Transcription from the polyoma late promoter in cells stably ...

MOLECULAR AND CELLULAR BIOLOGY, Apr. 1985, p. 797-807 0270-7306/85/040797-11$02.00/0 Copyright C 1985, American Society for Microbiology

Vol. 5, No. 4

Transcription from the Polyoma Late Promoter in Cells Stably Transformed by Chimeric Plasmids FRANCIS G. KERN AND CLAUDIO BASILICO* Department of Pathology, New York University School of Medicine, New York, New York 10016 Received 16 November 1984/Accepted 21 January 1985 We have examined the expression of chimeric plasmids containing coding sequences for the herpes simplex virus thymidine kinase (tk) gene or the TnS gene for neomycin resistance (neo) linked to the late promoter of polyoma DNA. Although polyoma late genes are generally not expressed in transformed cells containing only integrated viral DNA molecules, rat tk- or wild-type cells transfected with the tk- or neo-containing plasmids were capable of growing in medium containing either hypoxanthine-aminopterin-thymidine or G418, respectively, under conditions nonpermissive for extrachromosomal DNA replication, indicating that the tk or neo genes were fully expressed. Moreover, cells were capable of growth in either hypoxanthine-aminopterin-thymidine or G418, even in the absence of direct selection for this activity. Northern analysis indicated steady-state levels of tk or neo transcripts that approximated the levels of polyoma early transcripts. S1 analysis showed that these transcripts initiated within the late promoter of polyoma and that their 5' ends mapped at positions similar or identical to those utilized during late lytic infection. The effect of substitution of polyadenylation signals was examined. Although plasmids containing the polyoma early polyadenylation signal were more efficient in conferring to cells a stable G418-resistant phenotype than similar constructions using the late signal, both signals were found to be effectively utilized. This indicates that the inability to detect late transcripts in polyoma-transformed cells in the absence of free viral DNA production is not an effect of inefficient mRNA cleavage or polyadenylation. Our results suggest that late gene expression in integrated polyoma genomes is not regulated at the level of message initiation but, most likely, through posttranscriptional events.

The polyomavirus genome is a double-stranded, circular DNA molecule of ca. 5.3 kilobase pairs in size. It is organized into two transcriptional units, the "early" and "late" regions (48). Early region transcription has been extensively analyzed in mouse cells that are permissive for virus infection as well as in nonpermissive rat cells transformed by the virus. Si analysis reveals that the early 5' termini and splice junctions found in transformed cells are identical to those observed during early lytic infection (19, 29, 30). The location of the 3' termini in transformed cells, however, is dependent upon the location of the junction of host and integrated viral DNA sequences (19). As a consequence of early lytic transcription, three early proteins are synthesized before the onset of viral DNA synthesis. One of these proteins, the large T antigen, by virtue of its role in DNA replication, is responsible for the increase in the number of viral molecules that serve as templates for late region transcription. Late lytic transcription is characterized by a heterogeneous mixture of 5' initiation sites (13, 50). Additional diversity is generated by the splicing of giant multimeric nuclear transcripts that presumably result from repeated transcription of the circular genome (1). This splicing produces a noncoding leader sequence that contains a variable number of repeats of a 57-nucleotide segment and precedes the bodies of the messages for the three late viral proteins (21, 28, 32). It has been known for some time that nonpermissive cells transformed by polyomavirus or the related virus simian virus 40 contain very low or undetectable levels of late region transcripts, even when the corresponding coding regions can be shown to be integrated intact (17, 19, 29, 31, 48). In rat cells transformed by polyomavirus, late transcrip*

