(TP1 and LMP) in lymphocyte and epithelial cells

© 1991 Oxford University Press Nucleic Acids Research, Vol. 19, No. 9 2435

Differential expression of Epstein Barr viral transcripts fortwo proteins (TP1 and LMP) in lymphocyte and epithelialcells

Paul R.Smith and Beverly E.GriffinDepartment of Virology, Royal Postgraduate Medical School, Du Cane Road, London W12 ONN, UK

Received January 10, 1991; Revised and Accepted April 10, 1991

ABSTRACT

Studies presented here show that some functions ofthe human herpesvirus, EBV, may be transcriptlonallydifferentially expressed in two cell types which carrythe same (C15) Isolate of this virus. Of the 'latent' viralfunctions investigated, only one (TP2) of theepisomally-speclflc genes that encode terminalproteins (TP1 and TP2) is found to be expressed In theC15 epithelial cell tumour environment, whereas bothare transcribed—as different, but related,messengers—in a B-cell line generated with virus fromthe C15 tumour. The other gene investigated Is that forlatent membrane protein (LMP), which is found in thesame region of the EBV genome but on the oppositestrand. This gene, apparently transcriptionally silent inB-cell (Burkitt's) lymphomas, is expressed in the C15epithelial tumour, as well as in other nasopharyngealcarcinomas investigated. Promoter usage in thecarcinomas and B-cells appears, In some cases at least,to be cell-type specific. Expression may also begoverned by methylation since a chromosomally silentregion In the carcinoma (that encompassing TP1) ishighly methylated on CpG residues, whereas the activeregion (encoding TP2 and LMP) is virtually free of suchmethylation. Our data suggest that there may beselective transcriptional regulation of EBV genes in thetwo types of cells investigated. Thus, it may beunnecessary to invoke different virus genotypes toaccount for the two distinct malignancies—Burkitt'slymphoma and nasopharyngeal carcinoma—associatedwith EBV.

INTRODUCTION

One of the more unusual properties of the human herpesvirus,Epstein-Barr virus (or EBV), is the fact that it is associated witha variety of different diseases. These include infectiousmononucleosis—where a causative role has been assigned—andtumours of both B-lymphocytes and epithelial cells (reviewed,ref. 1). The best characterised of the B-cell malignancies is thecommon tumour of children in southern Africa, known asBurkitt's lymphoma (BL). In many Asian adult populations, EBVhas been implicated in the genesis of a poorly differentiatedcarcinoma of the nasopharynx. The precise biology of EBV with

regard to disease is uncertain, although it is clear from the abovethat the virus is able to infect and replicate in at least two distinctcell populations. Another feature of the EBV life cycle is alsoits ability to maintain a silent infection in its host, existing ina so-called 'latent state'; the precise site(s) of latency in vivo isas yet unknown. Description of latent gene expression has reliedupon studies with B-lymphoblastoid cell lines, where a restrictedset of EB viral genes encoding a number of nuclear proteins(EBNAs 1 -6) and three putative membrane proteins (LMP, TP1and TP2), as well as two small RNA species (EBERs), areexpressed (reviewed, ref. 2). Numerous restriction enzymeanalyses of EBV DNAs from various sources have been carriedout in an attempt to determine whether genetic diversity amongEB virions could account for the different host responses observedfollowing infection with this virus (3, 4). But, to date, there isno compelling evidence that this is the case. Although, notsurprisingly for such a large virus (-175,000 base pairs),alterations are observed among viral genomes from differentsources (5), none of these have been directly related to the courseof disease.

Several years ago we began to look into the likelihood thatEBV gene expression in different cellular environments mightvary and possibly be related to the pathologies associated withthis virus. Our first studies on this topic showed that an area ofEBV DNA not hitherto implicated in altering the properties ofB-cells—contained within a cosmid recombinant DNA,p31—could induce cellular immortalization (continuousproliferation) of primary epithelial cells in culture (6, 7). Then,in a comprehensive study of viral gene expression in anasopharyngneal carcinoma (NPC) passaged in athymic mice,we observed that transcriptional expression in this tumour wasmarkedly different from that previously reported for cell linesestablished from BL patients (8). Similar results were obtainedin biopsies from NPC patients (9). In particular, high levels oftranscription were observed from the region of EBV capable ofin vitro immortalization of primary epithelial cells. We and othersshowed that CpG methylation patterns over the region implicatedin expression of early viral nuclear antigens (EBNAs) alsodiffered in the two cell types and might be responsible for, orat least related to, the observed differential gene expression (10,11). Here we have examined the transcription of part of thegenome known to be involved in expression of a latent viralfunction(s)—initially designated terminal protein (or TP) (12)—in

