Vol. 11, No. 8
MOLECULAR AND CELLULAR BIOLOGY, Aug. 1991, p. 3915-3924
A Novel, Plasmid-Based System for Studying Gene Rearrangements in Mammalian Cells ROBERT S. KRAUSS'* AND I. BERNARD WEINSTEIN1' 2 Institute of Cancer Research and Comprehensive Cancer Center1 and Department of Genetics and Development,2 Columbia University College of Physicians and Surgeons, New York, New York 10032 Received 2 February 1991/Accepted 8 May 1991
We have developed a plasmid-based system for isolating gene rearrangements in mammalian cells by selection for reversion of a promoterless drug resistance gene. pNH4 contains the selectable marker gene neo under the control of the herpes simplex virus, thymidine kinase (tk) promoter and, upstream and in the opposite orientation, a dormant promoterless hygromycin B resistance gene (hph) that can be expressed following rearrangement events. An NIH 3T3 cell line stably transfected with pNH4 that has a spontaneous frequency of generation of Hphr colonies of -10-8 was isolated. Treatment of this line with ethyl methanesulfonate raised the frequency of Hphr colony formation -100-fold. Approximately 60% (21 of 35) of ethyl methanesulfonate-induced Hphr clones showed rearrangements detectable by Southern blot analysis within a 40-kb region surrounding the integrated construct, including a nonhomologous recombination event and, possibly, a large insertion. Additionally, three Hphr clones showed evidence of gene amplification. Northern (RNA) blot analysis of hph mRNA suggests that the rearrangements may provide a function that allows the tk promoter to initiate transcription off the opposite strand, thus yielding hph transcripts. Cell lines harboring pNH4, or modifications of it, may be valuable for studying recombination mechanisms responsible for the various types of genetic rearrangements found in cancer cells.
Clinical, epidemiological, and experimental evidence indicates that carcinogenesis is a multistage process, multiple genetic events being necessary for the neoplastic transformation of a normal cell (11, 14, 32, 62). Chemical carcinogens have been demonstrated to produce simple point mutations in several systems (51), and considerable emphasis has been placed on the role of this type of mutation. Carcinogenesis, however, also involves more complex changes in the genome. In fact, chromosomal rearrangements and genomic instability are typical of most advanced malignancies: aneuploidy, translocations, large deletions, gene amplification, and retrotransposition have all been observed in human and chemically induced rodent tumors (2, 17, 31, 46, 59). The recombinational mechanisms by which such phenomena arise are unclear, but some of the potential consequences are becoming apparent through the study of oncogenes and tumor suppressor genes. The trk and met oncogenes were isolated as a result of their activation by complex translocation events (24, 39), while N-myc and HER-2/erbB2 are frequently present in large amplified arrays of tumor cell DNA in human neuroblastomas and mammary adenocarcinomas, respectively (6, 52). Furthermore, the role of reciprocal translocation in the activation of the c-myc gene in Burkitt's lymphoma and the c-abl gene in chronic myelogenous leukemia is well established (2, 59). Recent studies provide evidence that putative tumor suppressor genes are inactivated in various malignancies by translocation, deletion, and other rearrangement events (13, 19, 54, 55). A wide variety of transformed rodent cell lines and chemically induced rodent tumors express high levels of endogenous retroviruslike mRNAs (for a review, see reference 63). These sequences are potentially capable of inducing inser*
tional mutations via reverse transcription and could therefore generate gene rearrangements in such cell lines and tumors. This hypothesis is supported by the presence of intracisternal A-particle sequences inserted into the c-mos gene in certain murine myelomas (16, 41) and into the interleukin-2 and Hox-2.4 genes in a murine monomyelocytic leukemia (4, 64). These insertions resulted in the transcriptional activation of the respective target genes. Recently, insertion of a long interspersed middle repetitive DNA element (LINE-1; a nonviral retroposon) into the c-myc locus in a human breast carcinoma (33) and into the factor VIII genes of two hemophilia A patients (22) has also been observed. While the importance of understanding the recombinational mechanisms underlying the generation of these various types of gene rearrangements is obvious, with the exception of gene amplification (53), there are few systems for the controlled selection and analysis of such events. It would be of particular interest to develop a system that could be applied to a variety of cell types (i.e., different transformed and tumor cell lines, DNA repair-deficient lines, etc.) and that could select for various types of rearrangements that are currently difficult to study. We report here the development of such a model system. Our strategy was to construct a plasmid vector (pNH4) that contains the selectable marker gene neo (for stable introduction into cultured cells) and a promoterless hygromycin B resistance gene (hph) that has the potential to be activated by a gene rearrangement that provides a transcriptional control sequence.
