Simian Virus 40 Revertant Enhancers Exhibit Restricted Host. Ranges for Enhancer Function. ALYSSA SHEPARD,t JENNIFER CLARKE, AND WINSHIP HERR*.
Vol. 62, No. 9
JOURNAL OF VIROLOGY, Sept. 1988, p. 3364-3370
0022-538X/88/093364-07$02.00/0 Copyright C 1988, American Society for Microbiology
Simian Virus 40 Revertant Enhancers Exhibit Restricted Host Ranges for Enhancer Function ALYSSA SHEPARD,t JENNIFER CLARKE, AND WINSHIP HERR* Cold Spring Harbor Laboratory, P.O. Box 100, Cold Spring Harbor, New York 11724 Received 25 January 1988/Accepted 16 May 1988
We have assayed the cell-specific activity of a matched set of four enhancers found in viral revertants derived from simian virus 40 (SV40) enhancer mutants. These enhancers all contain 71-base-pair duplications that span identical regions or, in one case, the same region shifted by 2 nucleotides. The four enhancers differ, however, in that each one either carries a different wild-type pair of the genetically defined SV40 enhancer A, B, or C elements, with the other two elements mutated, or carries all three elements mutated. The three enhancers carrying two copies of a wild-type element effectively enhance transcription in CV-1 and HeLa cells, but only the enhancer containing a duplicated wild-type C element exhibits activity in NIH 3T3 cells. These results show that the ability of the A, B, and C elements to compensate for one another is cell specific and that selection for enhancer function in one cell type can generate enhancers with different cell-specific activities. These results are consistent with the hypothesis that tandem duplication of multiple distinct enhancer elements, as in wild-type strains of SV40 (e.g., the 72-base-pair repeat), has the property of expanding the host range of an enhancer.
The ability of certain promoter elements, called enhancers, to activate transcription when positioned at a large distance either upstream or downstream of the transcriptional initiation site was first discovered in studies of the simian virus 40 (SV40) early promoter (4, 27). Enhancers have since been identified in numerous viral and cellular promoters (for a review, see reference 16) and direct both wide and narrow ranges of cell-specific gene expression. For example, the activity of the heavy-chain immunoglobulin gene enhancer is largely restricted to cells of lymphoid origin (3, 15, 28). In contrast, the activity of many viral enhancers, such as those found in SV40, polyoma virus, and Moloney sarcoma virus, is not limited to one cell type, although the relative activities vary between cell lines (10, 24). The SV40 enhancer has served as the prototype of a generally active enhancer and like many generally active enhancers, the wild-type SV40 enhancer contains a tandem
duplication (the 72-base-pair [bp] repeat). Mutational studies have shown that more than one region is responsible for the activity of the SV40 enhancer (19, 45). In one series of experiments, SV40 viral growth revertants were isolated in the African green monkey kidney cell line CV-1 from SV40 mutants carrying a single 72-bp element and debilitating enhancer point mutations; sequence duplications within the revertant enhancers defined three functional elements, called A, B, and C, which range in size from 15 to 22 bp (19, 20). When any two of these three elements are inactivated by point mutation, viral revertants arise which contain tandem duplications spanning the remaining wild-type element (8). Thus, in CV-1 cells, the duplication of any one of these three elements can compensate for loss of function in the remaining two elements. The individual activity of different segments of the SV40 enhancer has been tested by construction of enhancers composed of multiple tandem copies of short synthetic oligonucleotides (30, 34). Synthetic enhancers composed of
either the A, B, or C elements can activate transcription in CV-1 cells, but each displays a unique pattern of cell-specific activity (30; B. Ondek and W. Herr, unpublished results). By using such synthetic constructs, the number of individual elements required to obtain enhancer activity was tested with a series of B-element constructs. This experiment showed that a single copy of the B element does not effectively enhance transcription, whereas two or more copies do function as an enhancer (30). A requirement for two elements to create a functional enhancer is consistent with the appearance of duplicated A, B, or C elements in viral revertants to compensate for loss of function in the remaining elements. The different cell-specific activities of the synthetic A-, B-, and C-element enhancers and the requirement for two elements to enhance transcription led to the hypothesis that the wild-type SV40 enhancer 72-bp repeat, by duplicating multiple elements, could serve to expand the host range of the enhancer. One of the expectations from this hypothesis is that duplicated enhancers with mutated elements will display restricted host ranges for activity. To test this expectation, we have assayed the activity of a matched set of revertant SV40 enhancers containing 71-bp duplications spanning identical, or in one case nearly identical, regions but with only one pair of wild-type A, B, or C elements. Consistent with this hypothesis, the revertant enhancers with wild-type A and B elements have restricted host ranges. These experiments also show that enhancers selected for activity in one cell type can possess different patterns of cell-specific activity. MATERIALS AND METHODS Plasmid DNAs were transfected into CV-1, HeLa, and NIH 3T3 cells by coprecipitation with calcium phosphate (17). Plasmid DNAs were constructed as described previously (20, 38) and were purified by alkaline lysis (21), CsCl gradient centrifugation, and polyethylene glycol precipitation. DNAs were suspended in 10 mM Tris hydrochloride (pH 8.0)-0.1 mM EDTA, and a portion was diluted to 0.1 mg/ ml in the same buffer for transfection. The cell lines were as described previously (30). Cells were seeded 18 to 24 h prior
* Corresponding author. t Present address: Laboratory of Tumor Virus Genetics, Dana Farber Cancer Center, and Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA 02115.