Downloaded from https://academic.oup.com/nar/article-abstract/19/9/2435/2387351by gueston 18 February 2018

2436 Nucleic Acids Research, Vol. 19, No. 9

NPC tumours propagated in nude mice (13), and compared thedata with those obtained from a B-lymphoblastoid cell linegenerated from virus 'rescued' from one of these tumours (8,14) and with expression of another function designated latentmembrane protein (or LMP) which appears to be transcriptionallysilent in Burkitt's lymphomas (1).

TP is an unusual gene since, being encoded by informationfound in both extremes of a linearized viral genome (such as existsin EB virions), it can only be expressed off circular (episomal)forms of EBV or off tandemly repeated copies of the viralgenome. (There is no evidence that the latter exist.) The structureof two TP transcripts has been elucidated (12, 15) and recently,two differently-sized TP proteins have been demonstrated to existin B-cells (16, 17). We have found the region encoding the TPgene to be transcriptionally active in the NPC tumours examined,as well as in a B-cell line established with rescued virus fromone of the tumours. Our studies show that of the two TPtranscripts observed in the B-cell environment, however, onlyone of them (TP2) is found in the carcinomas propagated in nudemice, even though the viral genome in the tumours is presentas episomes and both promoters potentially are capable of beingexpressed. Further, differential expression can be shown to becorrelated with methylation of one of the two promoter regionsin the tumour. Investigation of the control of TP expression iscomplicated by the fact that the TP2 promoter region overlapsthat of LMP, although the latter lies on the complementary DNAstrand. The possible relationship between TP and LMP expressionis discussed.

MATERIALS AND METHODSCellsC15 and C17 are human nasopharyngeal tumours of NorthAfrican origin which have been passaged in nude mice (13, 18);NAD-C15-STO-B (hereafter referred to as NAD-C15) is a B-lymphoblastoid cell line derived from co-cultivation of C15 cellswith adult human B-cells (14, 19), M-ABA is a lymphoblastoidcell line derived from an NPC patient, B95-8 is an EBV-infectedmarmoset cell line, permissive for viral replication, and Raji isan EBV positive Burkitt's lymphoma cell line. All cell lines weregrown in RPMI1640 medium supplemented with 10% foetal calfserum and 2mM glutamine. Biopsies from nasopharyngealcarcinoma patients were obtained as described elsewhere (9).

DNA clonescDNA clones encoding TP were isolated from a C15 Xgt 10cDNA library (8). Recombinant plaques were screened aspreviously described (8); plaques positive when probed with theEBV EcoRl I fragment (20) were subcloned and sequencedaccording to standard protocols.

RNA extraction and Northern blottingTotal RNA was isolated by the guanidinium/caesium chloridemethod (21). Poly(A)+ RNA was selected by chromatographyon two sequential oligo (dT)-cellulose columns (Pharmacia) asdescribed by the manufacturer. For Northern blotting analysis,poly A+ RNA (5/ig) was loaded in each lane of a 1 % agarosegel containing 0.7 % formaldehyde and 5mM iodoacetamide, andsubjected to electrophoresis. RNA was transferred to Zetaprobemembrane (Bio-Rad) by capillary transfer and all subsequenthybridization steps performed according to the manufacturers'

instructions. DNA used to probe Northern blots was labelledusing the Amersham multiprime kit (Amersham pic).

Oligonucleotides for Polymerase Chain Reaction (PCR)A shared TP1/TP2 20-long oligonucleotide (TP1,2-R) 5'-TTT-CCTTTGTGCAGCGGCAT (from exon 2 of the TP genes),position 232-212 on the EBV B95-8 genome (22), and specificoligonucleotides for TP2 (exon 1) 5'-AGCGGCAGTGTAAT-CTGCAC (designated TP2-L), position 169819-169839 and forTP1 (exon 1) 5'-CGCCTTATGAGGACCCATAT (designatedTP1-L), position 166739, were synthesised using an AppliedBiosystems oligonucleotide synthesizer. With theseoligonucleotides, the predicted sizes of the PCR products fromthe respective mRNAs should be 351 bp for TP1 and 264 forTP2. Oligonucleotides used for PCR of LMP were synthesisedfrom exon a, 5'-CTTGTCCTCTATTCCTTTGCT, position169318 to 169297, and from exon c, 5'-GTCTGCCCTCGGT-TGGAGT, position 168651 to 168638; these should result in aPCR fragment of 513 bp if derived from spliced mRNA or 680bpif from unspliced RNA or DNA (23).

