VIROLOGY
184,795798
( 1991)
Quantitative
Detection
of Spliced E6-E7 Transcripts of Human Papillomavirus in Cervical Premalignant Lesions
HIROSHI SHIRASAWA, Department
HIDEKI TANZAWA,
of Microbiology, Received
School March
TADASHI MATSUNAGA,
of Medicine, 6, 199 1; accepted
Chiba June
University,
Type 16
AND BUNSITI SIMIZU Chiba
260,
Japan
72, 199 1
The splicing patterns of E6-E7 transcripts of human papillomavirus type 16 ( HPVl8) in cervical premalignant lesions were quantitatively analyzed by Sl nuclease protection assay. The major E6-E7 transcripts in HPV16-containing cervical lesions (four cervical intraepithelial neoplasias and one invasive carcinoma) were from spliced E6’l/ E7 mRNA. The unspliced E6/E7 mRNA, which can encode the full-length zinc finger protein E6, is expressed as 8 to 15% of E6-E7 transcripts. The spliced E6’ll/E7 mRNAs were expressed as 14 to 24% of E6-E7 transcripts in most tissues. However, in HPVl6-containing cell lines, the expression levels of spliced and unspliced E6-E7 transcripts were variable. CD1991 Academic
Press, Inc.
Human papillomaviruses (HPVs) are heterogeneous small DNA viruses that are divided into two groups, those associated with lesions of mucosal origin and those mainly found in cutaneous lesions ( 1). HPVs associated with malignant mucosal lesions, such as HPV16(2), HPV18(3), HPV31 (4), HPV33(5), HPV35 (4), and HPV56 (4), possess possible introns within the E6 open reading frame (ORF). These HPVs have the potential to express different E6 proteins, E6 *s (6)) by alternative splicing. In contrast, with the exception of HPV43 (7), HPVs, such as HPV6 and HPVl 1, associated with benign lesions cannot express E6* proteins (6). Spliced E6-E7 transcripts of HPV16 and HPV18 have been found in cancer-derived HeLa, SiHa, and CaSki cells, and in HPV16containing cervical cancer tissues (6, 8, 9). Major E6-E7 transcripts of HPVl6 in cervical carcinoma tissues and cell lines contain one of the introns within the E6 ORF (E6*I/E7 or E6*ll/E7 mRNA) (8, 9). The uniformity in the splicing pattern of the E6-E7 transcripts in cervical premalignant lesions and carcinomas has been shown by RNA polymerase chain reaction (PCR) ( 10). However, the ratios of the alternatively spliced E6-E7 transcripts in carcinoma tissues and cervical carcinoma cell lines seem to be variable (8, 9). It is of interest whether these ratios are also variable in premalignant tissues; however, due to the difficulty of recovering sufficient amounts of RNA from the often very small premalignant lesions, the quantitative aspects of this alternative splicing have not yet been determined. Several problems are associated with the PCR technique for quantitative analysis ( 10). Therefore, to quantitatively determine the expression levels of the spliced and unspliced E6-E7 transcripts in cervical tissues, we employed si-
