca1 mRNA was found to be ubiquitously distributed with different levels (Fig. 2), which were not necessarily consistent with their growth activities; it should be noted that heart harbored a low mRNA level, whereas jejunum with its high growth rate demonstrated the greatest amount of mRNA observed. Further examinations of cca1 mRNA level in the different types of cells in each organ are required to reveal association of the mRNA level with growth-arrested cell state in vivo.

Mouse sarcoma cells transfected with a cDNA for the liver cell adhesion molecule change their morphology from round or spindle like form to a cobblestone-like form after reaching confluence by increasing the cell adhesiveness through the formation of adherens and gap junctions (16). The mechanisms underlying the similar morphological changes observed in 3Y1 cells remain to be elucidated, but the present isolation of cca1 cDNA provides an important clue toward their elucidation.