THE JOURNAL OF BIOLOGICAL CHEMISTRY Val. 266, No. 34, Issue of December 5, pp. 23416-23421, 1991 Printed in U.S. A.

The Hepatitis B Virus S Promoter Comprises A CCAAT Motif and Two Initiation Regions*

(Received for publication, May 31, 1991)

Dao-Xiu ZhouS and T. S. Benedict Yen6 From the DeRartment of Patholoex Veterans Affairs Medical Center and University of California School of Medicine, Sun Francislo, California 94143-0506

The hepatitis B virus S (major surface gene) pro- moter is embedded in two overlapping open reading frames and specifies several transcripts with hetero- geneous 5’ termini. We have identified the cis-elements necessary for S promoter function. A single upstream CCAAT element is essential for high level expression in both liver and non-liver cells. No TATA box is present, but two regions surrounding the initiation sites can function separately as initiating elements. These results show that the S promoter has a simple requirement for upstream activating sequences, but a complex initiation region. Therefore, it may provide a suitable model system for studying transcription initi- ation in non-TATA type promoters.

Eucaryotic gene expression is largely regulated at the level of transcription. Numerous studies have shown that tran- scriptional activity is determined by trans-acting cellular fac- tors that bind to specific cis-elements (reviewed in Ref. 1). Some of these elements specify the sites of initiation, whereas others up- or down-regulate the rate of transcriptional initi- ation. One well studied initiation element is the so-called TATA box, which binds the factor TFIID (reviewed in Ref. 2). In higher eucaryotes, this element helps in specifying initiation at a site -30 base pairs (bp)’ downstream. However, many genes, including most housekeeping genes, do not have a recognizable TATA box. These genes frequently have het- erogeneous start sites, and most have GC-rich elements (3) within a few hundred bp of the start site. The initiation elements for these genes are less well characterized but appear to center around the initiation sites (4-7). The factors that bind to these elements are as yet unknown.

The hepatitis B virus (HBV) S (major surface gene) pro- moter specifies transcripts with heterogeneous 5’ termini (8- 10; reviewed in Ref. 11), with the largest transcript giving rise to the middle surface protein and the other transcripts giving rise to the small (major) surface protein (see Fig. 1). This

*This work was supported by a Veterans Administration merit award and a grant from the Council for Tobacco Research (to T. S. B. Y . ) . The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Partial salary support from the Centre National de la Recherche Scientifique, France. Present address: Laboratoire de Biologie Mole- culaire Vegetale, CNRS UA1178, University of Grenoble, F38402 Saint Martin d’Heres Cedex, France.

T o whom correspondence should be addressed: Dept. of Pathol- ogy, Box 0506, University of California, San Francisco, CA 94143- 0506. Tel.: 415-476-5334; Fax: 415-476-9672.

’ The abbreviations used are: bp, base pair(s); HBV, hepatitis B virus.

promoter is highly unusual, in that it is entirely contained within two overlapping open reading frames (for the large surface protein and the polymerase protein; see Fig. 1) and thus may have unique features. Since previous studies have only defined the rough boundaries of the S promoter (8-lo), we have started to examine the cis-elements in more detail. Our results show that a single upstream element is needed for high level expression of this promoter. However, two elements (with no homology to the TATA box) can function separately to direct transcriptional initiation. Therefore, the S promoter may provide a novel model toward the understanding of non- TATA initiator elements.

MATERIALS AND METHODS

Plasmid Construction-The large BglII fragment of HBV genomic DNA strain adw (12) was cloned into the BamHI site of pTZ19U and named pSAg. This fragment contains the preS1, S and X promoters, the entire surface and X open reading frames, and the polyadenylation site (see Fig. 1). Upstream deletion mutants of pSAg (see Fig. 1 for maps) were made as follows. For pSAg (ABst/Bsm), pSAg was di- gested with BstEII and BsmI, treated with T4 DNA polymerase, and religated (13). For pSAg(ABsm/Bsm), pSAg was digested with BsmI and religated. For pSAg(ABgl/Eco), the EcoRI-BglII fragment of HBV DNA was inserted into the BamHI site of pTZ19U, with the addition of an EcoRI-BglII linker (13). For pSAg(ABgl/Hin), pSAg(ABgl/Eco) was digested with HindIII, treated with T4 DNA polymerase, partially digested with EcoRI, and ligated to the HincII to EcoRI fragment of HBV DNA cut out from pSAg. For pSAg(-32), a double-stranded oligonucleotide corresponding to nucleotides 3191- 3221 of the HBV DNA (see Fig. 3B) was synthesized with EcoRI- compatible ends and ligated into the HBV EcoRI site of pSAg(ABgl/ Eco).