multaneous extraction of DNA and RNA from cervical tissues ( 1 l), followed by Sl nuclease protection assays using homogeneously labeled probes which allow high sensitivity detection. The splicing patterns of E6-E7 transcripts of HPV16 were evaluated by analyzing the E6-E7 transcripts in the cervical carcinoma cell lines QG-H, QG-U ( 12), and HPVl6-transformed mouse 1 OT1/2 cells, GT3 ( 13). Probes were prepared by in vitro synthesis of antisense single-stranded DNA using sense singlestranded DNA as templates which were cloned in the multicloning sites of Ml 3mpl8 or Ml 3mpl9, as described previously ( 14). Homogeneously labeled probes were synthesized with the Klenow fragment in the presence of [32P]-dCTP. The synthesized DNA was cut with a unique restriction enzyme, separated on a 5% acrylamide gel, then eluted from the gel, and recovered. The probes contained primer and polylinker sequences at both ends, in order to differentiate the undigested and fully protected Sl bands. The regions used as probes are shown in Fig. 1A. St nuclease protection assays of RNAs from QG-H and QG-U with probe a, which contains the P97 promoter and the splice donor site within the E6 ORF (Fig. 1A), showed that major mRNAs colinear with the E6 ORF in these cells are transcribed from the P97 promoter site and spliced at the donor site, nt 226, as described previously (9) (Fig. 1 B, probe a). The presence of E6*I transcripts in both QG-H and QG-U was revealed by Sl nuclease protection assays using probe b, which covers the P97 promoter, the splice donor site at nt 226, and the acceptor site for E6 *I at nt 409 (Fig. 1 B, probe b). When the 130-nt band representing E6*I or E6*II and the 96-nt band representing 795
0042.6822/91
$3.00
CopyrIght Q 1991 by Academrc Press, Inc All rights ot reproduction 1” any form reserved
796
SHORT
COMMUNICATIONS
A
544 65
556
E7
EB
I P97
+
226
I
I Acl
Do
I+
probe a 5*
303
AC2
+
+
409
526
201
wotected
r’ragmenta
probe b
R55
130
58
504 403
rotected
Fragments
130
probe c
96 ma
RRO
577 472 355
B MHU
probe c
probe b
probe a
PHU
t P 408-
t M
PtMHUG 577472-
-207
FIG. 1. Sl nuclease protection assay of HPV16 E6-E7 transcripts in cell lines. (A, top) Map of the E6-E7 region of the HPV16 genome indicating the positions of the P97 promoter, the splice donor site (Do), and the acceptor sites (Acl and Ac2 for E6*I and E6*ll, respectively). (A, bottom) Probes used for Sl nuclease protection assay. The Sl nuclease protected fragments are shown, respectively, below probes a, b, and c. (6) Sl nuclease protection assay of RNAs from cell lines QG-H (lanes H), QG-U (lanes U), HPVl B-transformed mouse 1 OT1/2 cells, GT3 (lane G), and tRNA (lanes t) using probes (lanes P) a, b, and c. The size markers (lanes M) were 3’-end-labeled fragments of Hpall-digested pBR322. The sizes of the markers, in nucleotides, from the top are 623, 528, 405, 308, 243, 239, 218, 202, 191, 181, 161, 148, 124, 1 11, 91, 77, and 68. The sizes of the protected fragments, in nucleotides, are given by horizontal numbers.
SHORT TABLE
COMMUNICATIONS
1
EXPRESSION LEVELS OF HPV16 TRANSCRIPTS COLINEAR WITH E6 ORF IN GIN, INVASIVE CARCINOMA, AND CELL LINE’ Expression Tumor or cell line Cl c2 c3 c4 IC Cell line QG-H QG-U GT3
levels
of E6 transcripts
E6
E6*I
E6*ll
15 8 8 8 11
85 76 68 74 76
ndb 16 24 18 14
8 20 44
67 80 56
25 nd nd
a Expression levels are expressed as a percentage of the total transcripts colinear with the E6 ORF. The values were evaluated by an image analyzing system PIP-4000 (ADS, Japan) based on autoradiographic images. The values were normalized by the length of the protected band. b nd; not detected.