PCROligonucleotide TP1.2-R was hybridised to poly A+ RNA (1/ig)in a volume of 20/il containing 2/xl of 10 x reverse transcriptasebuffer (250mM Tris/HCl pH 8.3, 25mM MgCl2, 1.25mM ofeach deoxynucleotide dATP, TTP, dCTP, dGTP, 20mM DTTand lOOmM KC1), and 10 units of RNasin (Promega).Hybridization was carried out at 60 °C for 20 minutes and thereaction cooled in ice for 2' prior to the addition of 12 units ofreverse transcriptase. The resulting solution was incubated at37°C for 60 minutes. The cDNA generated was divided into twoaliquots and the oligonucleotides TP1-L and TP2-L, specific forTP1 and TP2 transcripts, respectively, were used in conjunctionwith the common primer TP1.2-R, to amplify specific products.PCR analyses were performed in a volume of 50/tl containing10/xl cDNA, 5/*l 10X reaction buffer (200mM Tris/HCl pH 8.3,15mM MgCl2, 25OmM KC1, 200/iM of each dNTP) and 5 unitsTaq polymerase. The PCR conditions used were 94°C for 30seconds, 55°C for 30 seconds and 72°C for 2 minutes, over 30cycles. PCR generated products were separated by electrophoresison a 1.2% agarose gel, transferred to Bio-dyne paper and probedat 60°C for 20' using standard methods.

Analysis of DNA methylationTotal cellular DNA was digested with either HpaU or itsmethylation insensitive isoschizomer, Mspl, as described (10).In each case 20/ig of total cellular DNA was digested overnightwith lOunits enzyme//ig DNA. The digested DNA was separatedon a 1.2% agarose gel and transferred to nitrocellulose bycapillary transfer. The filters were then hybridised with 32Plabelled DNA probes corresponding to the putative promotersfor the two terminal proteins. For analysis of the TP1 promoter,a fragment with co-ordinates 166075-166502 from the EBVgenome was used as a probe, and for TP2 from 169477 -169848(22). The probes were generated by PCR from C15 tumourDNA.

RESULTS

Two transcripts have been identified with the expression ofterminal proteins. Data on these are summarised in Figure 1 andare derived from previously published observations in B-cells (12,


Nucleic Acids Research, Vol. 19, No. 9 2437

15) and, in part, from our own studies with the C15 NPC tumour(below). The transcripts use different promoters at the 'right hand'end of the conventionally-orientated EBV genome, as shown(Figure 1) and their first exons (exons 1) are not shared. In TP2,this exon is non-coding. The two transcripts, on the other hand,have identical 3'-exons (exons 2 -9) .

Analysis of cDNA clones isolated from a C15 XgtlO cDNAlibrary (8) showed sequences from both extremes of a linearisedEBV genome—such as would be found in virions—to be present.Since no linear forms of the DNA exist in the tumour, an obviousexplanation for the hybridization data was that TP transcriptsspanning the ends were being transcriptionally expressed. PutativeTP clones were therefore identified in the library by filterhybridization and the primary sequences of five of them weredetermined (24). [The only sequence alteration identified in theC15 clones, relative to that of the prototype B95-8 isolate DNA(22), involved a base change at position 78 in exon 2, whichwould result in a change from CCT (proline) to CTT (leucine)in a translation product.] The sequence data (not given) showedall of them to be consistent with the presence of only one of thetwo known transcripts of TP, that is, with the TP2 transcript(Figure 1). Of the five clones, three were spliced from the TP2exon 1 region. The other two which contained exons from the3' region of the transcript were unspliced 5' to exon 2. This couldreflect the usage of an alternative (third) promoter, but most likelyis a result of some unspliced mRNAs in the library, as notedearlier (8). No full-sized cDNA clones were isolated in this study;the largest clone analysed (designated D5) was 960bp andconsisted of a sequence whose 5' end was in exon 1 of TP2 andwhose 3' end was in exon 6 (see Figure 1). These findingssuggested that in the tumour either the region unique to TPl istranscriptionally silent or that the TP2 transcript is considerablymore abundant than that of TPl.