Clustered point mutants of pSAg(ABgl/Hin) (M1 and M2, see Fig. 3B for sequences) were made by oligonucleotide-directed mutagenesis (14) with a ki t from Bio-Rad.

pSAg(ABgl/Hin)R54, pSAg(ABgl/Hin)Ml was digested with SphI Downstream mutants of pSAg were derived as follows. For

and PstI and ligated to the SphI to PstI fragment of pSAg(ABgl/ Eco). This replaces a 54-bp segment of HBV DNA (Fig. 3 B ) with a 36-bp segment of polylinker from pTZ19U. For pSAg(ABgl/Hin)D26, pSAg(ABgl/Hin) was partially digested with EcoRI and PstI, treated with T4 DNA polymerase, and religated. For pSAg(ABgl/Hin)D85, pSAg(ABgl/Hin)Rh4 was digested with PstI and religated.

Cell Transfection, Primer Extension, and Radioirnrnunoassay- HUH-7 human hepatoma cells, which are free of HBV DNA (15), and CV-1 monkey kidney cells were grown in Dulbecco’s modified Eagle’s medium-H-21 medium with 10% fetal bovine serum under 6% COY a t 37 “C. Cells were transfected overnight with 10 wg of plasmid DNA, using the calcium phosphate precipitation method (16). Forty hours after initiating the transfection, total RNA was purified by the cesium chloride/guanidinium method (17) and subjected to primer extension analysis with a surface gene primer, as described previously (18, 19). The primer corresponds to nucleotides 111-85 in the HBV genome. To localize the 3‘ ends of the extended products, DNA sequencing reactions using the chain termination method (20) were performed with the same primer on HBV DNA and electrophoresed in adjacent lanes. In selected experiments, the cell media were harvested and

23416

Heput i t i s R Virus S P r o m o t e r E l e m e n t s 23417

assayed for HHY surface protein with a radioirnmunoassay kit from Ahhott Laboratories.

l).\‘o.sc I’rofccfion a n d (;rl-.shiff Assnys-HUH-; cell extracts were ohtained with the method of Oshorn r t ol. (21). For DNase I protection 1 3 ) . the HstEII to h’coHI fragment ofpSAg,?vas laheled at the :3’ end o f the top strand (shown in Fig. :1H) with [“PJdATI’and the Klenow l’ragment of DSA polymerase I (13) and incubated with DNase I in the presence of various amounts of HUH-7 extracts, as described previously (1X. 19). The digested products were then electrophoresed on a denaturing polyacrylamide gel. together with a (; + A scission reaction of the same D S A . as marker. For gel shift analysis (22). the NincII to KcoRi fragment of pSAg was laheled with I ‘“I’IdATI’ and the Klenow fragment o f DNA polymerase I. incuhated with HUH-7 extracts. and electrophoresed on a nondenaturing polyacrylamide gel, a s descrihed previously (18. 19). In some experiments. unlaheled competitor DS.4 at %-fold molar excess was included in the incuha- tion mixture.

S q c t c ~ n c c ~ Ann/ysi.s-HRY genomic sequences were retrieved from (;enRank and analyzed with Intelligenetics software on the Hionet Computer.

RESULTS

Localization of Upstrcam Cis-elements-The large HglII fragment of HBV DNA (Fig. 1) directs the transcription of a similar amount of S mRNA as the entire viral genome when transfected into hepatoma cells (data not shown) and hence presumably contains all of the cis-elements necessary for the function of the S promoter. To map any upstream elements important for S promoter activity, plasmids with various 5’ deletions were constructed (see Fig. 1) and tested for the effect on S transcription, as assayed by primer extension analysis of transiently transfected HUH-7 cells. As shown in Fig. 2A, deletion to the HstEII site (SAg(ABgI/Bst)) had no effect, whereas deletion to the HincII site (SAg(ARgl/Hin)) in- creased S transcription by 2- to 3-fold. This suggests the presence of a negative element between the RstEII and HirzcII sites. This element was further localized by deletion of suc- cessively smaller fragments to be between the two HsmI sites

Fres: -’ x E] Ptli.