3364
VOL. 62, 1988
3365
SV40 REVERTANT ENHANCERS AND HOST RANGES
to transfection at 1
x 106 CV-1 cells, 1 x 106 HeLa cells, and 3 x 106 NIH 3T3 cells per 100-mm-diameter plate in 10% fetal bovine serum in Dulbecco modified Eagle minimal medium (DME) with 100 U of penicillin per ml and 100 jig of streptomycin per ml. Cells were fed with fresh medium 3 to 6 h prior to transfection and were transfected with a total of 20 ,ug of plasmid DNA; for CV-1 and HeLa cells, 2 ,ug of ITSVHHI3A128 experimental plasmid, 1 ,ug of rrSVHPa2, and 17 ,ug of pUC119 DNA were used, and for NIH 3T3 cells, 6 or 8 ,ug of TrSVHHPA128, 2 ,ug of irSVHPa2, and 10 or 12 ,ug of pUC119 DNA were used. Each transfection included a control to check for saturation of ,-globin expression which contained half the concentration of irSVHHPA128 2x72 plasmid DNA; this sample gave 40 to 60% the level of expression of the undiluted sample, showing that the transfection was not performed in ,-globin plasmid excess. For transfection, the DNAs were diluted into 0.5 ml of 250 mM CaCl2-7.5 mM Tris chloride (pH 8.0)-0.075 mM EDTA, mixed dropwise into 0.5 ml of 50 mM sodium HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) (pH 7.1)-280 mM NaCl-1.5 mM sodium phosphate, and after 2 min at ambient temperature, were added to the cells. After 16 to 20 h, the cells were washed twice with DME and refed with DME containing 10% fetal bovine
serum.
Cytoplasmic RNAs were isolated 55 to 60 h after initiation of the transfection for CV-1 cells and 48 h for HeLa and NIH 3T3 cells, by lysis of the cells in 1 ml of 0.65% Nonidet P-40150 mM NaCl-50 mM Tris chloride (pH 8.0). After removal of nuclei by centrifugation, the RNAs were added to 1 ml of 7 M urea-1% sodium dodecyl sulfate-350 mM NaCI-10 mM Tris chloride (pH 7.4)-10 mM EDTA and extracted twice with phenol. Following two precipitations with ethanol, contaminating DNA was digested with 0.4 ,ug of DNase 1 in 200 ,ul of 50 mM Tris chloride (pH 7.6)-0 mM MgCl2 for 20 min at 37°C, followed by protease digestion for 30 min at 37°C by addition of 100 RId of 0.9 M sodium acetate (pH 5.3)30 mM EDTA-0.6% sodium dodecyl sulfate-150 ,ug of proteinase K per ml. The samples were prepared for hybridization by one extraction with phenol, ethanol precipitation, and suspension of 50 R1 of H20. One-fourth of the RNA was probed with single-stranded SP6 RNA polymerase-generated RNA probes as described previously (20, 46). The RNase A- and RNase T1-protected RNAs were fractionated by electrophoresis through 8 M urea-6% polyacrylamide thin sequencing gels. To produce the autoradiograms shown in Fig. 3, the levels of a-globin RNAs were equalized after examination of the autoradiogram of the first gel, and the remaining sample was fractionated on a second polyacrylamide gel. The levels of a- and ,B-globin RNA expression were determined by scintillation counting. The levels of ,-globin RNA were first normalized to the a-globin internal control and then normalized to the 2x72 sample. RESULTS A matched set of revertant enhancers containing 71-bp duplications. The location within the SV40 early promoter and nucleotide sequence of the A, B, and C elements are shown in Fig. 1. Within the sequence of each element, sequence motifs (e.g., core, sph motifs) are identified and the three sets of double point mutations (dpm), called dpml, dpm2, and dpm6 are shown; these mutations inactivate the A, B, and C elements, respectively (19, 20). Numerous viral growth revertants, containing sequence duplications, have been isolated from SV40 viruses harboring different combi-