tion is easily observed after the excision and extrachromosomal replication of integrated viral genomes, thus ruling out the possibility that these cells lack some factor necessary for viral late transcription (29). In simian virus 40-transformed or abortively infected rat or mouse cells, a low but detectable amount of late transcripts is observed in the absence of viral DNA replication. 5' end mapping of these transcripts indicates that they are less heterogeneous than the 5' ends of transcripts produced during the lytic cycle and are a subset of the 5' termini used during lytic infection (17). The inability to detect significant levels of late gene transcription in polyoma-transformed rat cells may be due to the fact that the viral late promoter is inactive in integrated genomes, perhaps because its function requires independent viral DNA replication. Alternatively, there may be defective processing of late transcripts. We have examined this question by transfecting rat cells with chimeric plasmids that contain the coding sequences of the herpes simplex virus thymidine kinase (tk) or the neomycin resistance (neo) gene positioned downstream of the polyoma late promoter. Our results indicate that, in cells stably transformed by these constructs, the polyoma late promoter is functional in an integrated state at levels approximately equal to those of the early promoter, suggesting that the regulation of late gene expression in polyoma-transformed cells may occur at a posttranscriptional level. MATERIALS AND METHODS Cells and cell culture. Rat 2 tk- cells (49) were provided by W. Topp and B. Ozanne and were maintained in Dulbecco modified Eagle medium containing 10% (vol/vol) calf serum. Rat F2408 fibroblasts were maintained in the same medium. Transformed cell lines were derived from colonies isolated after plating 105 cells in 60-mm tissue culture dishes in

Corresponding author. 797

798

KERN AND BASILICO

suspension in medium containing 0.34% agar, as previously described (40). G418-resistant colonies were obtained by plating 105 cells into 100-mm dishes in Dulbecco modified Eagle medium containing 10% calf serum and 200 to 250 ,ug of G418 per ml (Geneticin; GIBCO Laboratories, Grand Island, N.Y.). Cells were maintained at 39°C and fed with G418-containing medium. Isolated colonies were picked 10 to 12 days after plating. Transfection. We used the procedure of Graham and van der Eb (23), with the modifications of Wigler et al. (55) as previously described (15). Two micrograms of plasmid DNA and 18.0 ,ug of rat liver DNA were added routinely in a volume of 1 ml to cells that had been plated at 106/100-mm dish the previous day. F2408 cells were incubated overnight in the presence of the calcium phosphate precipitate at 33°C; rat 2 tk- cells were incubated for 8 h. The cells were then rinsed in Dulbecco modified Eagle medium and shifted to 39°C in the same medium containing 10% calf serum. Four to 24 hours later, the cells were trypsinized and plated in suspension in agar medium or into G418-containing medium. RNA extraction and Northern blot analysis. RNA was extracted by a modification of the guanidium-isothiocyanate method (36). Cells were rinsed and scraped into a cold Tris-buffered saline solution, centrifuged, and suspended into at least 10 volumes of a 4 M guanidium-isothiocyanate-0.1 M sodium acetate-5 mM EDTA solution (pH 5.0). The lysate was vortexed and layered on top of a 4.0-ml cushion of 5.0 M CsCl-0.1 M sodium acetate-5 mM EDTA. The sample was centrifuged in an SW41 rotor overnight at 31,000 rpm at room temperature. The supernatant was aspirated, and the pellet was resuspended into sterile 0.01 M Tris-Cl-0.001 M EDTA-0.05% (wt/vol) sodium dodecyl sulfate and stored as an ethanol precipitate at -20°C. Selection for polyadenylated [poly(A)+]RNA on oligodeoxythymidylate columns was performed as described by Maniatis et al. (36), and poly(A)+ RNA was stored as ethanol precipitates at -20°C. Five to 10 micrograms of poly(A)+ RNA was fractionated in 1.2 or 1.4% agarose gels containing 2.2 M formaldehyde. The gel was rinsed in 20x SSC (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and transferred to nitrocellulose in the same buffer (3, 47). Hybridization was performed as described by Wahl et al. (53) with 32P-labeled, high-specific-activity, nick-translated probes (41). Filters were washed and exposed to X-ray film (Kodak XAR-5) in the presence of intensifying screens at -70°C for 1 to 3 days. Si nuclease analysis. Mapping of 5' termini (6, 54) was performed as described by Favaloro (18). Purified restriction fragments were end labeled with [-y-32PIATP (3,000 or 7,000 Ci/mmol; New England Nuclear Corp., Boston, Mass.) and digested with a second internal restriction enzyme. Doublestranded probes were eluted from polyacrylamide gels and hybridized with 10 to 20 p.g of poly(A)+ RNA at 52°C for 3 h. The L strand of the end-labeled Bcll-BgI fragment of polyoma was isolated from a 5% denaturizing polyacrylamide gel containing 8 M urea. Hybridization with the

single-stranded probe was performed at 30°C overnight. Si digestions were performed at 12°C for 2 h at a concentration of 200 U/ml when the single-stranded probe was used; double-stranded probes were digested at 14°C for 2 h or 37°C

for 1 h with 100 U of enzyme per ml. With the doublestranded probes, the various temperature conditions had no effect on the size of the protected fragments observed. Protected fragments were size fractionated on either 8 or 12% polyacrylamide sequencing gels. The size of protected fragments was determined by comparison to the size of end-labeled pBR322 MspI digest markers and to A+G or

MOL. CELL. BIOL.