MATERIALS AND METHODS Construction of pNH4. A promoterless hph gene with a polyadenylation signal from the herpes simplex virus thymidine kinase gene (tk) was isolated on a 1.1-kb HindIIIBamHI restriction fragment from pRSV1.1 (37; kindly pro-
Corresponding author. 3915
KRAUSS AND WEINSTEIN
MOL. CELL. BIOL.
FIG. 1. Structure of pNH4. The transcriptional orientations of the sequences are represented by arrows.
vided by A. Murphy and A. Efstratiadis). This hph gene does not contain an out-of-frame ATG sequence that is present four bases upstream of the start codon in the wild type hph gene; the resulting gene is 100-fold more efficient than the wild-type gene in transfection assays (37). The 1.1-kb fragment was treated with T4 polymerase, BamHI linkers were ligated to the fragment and digested, and the gene was inserted into the BamHI site of pIBW with T4 DNA ligase. pIBW is a derivative of pBR322 and contains the G418 resistance gene neo under the transcriptional control of the tk promoter and polyadenylation signal (20). A recombinant plasmid containing the hph gene upstream of, and in the opposite transcriptional orientation from, the neo gene was isolated and designated pNH1. pNH1 was partially digested with BamHI, and linear molecules were isolated by agarose gel electrophoresis, treated with T4 polymerase, and recircularized by ligation with T4 DNA ligase. A plasmid lacking the BamHI site 3' of the hph gene was picked and designated pNH2. To lengthen the distance between the hph and neo genes, an intron cassette derived from the late region of the simian virus 40 (SV40) genome (map positions 519 [Hincdl] to 1493 [HindIII]) was isolated on a 1.0-kb ClaI-XhoI fragment from plasmid p96 (kindly provided by S. Obici and R. Axel). BamHI linkers were added to this fragment as described above, and it was inserted into the remaining BamHI site of pNH2. A recombinant plasmid containing the intron cassette in the same orientation as the hph gene was picked and designated pNH4. The final structure of pNH4 is diagrammed in Fig. 1. All recombinant DNA procedures were carried out with standard protocols (47), and all enzymes were from either Boehringer Mannheim or New England BioLabs. Cell culture and construction of 3T3pNH4-4 cells. NIH 3T3 cells were grown in Dulbecco modified Eagle medium plus 10% calf serum (Flow Laboratories). Cultures were maintained in a humidified incubator at 37°C with 5% CO2 in air and fed twice a week with fresh medium. One microgram of pNH4 plus 20 ,ug of NIH 3T3 carrier DNA was transfected into NIH 3T3 cells by the calcium phosphate coprecipitation technique as described by Hsiao et al. (20), and the cultures were selected in medium containing 200 ,ug of G418 (GIBCO) per ml. G418-resistant colonies were isolated with cloning cylinders, expanded into cell lines, and tested for sensitivity to hygromycin B (Boehringer). One of these cell lines, designated 3T3pNH4-4, was used for further study. 3T3pNH4-4 cells were maintained in medium containing 50 ,ug of G418 per ml.