~72 bp
|
21bp
21bp |21bp
300
ori 0/5243
A
B
C A A
A
ITT
|GTGGAATG|TGTGTCAGTTAGGgtg cc
IT
TGGAAAGITCCCCA
C
T
B
GCAGAAGTIATGCAAA
A G x
-
ITT
CATGCA x X-
FIG. 1. Location and nucleotide sequence of the SV40 enhancer A, B, and C elements. The upper portion shows a diagram of the SV40 early promoter with the early-early transcriptional start sites shown to the right (-). From right to left are shown the position of the AJT-rich region which serves as a TATA box, two perfect (21bp) and one imperfect (21bp') 21-bp repeated sequence with the location of GC boxes (_) shown, and the 72-bp sequence that is tandemly duplicated in SV40 strain 776. The three stippled boxes labeled A, B, and C are aligned with the diagram of the SV40 early promoter to show the location of the A, B, and C elements. The lower portion of the figure shows the nucleotide sequence of each element as defined previously (19, 20). The A-element sequence in lowercase letters, gtg, identifies a 3-bp terminal redundancy in the sequence of this element. The boxed sequences are homologous to the core motif (24, 44) in the A and C elements and to the octamer motif (12, 31) in the B element. An imperfect 9-bp repeat (sph motifs I and II) within the B element is identified by the arrows with the letter X showing the location of mismatch. The base changes created by the dpml, dpm2, and dpm6 mutations are shown with circled numbers above the sequence of each element.
nations of the dpm mutations. These revertants were selected by transfection of large amounts of mutant viral DNA into CV-1 cells, followed by passage and purification of the revertant viruses by plaque formation. In every case tested, these growth revertants have restored enhancer potential. Revertants have been isolated from the double mutants dpml2 (20), and dpml6 and dpm26 (8). One particular 71-bp revertant duplication (rd) arose multiple times during the isolation of these revertants. Figure 2 shows the structure of these revertants, along with the wild-type SV40 enhancers containing either one (lx 72) or two (2 x72) copies of the 72-bp sequence. The 1 x 72 enhancer contains one copy of each of the A, B, and C elements; the 2x72 enhancer contains two copies of both the B and C elements and a single copy of the A element. Revertants dpm26 rd71 and dpml2 rd7l contain duplications that span identical regions within the enhancer, creating two copies of the wild-type A element in dpm26 rd7l or two copies of the wild-type C element in dpml2 rd7l (Fig. 2). The 71-bp duplications in dpml2 rd7l and dpm26 rd7l do not duplicate the four ori-proximal B-element nucleotides that are mutated by the dpm2 mutations (Fig. 1). In dpml6 rd7l, the 71-dp duplication is shifted by 2 bp to encompass all but the terminal dinucleotide CA in the previously defined B element (19). The junction of the dpml6 rd7l duplication replaces the B-element nucleotides CA by TG. It is not evident whether the TG dinucleotide can functionally compensate for the loss of the CA sequence, or whether the boundaries of the B element can be refined to a sequence lacking the terminal two nucleotides. Whichever is the case,
SHEPARD ET AL.