I . . . . . . . . . . . . . . . I I

AAi A * ’.’::-::. ? r e a . . . . . . . . . . . . . . . An D ? ~ A

BH H H

SA9

A SAg(ABet/Bern)

A SAg(ABsrn/Bsrn)

SAglABgl/Bst)

SAg(ABgl/Hin)

SAg( -32 )

FIG. 1. Diagrammatic representation of the large BglII fragment of H R V genome and plasmid constructions used in this paper. The centrnl rrctnnglr represents HHV DNA, with open reading frames shaded: the overlapping polymerase open reading frame, which starts upstream of the k f t HglII site and terminates within the S gene. is not shown to preserve legihility. The three forms of the surface protein are translated from the same reading l‘rame and differ in the initiating AUG codon used: thus, the small surface protein includes only the shaded region marked “Surfncc“: the middle surface protein includes in addition the “I’rrS2” region. whereas the large surface protein also includes the “PrrSl” region. I he major transcripts synthesized from the three promoters in this l’ragment are shown ahove the D S A as nrrocc’s. The chevron indicates the location of the oligonucleotide used for primer extension analysis o f S transcripts. The fragments of HHV DNA present in the deriva- tives of SAg are shown as hrncy linw helow. A,,, the HRV polyade- nylation site: H , HstEII site: M , HsrnI sites: H , HincII site. See Fig. 3 H for the detailed sequences of the S promoter.

,.

A

2 VI

cv m I v

VI 2

B cv I

m v

2 VI

9 FIG. 2. Transcription from 5’ deletional mutants of the S

promoter. A, the indicated plasmids, containing fragments 0 1 the HH\’ genome as shown in Fig. 1, were transfected into HUH-7 cells and the S transcripts analyzed hv primer extension. The live extended fragments marked with “s” represent the five major S transcripts, hased on the size estimate from comparison with molecular weight markers and on their ahsence in nontransfectetl cells (data not shown); their 5’ ends correspond to nucleotides 3197, 3217, 3220. 7 . and 10 in the HRV genome and are similar to start sites mapped previously by other investigators (X, 36). The smaller species are probahly an artifact of primer extension, since their intensity varied I’rom experiment to experiment, even on identical samples. Similar results were ohtained with three independent transfections. 13. a 20- fold longer exposure of extended products from SA:.(-:??)), revealing a low remaining level of correctly initiated transcripts.

(Fig. 2 A ) . Further deletion past the HincII site to within 10 bp upst,ream of the most upstream start site resulted in a dramatic decrease in S transcription (SAg(-32), Fig. ZA) , implying the presence of an important up-regulatory element downstream of the HincII site.

T o localize DNA elements that may bind cellular trans- acting factors, DNase I protection assay was performed on the S promoter fragment from the RstEII site to the EcoRI site, which is in the center of the transcription initiation sites. As shown in Fig. 3A, five protected regions were present. Two (footprints IV and V) are within the region between the two RsmI sites that mediate the negative regulation of the S promoter. Therefore, one or both could bind a negative-act,ing cellular factor. NF-I has been shown previously to bind to the footprint V region (9) and hence may mediate t.his negative regulation.

The third protected region (footprint 111) is transected hy the HincII site (Fig. 3 R ) and thus presumably does not hind a cellular factor with a significant effect on S promoter strength. The two most downstream protected regions (foot- prints I and 11) lie in t.he region that is important for S promoter activity by our deletional analysis, and thus one or b0t.h probably bind to positive-acting cellular factor(s1. To confirm the binding of cellular factor(s) to these regions, we performed gel-shift analysis of the DNA fragment from the HincII to EcoRI sites. A single major shifted band was seen with HUH-7 hepatoma cell extracts (Fig. 41, as well as with extracts from several other epithelial and lymphoid cell lines (data not shown). This represented a sequence-specific inter- action, since the band was competed hy excess unlaheled

23418 Hepatitis I3 Virus S Promoter Elements

FIG. 3 . Footprint analysis of the S promoter. A . a labeled fragment of S promoter extending from the HstEII to t h e I;coliI sites was partially digested with DSase I in the presence of RSA t lanc “ 0 ” ) . or 5 and 10 pg of HUH-7 cell nuclear extract (lanes “5” and “ I O ” , re- spectively) and electrophoresed on a de- naturing gel. Clearly footprinted regions are highlighted hy bars marked with Ro- man numemls 1-1’. The lane marked “G +A“ represents chemical cleavage of the same fragment at the G and A residues, as size markers. H . the sequence of the S promoter from the HstEII to the EcoRI sites is presented. as well as the five footprinted regions (undrr l ines, with numbering corresponding to part A ) . ‘The point mutations introduced into re- gions I and 11, in the mutant plasmids 511 and “2, respectively, are shown in italics above the wild-tJpe sequence. The putative variant TATA box of deMedina (’t a(. (9) is shown in outline type, whereas the most upstream nucleotide of the S promoter present in SAg(-GZ) is ouer- lined with an nrrou‘. The three start si tes present in the sequence shown are indi- cated with fil led circlrs over the initiating nucleotides.