C + T sequencing ladders generated by partial chemical cleavage of the hybridization probes (37). Plasmid isolation and construction. Standard recombinant DNA techniques were employed in the construction of all plasmids. The alkaline lysis method of Birnboim and Doly was used to rapidly prepare small amounts of plasmid DNA to determine the orientation of plasmid inserts (8). Large amounts were prepared by the alkaline lysis method and cesium chloride-ethidium bromide gradient centrifugation (36). The construction of tsA PypML has been described previously (14). pPyTK3 and pPyTK22 were constructed by ligation of the 2.8-kilobase (kb) BglII-BamHI fragment of the HSV tk gene that lacks the 5' gene control region (38) to BclI-cleaved tsA Py DNA. The ligated DNA was digested with BamHI, and the 8.1-kb linear fragment was isolated from an agarose gel and ligated to BamHI-cleaved and bacterial alkaline phosphatase-treated pML bacterial vector DNA. Competent HB101 were transformed, and individual colonies were picked and screened. pPyNeo was constructed in a similar fashion. Plasmid pIBW3 (a gift of A. Pellicer) contains the Tn5 gene for neomycin resistance (5, 11, 45) under the control of the HSV tk promoter and linked to the HSV tk polyadenylation signal. This plasmid was digested with BamHI and BglII to remove the HSV tk promoter. The gel-eluted 4.6-kb fragment containing the neomycin coding sequences, HSV tk polyadenylation signal, and BamHI-to-PvuII fragment of pBR322 was then ligated to a 4.9-kb BamHI-BclI fragment of tsA polyoma DNA. The ligation mix was used to transform HB101. pPyNeoA-2 was constructed as follows. A 1.15-kb BgIII-SalI fragment containing neomycin coding sequences and 3' noncoding sequences was isolated from plasmid pL088 (a gift of J. Lupski), which contains TnS inserted into pBR322. This fragment was filled in by using the Klenow fragment of Escherichia coli polymerase I, and BglII linkers were attached. The fragment was then inserted into the vector pML containing a BglII linker at the BamHI site. The 1.15-kb BglII insert was then gel eluted. The dam- dcm- bacterial host 1255 (a gift of J. Lupski) was transformed with tsA PypML, plasmid DNA was extracted, and a 4.9-kb BclIBamHI fragment was isolated and ligated to the 1.15-kb BglII fragment. The ligase was heat inactivated, and the ligated DNA was double digested with BamHI and BgIII. A 6.0-kb fragment was isolated and ligated to the pML vector containing the BglII linker. The ligation mix was then used to transform HB101, giving rise to pPyNeoA-2 that contains a single BglII site 343 base pairs (bp) downstream of the neomycin 3' coding sequences. A 622-bp Sau3A I fragment from nucleotides 2763 to 3384 of tsA PypML was inserted into this BglII site. This fragment contains both the polyoma early and polyoma late polyadenylation signals on opposite strands (12, 44). Insertion in one orientation gives neomycin coding sequences linked to the early polyadenylation signal (pPyNeoAE-2); insertion in the opposite orientation gives sequences linked to the late signal (pPyNeoAL-9). pCB7 contains neomycin coding sequences and an HSV tk polyadenylation signal linked to the polyoma early promoter at nucleotide 153 through a BglII linker. It was constructed in the following manner. Plasmid pIBW3 was digested with BamHI and BglII to remove the HSV tk promoter sequences. This generates a 4.6-kb fragment described above. A 0.8-kb BamHI-HphI fragment of polyoma DNA containing the origin of replication and early and late control regions was blunt-ended with T4 DNA polymerase. BgIIT linkers were attached (restoring the BamHI site), and after BgIIh and BamHI double digestion, the 0.8-kb BamHI-BglII frag-

799

POLYOMA LATE TRANSCRIPTION

VOL. S, 1985 L

X-_

L

s,, ,I/BI O.