RNA extraction and Northern blottingNorthern blot analyses were carried out using (polyA+) mRNAfrom the C15 tumour and, as probe, the D5 cDNA clone (above).The data have been compared with results obtained using several

well-characterised EBV-positive B-cell lines, including the line,NAD-C15, established from virus induced from the C15 tumour.The results obtained are shown in Figure 2. In summary, in allthe B-cells analyzed—except Raji (track B) which has a deletionthat precludes expression of TPl transcripts (12; see Figure 1)—two TP transcripts were found. Quantitatively, however, noconsistent pattern was observed in B-cells. In some (see tracksC and D, Figure 2) the most prominent RNA was the larger ofthe two transcripts (corresponding to TPl, about 2.0 kb in length),whereas in others, the smaller transcript (TP2, about 1.7 kb) wasthe more prominent (see track A, Figure 2). In the nude mousetumours, C15 (track E), as in Raji cells, on the other hand, onlythe TP2 transcript was observed.

PCR AnalysisThe failure to find TPl transcripts in C15 (or the other NPCpropagated in nude mice, C17) by Northern blot analyses, orin a cDNA library, might be a genuine result or merely reflectthe fact that this message is one of very low abundance.

B C

I •TPl

Exon 1

I fcr -

B R aJ'deletion'&«#»>,

166kb 167

TP2 -,Exon I 2T4 16 i

rp*- \ I I I I I 1

c b a

LmP

Terminalrepeats

1 1 1168 169 170 0 1

3

2 3

—^_

4

9

5 6

Figure 1. Structure and locations of terminal (TP) and latent membrane (LMP)protein transcripts relative to the physical map of the EBV genome. A. Locationof TP exons (1—9) showing exons common to both TPs (exons 2—9) and thoseunique to TPl and TP2 (exons 1). The location of the cDNA clone D whichencompasses (part of) TP2 exon 1, exons 2 - 5 , and (part of) exon 6 is indicated( • ) . Also shown are exons (a-c) that comprise the LMP transcript. Promotersare indicated — ('rightward') or — ('leftward'). B. Part of the physical mapof the EBV genome (map units 160-6) as h would appear in a circular (episomal)form of the virus. The locations of the 3'- end of a deletion in the BL strain,Raji, and direct (terminaJ) repeats at the termini of linear viral genomes (whichexist in virions), relative to the EBV DNA sequence (in kb), are given.

Figure 2. Northern blot analysis of TP transcripts. The slower migrating of thebands (-2.0kb) identified using recombinant cDNA clone D (see Figure 1)corresponds to TPl transcripts and the faster migrating band (~ 1.7kb) to TP2transcripts. The tracks contain poly (A+) cellular RNA from A, B95-8; B, Raji;C, M-ABA; D, NAD-C15andE, C15 tumour. Two distinct transcripts are presentin tracks A, C and D, but only one (TP2) in tracks B and E. [In track C, thesizes of both TPl and TP2 appear to be larger than the corresponding messengerRNAs in B95-8 (track A), NAD-C15 (track D) or C15 (track E). Without completesequence data, the significance of this, if any, remains unclear.]



A B C D A B C D

BFigure 3. PCR amplification of TP transcripts. (Poly A+) RNA was hybridisedto an oligonucleoCide common to both TPl and TP2 and, in a second hybridization,to oligonucleotides specific either to TPl or TP2. The latter were selected suchthat a TPl amplified product should be 355bp and a TP2 product 264bp. A.Ethidium bromide stained bands of putative amplified products of TPl and TP2,respectively, of mRNAs from (two tracks) of A, C15 tumour; B, NAD-C15;C, Raji; D, B95-8, after separation by electrophoresis in agarose gels. B. Samples,as above, transferred to nitrocellulose and probed with 32P-labelled cDNA cloneD. PCR products for both TPl and TP2 are present in both tracks of B and D,but only that for TP2 in tracks A and C.