3366
J. VIROL. A C r---
X72
C
A 2X
I
B
rd7I
dpml2 rd71
---
B
A
A dpm26
C
B
I1
- 2 21bp-2
:
C C r21i|@ bpl2bp
B dpml6 rd71
F
F-7,77r,77r
B
mI
F-)7 1
dpml26 rd7I
b 2bp
FIG. 2. Structure of wild-type and revertant SV40 enhancers. The structure of each enhancer is shown with the 21-bp repeats positioned
to
the
right. Mutated
elements
( El) and wild-type
elements (LIZ) are shown. The black arrows identify the sequences involved in the 72-bp duplication of wild-type SV40 strain 776 (2x72) and the 71-bp duplications of the rd7l series of revertants. The 71-bp duplication of dpml6 rd7l is shifted 2 bp to the right with respect to the identical duplications of the dpm26, dpml2, and dpml26 rd'71 revertant enhancers.
the 2-bp shift in the dpml6 rd7l duplication compared with dpml2 and dpm26 rd7l is consistent with a requirement for duplicated B-element function to restore activity in the
dpml6 mutant. In addition to the dpml2 and dpm26 rd7l duplications, the same 71-bp duplication arose during isolation of revertants of the triple mutant dpml26 (19; Fig. 2). dpml26 rd7l contains all three of the A-, B-, C-element mutations and consequently is a relatively weak enhancer (19). The dpml26 rd7l revertant serves as a comparative base line with which to determine the respective contributions of the remaining pair of wild-type elements in the dpml2, dpml6, and dpm26 rd7l revertants. These closely matched revertant SV40 enhancers offer several advantages to the analysis described here. (i) Among the revertants tested in each class, except for dpml26 rd7l, these revertants exhibited the best viral growth and enhancer function in CV-1 cells (8, 20). (ii) Because of the identical size and similar sequence of the rd7l duplications, any differences in the activity of the associated enhancer reflect the effect of the point mutations and duplicated wild-type elements. Furthermore, (iii) these enhancers were all selected in vivo as SV40 variants and thus represent authentic SV40 enhancers that have evolved to compensate for deleterious mutations. Assay of the 71-bp revertant enhancer duplications in CV-1, HeLa, and NIH 3T3 cells. To assay the activity of the rd7l collection of revertant SV40 enhancers separately from the entire SV40 early promoter, the ability of revertant enhancers to enhance transcription of the heterologous human ,-globin gene was assayed after transient expression. The levels of ,-globin gene expression were measured by using an RNase protection assay of an internally labeled complementary RNA probe generated by SP6 RNA polymerase. By using this assay, the levels of correctly initiated ,-globin RNA can be discriminated from RNAs arising from other initiation sites. Figure 3A shows the structure of the P-globin
gene expression plasmid TrSVHHPA128 (20, 38) used to assay the various enhancers; in this construction, the revertant enhancers are positioned within the SV40 control region but 1.2 kilobases upstream (2.2 kilobases downstream) of the 3-globin transcriptional initiation site. The P-globin constructs were transfected by calcium phosphate coprecipitation, along with a plasmid carrying the human a-globin gene (Fig. 3A, bottom), which served as an internal control for transfection efficiency. Figure 3B through D shows the results of assay of the matched set of 71-bp revertant enhancer duplications in three separate cell lines: CV-1 cells, the African green monkey kidney cell line used for the SV40 revertant isolations; HeLa cells, a human epitheloid cervical carcinoma cell line; and NIH 3T3, a mouse embryo fibroblast cell line. The correctly initiated and processed a- and P-globin RNAs protect 132 and 350 nucleotides of the respective RNA probes. Table 1 shows the quantitation of the relative levels of correctly initiated P-globin RNA from two (NIH 3T3) or three (CV-1 and HeLa) separate assays. In these assays, there are two incorrect P-globin transcripts (labeled itl and it2 in Fig. 3). The more prominent itl product is apparently the result of readthrough transcription; the production of a shorter protected fragment by this readthrough transcript, rather than the expected longer fragment of nearly probe size, is due to the splicing of the readthrough RNA at a cryptic splice acceptor about 30 nucleotides downstream of the P-globin transcriptional initiation site (11). This itl readthrough transcript could result from initiation at (i) the SV40 early promoter (Fig. 3A), (ii) a cryptic promoter, or (iii) the ,-globin promoter itself if the RNA polymerase can tran-