Mutagenesis of 3T3pNH4-4 cells. 3T3pNH4-4 cells were inoculated onto eight 150-mm plates (3.6 x 106 cells per plate) and 8 h later were treated overnight with 2.5 mM ethyl methanesulfonate (EMS; Sigma). The following day, the cultures were refed and allowed to recover for 3 days, at which time they were trypsinized, counted, and inoculated into medium containing 50 ,ug of G418 and 300 ,ug of hygromycin B per ml. Hygromycin B-resistant (Hph9) colonies were counted and isolated with cloning cylinders 12 to 14 days later. The colonies were expanded into cell lines and analyzed by Southern blotting. Fifteen Hphr lines were then subcloned for more detailed analyses. All Hphr lines were cultured in medium containing 50 jig each of G418 and hygromycin B per ml. The viability after EMS treatment was determined by a colony-forming efficiency assay, in which 300 control or mutagenized cells were seeded onto 90-mm plates in Dulbecco modified Eagle medium plus 10% calf serum. Colonies were counted 7 to 10 days later. Northern (RNA) and Southern blot analyses. RNA was isolated by centrifugation through a cushion of cesium chloride according to the method of Chirgwin et al. (8). The polyadenylated [poly(A)+] fraction was isolated following two rounds of selection through an oligo(dT)-cellulose column (1). RNA samples were electrophoresed through 1% agarose-6% formaldehyde gels and blotted onto Hybond N nylon membranes (Amersham) in lOx SSC (lx SSC is 0.15 M sodium chloride plus 0.015 M sodium citrate). Cellular DNA was isolated as described by Gattoni et al. (15) or by removing the DNA plug from the cesium chloride cushion following RNA isolation and dialyzing it against 10 mM Tris-0.1 M sodium chloride-1 mM EDTA, pH 8.0. The dialyzed DNA samples were treated with 0.2% sodium dodecyl sulfate and 100 ,ug of proteinase K per ml overnight at 37°C, and the samples were then extracted with an equal volume of phenol and precipitated with 2 volumes of ethanol. All DNA samples were dissolved in 10 mM Tris-1 mM EDTA, pH 8.0. DNA samples (5 ,ug) were digested with various restriction enzymes according to the manufacturer's recommendations and were then electrophoresed through 1% agarose gels and blotted onto Hybond N nylon filters. Northern and Southern blots were hybridized to nick-translated probes (42) by the procedure of Wahl et al. (60), washed at 68°C with 0.1 x SSC, and autoradiographed. The probes used represent the individual full-length sequences of the hph gene, the SV40 intron, the tk promoter, and the neo gene. For some Northern blots, single-stranded RNA probes were used in place of nick-translated probes. Sequences corresponding to the full-length tk promoter or full-length neo gene were cloned into the Bluescript vector pKS(Stratagene), and 32P-labelled RNA molecules were synthesized from the T4 or T7 promoter within the plasmid. An RNA transcription kit (Stratagene) was used according to the manufacturer's protocol. Isolation of a rearrangement breakpoint by inverse PCR. Three hundred micrograms of cellular DNA from the Hphr line EMS-26 was digested with Bglll and electrophoresed through a preparative 1% agarose gel. The gel was sliced in the size range of 1.2 to 1.6 kb (the size range containing the breakpoint; see Results), and the DNA was isolated by electroelution. Inverse polymerase chain reaction (PCR) was carried out as described by Ochman et al. (38). Briefly, 200 ng of size-fractionated EMS-26 DNA was treated with T4 DNA ligase in a final volume of 400 ,ul to yield circular molecules. The reaction was extracted with equal volumes of phenol, phenol-chloroform, and chloroform, the aqueous phase was removed, and the DNA was precipitated with
VOL. 11, 1991
GENE REARRANGEMENT IN MAMMALIAN CELLS
TABLE 1. Effect of EMS treatment on the frequency of Hphr colony formation in 3T3pNH4-4 cells Treatment
EMS (2.5 mM)b
Colony-forming efficiency (%) 42.8 41.4 23.2
No.treated of cells
No.selective of cells medium plated into
of Hphr No. colonies
2.9 x 107
4.7 x 107 8.0 x 107 1.7 x 108
0 1 115