Therefore, PCR was utilised to assess whether there was lowlevel expression of TPl in the tumour. The data obtained, as givenin Figure 3, complement those obtained by the other methods,however. That is, the B-cell lines, B95-8 (panels B, lanes 2 and1) and NAD-C15 (panels D), show bands derived from both setsof oligonucleotide primers, whereas Raji (panels C) and C15(panels A) or C17 (not shown) gave a single band, correspondingto the use of an oligonucleotide specific for an exon from TP2but not to one using sequence from a specific exon in TPl. Similardata were also obtained with mRNA of one biopsy from anasopharnygeal carcinoma patient from Hong Kong. This wasnot the case with three other biopsies, however, where both TPland TP2 products were identified (data not given). [The data onprimary NPC biopsies cannot be assumed to represent expressionin epithelial cells, however, since these tumours (often referredto as lymphoepitheliomas) contain numerous invasivelymphocytes, some of which could be EBV-positive B-lymphocytes]. No PCR products were obtained in the control(the EBV negative B-cell line, Ramos) with either set ofoligomers.

Restriction Enzyme DigestionsThe simplest explanation for lack of TPl expression in C15 orother tumours is that—similar to the situation observed in Rajicells—a deletion is present in the viral genome which precludessynthesis of TPl. Restriction enzyme analyses carried out on totalchromosomal DNA from C15 and probed with viral sequencescorresponding to those deleted in Raji (see Figure 1) indicated,however, that there was no gross deletion in the C15 tumour inthis region of the genome to account for the experimentalobservations (data not shown). These findings are consistent,moreover, with the fact that the NAD-C15 cells, derived fromvirus 'rescued' from C15, express both TPl and TP2 transcripts.[Whereas it is conceivable that the EBV genome in NAD-C15is different from that in C15, this seems unlikely since sequencing

Msp I sites in B95-8 EBV genome

* TP1 Exon I

42130 195 81862

164 7211

667 140

165,000 166.000 167.000

TP1 Promoter Region

293 57 357 504 I 177 7635

169,000 170,000

TP2 Promoter Region

Figure 4. Location of CCGG restriction endonuclease sites in the region of EBVDNA that encompasses the promoters and (part of) the coding sequences of TPl,TP2 and LMP (between nucleotides 165-175Kbp). Potential HpaU and Msplcleavage sites are indicated for TPl (in A), similarly for TP2 and LMP (in B),by vertical lines, and their sizes [as deduced from the sequence of the B95-8viral genome (22)] given. The hatched areas show the locations of the probesused in Figure 5.

of other areas of the genomes, including that of a known highlypolymorphic region of EBV encoding the nuclear antigenEBNA-1 (9, and our unpublished work), has showed the twoDNAs to be indistinguishable.]

Methylation of TPl and TP2 promoter regionsAn alternative explanation for the failure of TPl transcriptionis that the corresponding promoter is transcriptionally silent inthe tumours. Earlier studies had shown tumour DNA to be highlymethylated in viral regions encoding promoters for some EBVgenes (notably EBNAs 2-6) (10, 11, 19). Methylation of CpGresidues in promoter regions has often been associated with afailure of these regions to be transcribed (25, 26). Therefore,utilisation of methylation sensitive and insensitive isoschizomers,such as HpaU (sensitive) and Msp I (insensitive), should givedata concerning the methylation status of promoter regions andan indication of the 'transcriptional status' of these promoters.To assess methylation of TPl and TP2 promoter regions amongthe various cells under investigation, chromosomal DNAs fromseveral of them were cleaved with this isoschizomer restrictionenzyme pair. Figure 4 shows the theoretical MspVHpaU mapof both the TPl and TP2 promoter regions: use of TPl and TP2promoter specific probes, as indicated in the figure, shouldidentify fragments of 818 and r64bp in the TPl promoter regionand 146, 107 and 504bp in the TP2 promoter regions, respectively(23), assuming complete enzymatic digestion. Figure 5 showsthe actual results obtained after probing one digest. With a probefor the promoter region of TPl (Figure 4, panel A), both B95-8and NAD-C15 were cleaved to give an 818bp fragment,indicative of complete digestion with HpaU and no methylationat CpG sites within this region. In contrast, with C15 tumourDNA, the corresponding probe recognised only a very large(> 10kb) fragment, indicative of extensive CpG methylation. Toconfirm that incomplete digestion with HpaU could not accountfor the results obtained, the experiment was repeated, using anindependently isolated sample of C15 DNA and a different batchof HpaU enzyme. Identical results were obtained (not given).