E6*I were compared, the relative intensity of the 130nt protected fragment in QG-H was stronger than that in GIG-U, indicating more E6*ll mRNA was transcribed in QG-H than in QG-U (Fig. 1 B). To determine the relative amounts of these spliced and unspliced E6-E7 transcripts, probe c, which covers the splice acceptor sites within the E6 and E7 ORFs (Fig. 1A), was used for Sl nuclease protection assays of RNAs from QG-H, QG-U, and HPVl6-transformed mouse 1OT1/2 cells, GT3. In QG-H and QG-U cells, as shown in Fig. 1 B, the 472-nt bands representing the E6*I/E7 mRNAs were the most prominent, but in GT3 cells the intensities of the 472- and the 577-nt bands were almost equal, indicating that unspliced E6 mRNA and E6*I mRNA were expressed at nearly equivalent levels in GT3 cells. E6*ll/E7 mRNA was easily detected in RNA from QGH, and barely detected in RNAs from QG-U and GT3 (Fig. 1 B, probe c). To estimate the relative expression levels of these unspliced and spliced transcripts, the radioactivities of Sl bands were evaluated by an image analyzing system PIP-4000 (ADS, Japan) after reading the autoradiographic image by a photodensitometer PAD-65 (Sakura, Japan). As summarized in Table 1, the expression levels of each transcript colinear with the E6 ORF were variable in these cell lines. Further analyses for the HPV16 transcripts in cervical lesions were performed using only probe c due to the limited amount of specimens. To analyze the splicing patterns of E6-E7 transcripts of HPV16 in cervical tissues, DNAs from 18 cervical lesions of ClNs and invasive carcinomas were
797
screened for the presence of HPV16 by Southern blot analyses using HPV16 DNA as a probe. As shown in Fig. 2A, four ClNs and one invasive carcinoma were found to contain HPV16 genomic DNA judging from its typical fstl restriction fragment pattern. In C4, an extra band between the 2.8- and the 1.8-kb fragments was observed. Further Southern blot analysis of the C4 DNA, which was undigested or digested with HindIll (an uncut enzyme for HPV16) or digested with BamHl (a single-cut enzyme for HPV16), showed that most of the HPV16 genomic DNA existed as plasmid DNA and only a small percentage existed integrated in the host DNA (data not shown). Simultaneously obtained RNAs from tissues that had been confirmed to contain HPVl6 DNAs were analyzed by Sl nuclease protection assay using probe c (Fig. 2B). The most prominent Sl bands in the ClNs and invasive carcinoma were the 472-nt bands representing the E6*I/E7 mRNA, while the 577~nt bands representing the unspliced E6-E7 transcripts were detected at very low levels. The E6*ll/ E7 transcripts were detectable in most cervical lesions except the CIN tissue Cl, in which other Sl bands were also faint. A 390-nt band was consistently observed between the 472- and the 355-nt St bands. This band did not seem to represent spliced transcripts from an intron within the E6 ORF, because no corresponding spliced transcripts were detected when RNA PCR was performed with QG-H RNA (data not shown) The expression levels of the transcripts colinear with the E6 ORF in these tissues are summarized in Table 1. These data showed no obvious difference in the splicing patterns of E6-E7 transcripts between premalignant lesions and the one carcinoma tissue tested. The presence of two spliced transcripts (E6*I and E6 *II) in cervical premalignant lesions and carcinomas has been shown by RNA PCR ( 10). In this study, the E6*ll/E7 transcripts were not detectable in several RNAs from tissues and cell lines by Sl nuclease protection assay. This could be due to the limited amount of RNA tested and the limited sensitivity of the Sl nuclease protection assay. However, it seems likely that RNA-derived E6 *II PCR product should have been easily amplified due to its relatively small size compared to the other E6-E7 transcripts. When cDNAs of E6”l and E6*ll mRNA in QG-H cells were amplified by PCR, the amounts of the PCR products did not reflect the actual amount of each mRNA. The amount of E6*ll mRNA seemed to be overestimated by PCR (data not shown). In addition, it is difficult to exclude the possibility that PCR products corresponding to unspliced transcripts might be generated from trace DNA contamination in the RNA specimens. The expression levels of alternatively spliced transcripts in premalignant lesions were not variable as
798
SHORT
A H Cl C2
COMMUNICATIONS
6
IC C3 C4
P
t
MC1
P
t
M C2 IC
H
M H ICC3C4
1 -577 -472
-4 4 4
FIG. 2. Southern blot analysis and Sl nuclease protection assay of HPVl6-containing cervical tissues. (A) DNAs from cervical carcinomaderived cell line QG-H (lane H), cervical lesion ClNs (lanes Cl, C2, C3, C4), and one invasive carcinoma (lane IC) digested with Pstl, separated on a 1% agarose gel, and hybridized with 3ZP-labeled HPVl6-DNA. &&digested HPVl B-DNA fragments are indicated by arrowheads. The sizes of the fragments, in kb, from the top are 2.8, 1.8, 1.5, 1 .l, and 0.5. An extra band observed in the C4 is indicated by an arrow. (B) Sl nuclease protection assay of simultaneously extracted tissue RNAs from cervical lesion ClNs (lanes Cl, C2, C3, C4) and an invasive carcinoma IC (lane IC), and RNAs from QG-H (lanes H) and tRNA (lanes t) by using probe c (lanes P). The size markers (lanes M) are the same as in Fig. 1. The sizes of the protected fragments, in nucleotides, are given by horizontal numbers.