EX-

py

hu1w

pr

Py

-

FIG. 1. Chimeric recombinant plasmids pPyTK3, pPyTK22, and pPyNeo. A promoterless HSV tk coding sequence and polyadenylation signal was linked to the late control region of polyoma DNA. pPyTK3 contains the HSV coding sequences contiguous with the polyoma late strand; pPyTK22 contains the tk coding sequences in the opposite anti-late orientation. The orientation of the pML sequences is identical in both constructs. For pPyNeo, neomycin resistance coding sequences from the bacterial transposon Tn5, together with 3' noncoding sequences and a region of the HSV tk gene containing the polyadenylation signal, were linked to the polyoma late control region as described in the text. neo sequences are in the same orientation as the interrupted late coding sequences of polyoma. ment was ligated to the 4.6-kb BamHI-BglII fragment of

pIBW3. RESULTS Our initial experiments were performed by using the plasmids (Fig. 1) in which a promoterless herpes simplex virus tk gene (2.8-kb BglII-BamHI fragment) was ligated in both orientations to the unique Bcll site of a linear molecule of tsA polyoma, which produces a thermolabile large T antigen (22, 26, 27), and inserted into the BamHI site of the vector pML (35). Rat 2 tk- fibroblasts were transfected with these plasmids, and transformed cells were selected in agar medium at 39°C, the temperature that is nonpermissive for large T antigen function. Colonies were isolated and tested for their dbility to grow in medium containing hypoxanthineaminopterin-thymidine at the same temperature. Since late polyoma transcription is not detected at appreciable levels in polyoma-transformed rat cells in the absence of viral DNA excision and replication (19) and these phenomena require a functional large T antigen (4, 56), we were surprised to observe that seven of eight cell lines with the tk sequences in the late orientation were able to survive in hypoxanthineaminopterin-thymidine medium. In contrast, out of seven cell lines containing tk in an anti-late orientation, none was able to survive. We expanded on this finding by using a similar construct (Fig. 1). In this case, neomycin resistance coding sequences and an HSV tk polyadenylation signal were ligated to the Bcll site of polyoma in a late orientation. Rat F2408 cells

were transfected, and again colonies were selected on the basis of ability to grow in soft agar at 39°C. Colonies were expanded, and plating efficiencies were determined in regular medium and in medium containing the drug G418 (11, 45). Five of six cell lines tested were capable of forming colonies in G418 (Table 1). Northern blot analysis of tk or neo transcription in rat cells transfected with chimeric plasmids. The results of the transfection experiments indicated that the ability to grow in hypoxantine-aminopterin-thymidine medium was dependent upon the orientation of the tk insert. Taken together with the

TABLE 1. Plating efficiency of pPyNeo-transformed cells in

G418-containing mediuma Cell line

A6

B1 B2 B8 22H2 22H3

Frequency of colony formation DME-1O% CS +G418

3.4 10.5 25.9 1.8 13.2 2.1

x 10-2 x 10-2 x 10-2

x 10-2 x 10-2 x 10-2

3.2 x 10-3 5.8 x 10-2 17.5 x 10-2 0 9.1 x 10-2 3.9 x 10-2

Ratio (48DE (G418/DME)

0.09 0.55 0.68 0 0.69 1.86

a Cell lines were derived from colonies picked from agar after transfection of rat F2408 cells with plasmid pPyNeo (Fig. 1). Cells (5 x 102) were plated into 100-mm dishes in Dulbecco modified Eagle (DME) medium containing 10%o calf serum (CS) or DME-100 CS plus 250 1Lg of G418 per ml. Cells were incubated at 39°C, and 10 days after plating the colonies were fixed, stained, and counted. Numbers represent the average of five plates.

800

MOL. CELL. BIOL.