B95 8 NAD CIS

H M H M H M

B95-B

H M

NAD

H M

CIS

H M

Nucleic Acids Research, Vol. 19, No. 9 2439

CIS RAMOS7 C17

|-«513bp

Figure 6. PCR amplification of LMP transcripts. As in Figure 3; primers werechosen to amplify mRNAs corresponding to 513bp of LMP transcripts. (PolyA+) RNA was derived from NPC patient biopsy materials (tracks 1—7), fromtwo NPC biopsies passaged in nude mice, C17 and C15 (13) and from the EBVnegative B-cell line, Ramos.

818504

TP1PROMOTER

TP2PROMOTER

positive for LMP transcripts, further confirming thetranscriptional activity of this region of the viral genome in NPCs.Interestingly, LMP expression has not yet been identified inBurkitt's lymphomas (1), therefore the reservation aboutinterpretation of data on primary NPC biopsies, expressed abovefor TP gene expression, is of less relevance where LMPexpression is concerned.

Figure 5. Methylation of TP and LMP promoters. Cellular DNA from B-lymphoblastoid lines B95-8 and NAD-C15, and from the C15 tumour, as indicated,was cleaved with HpaU (H) or Mspl (M) and the resulting fragments separatedby electrophoresis on a 1 % agarose gel. Bands were identified on Southern blotsusing, in A, a probe shown in Figure 4A (for TP1) and B, a probe (for TP2,LMP) shown in Figure 4B. A. In B95-8, a single band is observed at the expectedposition, 818bp, after digestion with both enzymes. (The 164bp fragment, notseen on this blot, was observed when fragments were separated on a higherconcentration agarose gel.) A similarly-sized band is observed when both enzymesare used to cleave NAD-C15 (tracks M and H) and in the Atspl-cleavage track(track M) in C15. The latter, cleaved with HpaU (track H), gave only one verylarge band (> 10kb, indicated by arrow) indicative of extensive CpG methylationover this region in the tumour DNA. The higher molecular weight (~ lkb) minorbands in NAD-C15 (tracks H and M) and C15 (track M) may be indicative ofpartial methylation of one or more of the sites either side of those producing the818bp fragments. B. In all tracks, the major band observed is that correspondingapproximately to 504 and (146+107) bp fragments, obtained by complete digestionwith either HpaU or Mspl. (Slower migrating bands in two of the HpaU tracksmay be due to incomplete digestion or some upstream CpG methylation in thesesamples).

When the probe used corresponded to the TP2 promoter region(panel B) more or less complete digestion of the DNA with bothenzymes was observed in all three cell lines investigated, implyinglittle or no CpG methylation in this region of the genome. Theresults are consistent with the findings regarding transcriptionin the tumours, as compared with the B cell lines.

Transcriptional Expression of LMPA second viral transcript, that for a membrane-associated latentprotein (LMP; 23), has been identified in the TP1/TP2 promoterregion of the EBV genome (see Figure 1). But, whereas TP1and TP2 are transcribed in a 'rightward' direction, LMP istranscribed 'leftward' from the opposite strand. To confirm thatthe latter region was transcriptionally active in the NPCs, theability of the LMP gene to be transcribed was investigated. Thisseemed especially relevant since Western blot analyses using serafrom NPC patients have suggested that the LMP protein isexpressed in only about 30-65% of NPC biopsies examined (14,27). We have used PCR to assess the presence of the LMPtranscripts in the C15 tumour, as well as in a variety of samplesfrom NPC patients. The findings are shown in Figure 6. Althoughours is a limited study—due to the paucity of available biopsymaterials—without exception all the isolates examined were