were the expression levels in cell lines. Since HPVlG genomic DNA in the cervical tissues tested in this study seemed to exist mainly as plasmids, the variable ratios of alternatively spliced transcripts in the cell lines tested in this study, as well as in SiHa, CaSki cells, and carcinoma tissues (8, 9), may be due to the unregulated transcription of HPV16 genome integrated into host DNA. ACKNOWLEDGMENT We thank Mrs. N. Motoji. Japan, for the image analysis
Institute of Whole of autoradiography.
4. 5. 6. 7. 8. 9. 10.
Body
Metabolism, 11.
REFERENCES 12. DE VILLIERS, E. M., J. Wol. 63, 4898-4903 (1989). 2. SEEDORF. K., KR&MMER, G.. DORST, M., SUHAI, S., and R~WEKAMP, W. G., virology, 145, 181-185 (1985). 3. COLE, S. T., and DANOS, O., J. Mol. t?io/. 193, 599-608 ( 1987). 1.
13. 14.
GOLDSBOROUGH, M. D., DISILVESTRE, D., TEMPLE, G. F., and LORINCZ, A. T., virology, 171, 306-311 ( 1989). COLE, S. T., and STREECK, R. E., J. L/ire/. 58, 991-995 (1986). SCHNEIDER-G~~DICKE, A., and SCHWARZ, E., EMBO 1. 5, 22852292 (1986). LORINCZ, A. T., QUINN, A. P., GOLDSBOROUGH, M. D., SCHMIDT, B. J., and TEMPLE, G. F., /. viral. 63, 2829-2834 (1989). SMOTKIN, D., and WE‘ITSTEIN, F. O., Proc. Nat/. Acad. Sci. USA 83,4680-4684 (1986). SMOTKIN, D.. PROKOPH, H., and WETSTEIN, F. 0.. J. Viral. 63, 1441-1447 (1989). CORNELISSEN, M. T. E.. SMITS, H. L.. BRI~T, M. A., VAN DENTWEEL, J. G., STRUYK, A. P. H. B., VAN DER NOORDAA, J., and TER SCHEGGET, J., /. Gen. Viral. 71, 1243-1246 (1990). SHIRASAWA, H., TOMITA, Y., KUBOTA, K., KASAI, T., SEKIYA, S., TAKAMIZAWA, H.. and SIMIZU, B., J. Viral. 62, 1022-1027 (1988). SHIRASAWA, H., TOMITA, Y., SEKIYA, S., TAKAMIZAWA, H., and SIMIZU, B., J. Gen. l/irol. 68, 583-591 (1987). TANZAWA, H., SHIRASAWA, H., SATO, K., and SIMIZU, B., Chiba Med. J. 66, 327-333 ( 1990). BURKE, J. F., Gene 30, 63-68 (1984).