KERN AND BASILICO 1

3

2

1

4

2

o. -we

2.92. 5-

2.9 2.5S

0

l

- 1.3

W-1.2

-0.8

p

H S V -T K

yon-la Ear y

A

Neo

Pol yoma Ear y

B

FIG. 2. Steady-state levels of polyoma early, HSV tk, and neo mRNA in cell lines transformed by plasmids containing tk or neo linked to the polyoma late promoter. Poly(A)+ RNA (5 pLg) was size fractionated on formaldehyde-agarose gels and transferred to nitrocellulose. (A) Lane 1, Poly(A)+ RNA from cell line 12L11, derived from an agar colony obtained after transfection of Rat 2 tk- cells with pPyTK3, hybridized to a nick-translated 1.8-kb PstI 3 fragment containing the early region of polyoma. Lane 2, Poly(A)+ RNA from cell line 31L2, derived from an agar colony obtained after transfection of Rat 2 tk- cells with pPyTK22, hybridized to the PstI 3 probe. Lane 3, Poly(A)+ RNA from 12L11 hybridized to a nick-translated 3.4-kb HSV tk BamHI fragment. Lane 4, Poly(A)+ RNA from 31L2 hybridized to the HSV tk probe. (B) Lane 1, Poly(A)+ RNA from cell line 22H2, derived from an agar colony obtained after transfection of rat F2408 fibroblasts with pPyNeo, hybridized to the polyoma early region PstI 3 probe. Lane 2, Poly(A)+ RNA from 22H2 hybridized to a nick-translated 1.1-kb BgllI-SaIl fragment that contains neo coding and 3' noncoding sequences.

neo plasmid transfection data, this suggested that transcription in the late direction took place in polyoma-transformed cells, even in the absence of direct selection for this transcription to occur. To verify this hypothesis, we extracted total poly(A)+ RNA from two representative cell lines containing tk in late or anti-late orientations and from a third cell line containing neo in the late orientation. We size fractionated this RNA on agarose gels containing 2.2 M formaldehyde, transferred the RNA to nitrocellulose filters, and hybridized to either tk- or neo-specific nick-translated probes. Duplicate lanes from the same gel were hybridized to a probe specific for the early region of polyoma DNA. Two prominent bands of ca. 1.3 and 0.8 kb are observed when RNA from the cell line containing tk in the late orientation is hybridized to a tk probe (Fig. 2A). The 1.3-kb band is not observed in the RNA from the cell line containing tk in an anti-late orientation. If the assumption is made that the specific activities of the tk and polyoma early region probes are approximately equal, the results would also indicate that the steady-state level of transcription in the early and late directions is approximately the same. Figure 2B shows the results of Northern analysis on a cell line containing neo sequences in a late orientation. In this case, with a neo-specific probe, a 1.2-kb band is observed that, again making the same assumption regarding the specific activities of the probes, is present at the same levels as early transcripts. Southern blot analysis of DNA taken from cells

transfected with the tk-containing plasmid indicate that the plasmid sequences are integrated, and no extrachromosomal tk-containing DNA could be detected in cells grown at the nonpermissive temperature for large T antigen (data not shown). Moreover, we have used additional constructs similar to pPyNeo in which we have deleted either the complete polyoma early region or the viral origin of replication. Such deletions do not affect the ability to form colonies in G418-containing medium (manuscript in preparation). Therefore, we are certain that the tk- or neo-containing mRNAs are transcribed from an integrated template and that late promoter activity has no absolute requirement for large T antigen function. Si analysis of tk and neo transcripts. An unspliced 1.3-kb tk- or 1.2-kb neo-containing transcript would place the approximate start of these messages at the junction with polyoma sequences (5, 52). We mapped the 5' termini of these transcripts more precisely by the method of Berk and Sharp (6) to determine whether the initiation sites were occurring within the adjacent polyoma late promoter region. Alternatively, the transcription we observed could have been due to the utilization of possible cryptic promoter sequences in the tk or neo sequences caused by the presence of the upstream polyoma enhancer region (16, 51). We first used a double-stranded 361-bp TaqI-HpaII fragment to analyze the transcription in the cell line 12L11, which contains tk in the late orientation. This fragment was

POLYOMA LATE TRANSCRIPTION

VOL. 5, 1985

5' end labeled at the TaqI site, which is located 92 bp downstream of the junction between the tk and polyoma sequences (44, 52). Therefore, any protected fragment larger than 92 bp would indicate the initiation of transcription or the location of a splice site within the adjacent polyoma sequences (Fig. 3). Three protected fragments map to sites within the sequences that form the reiterated leader observed during polyoma lytic infection (13, 21, 28, 32, 50).

801

622-

404-

4

309242-

ol