DISCUSSION

The terminal protein (TP) genes—TP1 and TP2—encoded by thefused terminal repeats of the viral genome prove to be aninstructive model for studying control of viral gene expressionin EBV-carrying cells. Although biological functions for thesetwo related genes have not yet been identified, they use differentpromoters and specify differently-sized (40 and 53kD) proteins(16, 17). Presumably these proteins have different activities. InEBV positive B-lymphoblast cell lines, several studies have shownboth viral genes to be transcriptionally expressed (12, 15). Ourhybridization studies on a cDNA library derived from the C15NPC tumour further suggested that TP was transcriptionallyactive in carcinoma cells as well as in B-cells (8). Subsequentsequence analysis of a number of clones in the library confirmedthis, but failed to reveal any that would correspond to the largerof the two TP transcripts, that for TP1; all clones analysed werederived from TP2 transcripts. A variety of different experimentswere therefore carried out to investigate whether TP1 was indeedtranscriptionally silent in the tumour or whether there might bea trivial technical explanation for this observation. All methodsused confirmed, however, that the TP1 gene was silent in theNPC C15 tumour investigated, although it was transcriptionallyexpressed in a complimentary B-cell line (NAD-C15) derivedfrom virus 'induced' from this same tumour. The TP1 genetranscript was also found to be absent (by PCR) in the C17 NPCtumour (13), propagated in nude mice. As a corollary to theseexperiments, the transcriptional expression of a membraneprotein, designated latent membrane protein, (or LMP), encodedwithin the same area of the EBV genome but from the strandcomplementary to that encoding TP genes (Figure 1), has alsobeen investigated in the C15 tumour, as well as in C17 and anumber of biopsies from NPC patients. Transcriptionalexpression of LMP was identified by PCR in 100% of the NPCtumours studied. The TP2 gene was also transcriptionallyexpressed in all cells examined.

It is outside the focus of this paper to dwell in depth on theputative role(s) of EBV genes in tumour formation. However,it is noteworthy that LMP has been identified as a transformingoncogene in in vitro experiments and that the transformed cellsproduce tumours in immunodeficient (nude) mice (28). Thus,whereas there may be no a priori reason why TP genes need



be expressed in tumour cells, because of the juxtaposition of thepromoters for LMP and TP2, if LMP expression is essential forthe initiation or maintenance of a carcinoma (ie, NPQ, this regionof EBV—which also includes the promoter for the TP2 gene-must be transcriptionally active. The fact that the other TP gene,TP1, is not always expressed suggests that this gene may notbe involved in the tumorigenesis process. If this hypothesis iscorrect, and if LMP expression is not required in the case ofBurkitt's lymphomas (1), one would predict that TP geneexpression might also be silent in these tumours. To ourknowledge, this has not yet been explored. Again, as we (10)and others (11) have shown earlier, CpG methylation appearsto play a key role in producing, or maintaining, transcriptionallysilent regions of the viral genome in EBV-associated tumours:A probe containing the TP1 promoter sequence (Figure 4)identifies no HpaD. sensitive sites over a region in excess of lOkbpin the C15 tumour, in a region of the viral genome where thereare numerous restriction sites, a number of which are shown.Indeed, these data would suggest that virtually no promoter inthe region of the EBV genome immediately 5' to the TP1genome, contained within the BamHl A (12kbp) restrictionfragment of the genome, would be transcriptionally active.Notably, this region contains a number of genes, including theviral DNA polymerase, associated with either replication orencapsidation (29) and our earlier data showed them to betranscriptionally silent in the tumour (8). The most compellingconclusion allowed by data obtained in this study is that relatingto promoter usage and possible cellular control overtranscriptional expression of viral genes. Thus, differentialtranscriptional expression within a host cell may be sufficient toaccount for the varying pathologies associated with viral infectionfor a multigene virus like EBV.

13. Busson.P., Braham,K., Ganem.G., Thomas.F., Grausz.D., Lipinski.M.,Wakasugi.H. and Tursz,T. (1987) Proc. Natl. Acad Sd, USA, 84,6262-6266.

14. Fahraeus.R., Fu.H.L., Emberg.I., FinkeJ., Rowe.M., Klein.G., Falk,K.,Nilsson.E., Yadav.M., Busson.P., Tursz.T. and Kallin.B. (1988) Int. J.Cancer 42, 329-338.

15. Sample,.!., Lkbowicz.D. and Kieff.E. (1989) /. Virol., 63, 993-997.16. Frech.B., Zimber-Strobl.U., Suentzenich,K-O., Pavlish.O., Lenoir.G.,

Botnkamm.G.W. and Mudler-Lantzsch.N. (\99O)J.ViroL 64, 2759-2767.17. Rowe,D., HaU.L., Joab.I. and Laux.G. (1990)/.Virol., 64, 2866-2875.18. Busson.P., Ganem.G., Flores.P., Mugneret.F., Clausse.B., Caillou.B.,

Braham.K., Wagnasugi.H., Lipinski.M. and Tursz.T. (1988) Int. J. Cancer42, 599-606.