scribe the entire plasmid. We have not discriminated bepossibilities. The itl transcript is particularly prominent after transient expression in NIH 3T3 cells (Fig. 3D) and appears to be enhancer independent. Although the origin of this RNA is unknown, it does not affect the interpretation of the results in NIH 3T3 cells, because it can be separated from the correctly initiated P-globin RNAs. In each of the three cell lines shown, the wild-type 2x72 enhancer-containing plasmid (Fig. 3B through 3D, lanes 7) produces 50- to 100-fold more 3-globin RNA than the enhancer deletion plasmid ppenh-(lanes 1), indicating that the P-globin gene responds equally to the SV40 enhancer in these three cell lines. The levels of P-globin RNA produced by the 1 x 72 wild-type SV40 enhancer are three- to fourfold less than those produced by the 2x72 enhancer in all three cell lines (compare lanes 6 and 7 in Fig. 3B through D; see Table 1). In contrast to these wild-type SV40 enhancers, not all of the 71-bp revertant duplications of the dpml2, dpml6, dpm26, and dpml26 mutants function similarly in each of the three cell lines. As reported previously, these four enhancers can stimulate ,-globin gene expression to various degrees in CV-1 cells (8, 19). The dpml2 rd7l enhancer is nearly as active as the wild-type 2x72 enhancer, whereas the dpml6 and dpm26 rd7l enhancers direct somewhat less P-globin expression but still more than the 1 x72 enhancer (Table 1). All three enhancers are four- to eightfold more active than the dpml26 rd7l enhancer; thus in CV-1 cells, duplication of either of the wild-type A, B, or C elements can compensate for the loss of function in the others. The weak but evident activity of the dpml26 rd7l revertant suggests that either there are additional weak elements within the SV40 enhancer that have gone undetected, or the junction formed by the 71-bp duplication creates an active element (e.g., the B element [Fig. 2]). If the latter possibility is indeed the case,
tween these
VOL. 62, 1988
SV40 REVERTANT ENHANCERS AND HOST RANGES
AQ
3367
D.
C'
rd 71
rd71 rd 7
rd 71 (0
E PE E E a aL a
ci, -0 -a -o
q
a0x _0
m
_
-
ma
FQ X
I
c
c
O
\
\
x Q
-
de-
6 7 8
3 4
V
O
Ca
-'5
L 6 HeLa
7 8
XX
X _V
Wm-
-
2 3H4
cv-I
E E E
C} V -
soo s0 0~~~
aE
-
2
J N
Cu -E -F E -E r, r0-Q CaQ Qx X
E
-
04gI
- tI
2 3 4 5/ 6 7
NIH/3T3
FIG. 3. Assay of wild-type and revertant SV40 enhancer function in CV-1, HeLa, and NIH 3T3 cells. (A) Structure of the experimental 3-globin and internal reference a-globin plasmids. These plasmids have been described previously (20, 38). Symbols: ITVX vector -,
sequences;
(_)
exon
,
globin
sequences
gene sequences, above which
linked
by intron
are
shown the
(A). The
sequences
transcriptional initiation site (-) and untranslated (
SV40
sequences
( EZ)
are
shown
along with
the
) and translated
location
of the
early
transcriptional initiation site (-), and 21 bp (W) and 72-bp repeats fI. The SV40 enhancer sequences contained in the wild-type and revertant ,-globin plasmids extend from the ori-proximal Hindlll site to the unique SV40 HpaII site. In the enhancerless control plasmid (enh-), the enhancer sequences (SphI-HpaII) are missing. Kb, Kilobase. (B, C, and D) The results of expression assays in CV-1, HeLa, and NIH 3T3 cells are shown with the correctly initiated a- and 3-globin RNAs, labeled a and lB, respectively, and two other ,-globin transcripts labeled itl and it2. The assays are the result of the RNase protection of SP6 RNA polymerase-generated single-stranded a- and P-globin RNA probes after hybridization to total cytoplasmic RNA isolated from transfected cells as described in the Materials and Methods. The exact experimental P-globin plasmid used in each transfection is shown at the top of each panel. The Mock samples lacked either a- or P-globin plasmid DNAs in the transfection. The series of light bands in panel C just below the it2 transcript is derived from the a-globin probe.
it offers one explanation for the frequent selection of this revertant duplication. The relative activity of the rd7l revertants in HeLa cells is similar to their activity in CV-1 cells, except that the dpml6 and dpm26 rd7l enhancers are twofold less effective in HeLa cells than CV-1 cells compared with the 2x72 enhancer. TABLE 1. Relative activities of revertant SV40 enhancers in CV-1, HeLa, and NIH 3T3 cells Relative enhancer activity'
Enhancer CV-1
2x72 1x72 dpml2 rd7l dpml6 rd7l dpm26 rd7l dpml26 rd7l enh-
0.30 0.95 0.47 0.54 0.12 0.03
1 ± 0.10 ± 0.11 ± 0.06 ± 0.07 ± 0.03 ±