19. Emberg.I., Falk.K., MinarovitsJ., Busson.P., Tursz.T., Masucci.M.G. andKlein.G. (1989) J.gen.Virol. 70, 2989-3002.

20. ArrandJ.R., Rymo.L., WalshJ.E., Bjdrck.E., Lindahl.T. and Griffin.B.E.(1981) Nucleic Adas Res., 9, 2999-3014.

21. Maniatis.T., Fritsch.E.F. and SambrookJ. (1982) Molecular Cloning: ALaboratory Manual. Cold Spring Harbor Laboratory.

22. Baer.R., Bankier.A.T., Biggin,M.D., Deininger.P.L., Farrell, P.J.,Gibson.T.J., Hatfull.G., Hudson,G.S., Satchwell.S.C, Seguin.C,TufrheU.P.S. and Barrell.B.G. (1984) Nature 310, 207-211.

23. Liebowitz,D., Wang.D. and Kieff.E. (1986) J.Virol., 58, 233-237.24. Sanger.F., Nicklen.S. and Coulson,A.R. (1977) Proc. NatL Acad. Sd, USA.,

74, 5463-5468.25. Doerfler.W. (1983) Ann. Rev. Biochem. 52, 93-124.26. BirdAP. (1986) Nature 321,209-213.27. Young.L.S., Dawson.C.W., Clark.D., Rupani.H., Busson.P., Tursz.T.

Johnson,A. and Rickinson,A.B. (1988)7. gen. Virol., 69, 1051-1065.28. Wang.D., Liebowitz.D. and Kieff.E. (1985) Cell 43, 831-840.29. FarreU, P J . (1987) In Genetic maps 1987, (O'Brien, S.J. (ed.), Cold Spring

Harbor Laboratory, pp. 99-107.

ACKNOWLEDGEMENTS

We wish to thank The Cancer Research Campaign for generoussupport of this work. We also diank Dr. P.Busson for originalnude mouse NPC isolates, Dr. I.Ernberg for the NAD-C15 cellline, Dr. M.D.Jones for helpful discussions, and Ms. D.Liyanagefor preparation of oligonucleotides.

REFERENCES1. Klein.G. (1989) Cell, 58, 5 - 8 .2. Speck,S.H. and StromingerJ.L. (1989) In, Advances in Viral Oncology

(klein, G. ed.) Raven Press Ltd., New York, pp. 133-150.3. Rymo.L., Lindahl.T. and Adams, A. (1979) Proc. Natl. Acad. Sd. USA,

76, 2794-2798.4. Bornkamm.G.W., v Knebel-Doeberitz.M. and Lenoir.G.M. (1984) Proc.

Natl. Acad. Sd. USA, 81, 4930-4934.5. Zimber.U., Adldinger.H.K., Lenoir.G.M., Vuillaume.M., v Knebel-

Doeberitz.M., Laux.G., Desgranges.C, Wittmann.P., Freese,U-K.,Schrcider.U. and Bomkamm.G. (1986) Virology, 154, 56-66.

6. Griffin.B.E. and Karran.L. (1984) Nature 309, 78-82.7. Karran.L., Teo.C.G., King,D., Hitt.M.M., Gao.Y., Wedderburn.N. and

Griffin.B.E. (1990) Int. J. Cancer 45, 763-772.8. Hitt,M.M., Allday.M.J., Hara.T., Karran.L., Jones.M.D., Busson.P.,

Tursz.T., Emberg.I. and Griffin.B.E. (1989) EMBOJ., 8, 2639-2651.9. Smith,P., Karran.L., Griffin,B.E., ShamJ., Choy.D., Lung.M. and

Ng.M.H. (in press) First Hong Kong Symposium on NPC.10. Allday.M.J., Kundu.D, Finerty.S. and Griffin.B.E. (1990) Int. J. Cancer

45, 1125-1130.11. Masucci.M.G., Contreras-Salazar.B., Raguar,E., Falk.K., MinarovitsJ.,

Ernberg.I. and Klein.G. (1989) / Virol., 63, 3135-3141.12. Laux.G., Perricaudet.M. and FarreU.P.J. (1988) EMBOJ., 7, 769-774.


Documents

(TP1 and LMP) in lymphocyte and epithelial cells