I x “=- *

B 2825 BstEII GTCACCATATTCTTGGGAACAAGAGCTACAGCATGGGAGGTTGGTCATCAAAACCTCGCAAAGGCATGGGGACGAA

TCTTTCTGTTCCCAACCCTCTGGGATTCTTTCCCGATCATCAGTTGGACCCTGCATTCGGA~AACT~AAACAAT BsmI

V

CCAGATTGGGACTTCAACCCCATCAAGGACCACTGGCCAGC AGCCAACCA3.GTAGGAGTGGGAGCATTCGGGCCAG IV

B8mI

GGCTCACCCCTCCACACGGCGGTATTTTGGGGTGGAGCCCTCAGGCTCAGGGCATATTGACCACAGTGTCAACAAT HincII

111

TGG C T mTCCTCCTG-CGGCAGT CAFGAAGG€AGCCTACTCCCATCTCTCCACCTC%&&@&GACAGTCA

“f

I1 I

TCCTCAGGCCATGCAGTGGAA EcoRI

3221/0

homologous DNA but not by an irrelevant fragment of DNA (Fig. 4).

To determine if either footprinted regions I or 11, or both, were being bound in this gel-shift assay, we introduced clus- tered point mutations into these sites (see Fig. 3R for se- quences) of pSAg(ABgl/Hin). These mutants (M1 and M2, respectively) were then used as unlabeled competitors in the gel-shift experiment. As seen in Fig. 4, mutant M1 still competed effectively in this assay, but mutant M2 did not. This result implied that footprinted region I1 was responsible for the strong sequence-specific binding of a cellular factor detected with the gel-shift assay. It is not clear why no shifted band was seen for region I, although i t may be due to a weaker affinity for cellular factors.

Footprinted Region I I as Up-regulatory Element-To deter- mine if either region I or I1 is functionally important for the S promoter, we determined the effect of clustered mutations M1 or M2 on this promoter’s activity. As shown in Fig. 5, the M1 mutations had no significant effect on the amount of S transcripts synthesized in HUH-7 cells, whereas the M2 mu- tations produced a >20-fold decrease. Interestingly, the M2

mutations had a similar effect in CV-1 kidney cells (Fig. 5 ) . Therefore, it appears that footprinted region I1 binds one or more factor(s) present in both liver and non-liver cells that is critical for its activity. Inspection of this region reveals the sequence CCAAT, a motif known to bind many transcription factors (23-31).

To confirm the results from the primer extension assays, we also performed radioimmunoassays for surface protein secretion from cells transfected with some of the plasmids described above. As shown in Fig. 6, SAg(ABgl/Hin) directed the synthesis and secretion of a large amount of surface protein, whereas the amount of surface protein secretion directed by SAg(-32) was below the sensitivity of the assay. Furthermore, the M1 mutations had only a small effect on the amount of surface protein secreted, but the M2 mutations reduced i t to undetectable levels. Therefore, the amount of surface protein translated from the deletion and clustered mutant constructs correlated well with the amount of S tran- scripts detected with primer extension assays.

Mapping of Initiation Elements-Even when the region encompassing the CCAAT sequence is deleted, as in the

Extract

Competitor

T- -

SB-

FP -

FIG. 4. Gel-shift analysis of the S promoter. T h e S promoter from the HincII to the I.:coHI sites was end-laheled. incubated with o r without HUH-7 nuclear extracts, and electrophoresed through a native gel. T indicates the top of the gel: FP. the free probe; and SB, the shifted hand. In some of' the lanes, there was a 50-fold molar excess of unlabeled competitor DNA fragment. WT, the wild-type HincII to EcoRI fragment; M I , HincII to GcoRI fragment with the M1 clustered mutations shown in Fig. 3H: M2, the same fragment with the M:! clustered mutations shown in Fig. 3H; N S , a nonspecific competitor, derived from the NcoI-StuI fragment of the HRV S gene.

plasmid SAg(-32), there remains a very low level of correctly initiated S transcripts (Fig. 2B), presumably specified by initiation elements downstream of the CCAAT sequence. T o confirm this inferrence, we deleted an 85-bp segment down- stream of the CCAAT sequence (Fig. 7 A ) that spanned all of the initiating sites. As seen in Fig. 7I3, this deletion (SAg(ABgl/Hin)D85) resulted in the loss of all correctly ini- tiated transcripts. Therefore, the CCAAT sequence cannot act alone, but depends on downstream initiation elements to activate transcription. It should be noted that these elements cannot include a putative variant TATA identified previously by deMedina et al. (9), since this sequence motif is partially deleted in SAg(-32) (Fig. 3B).

To map these putative initiation elements in more detail, we constructed plasmids with smaller changes in the initiating region. When only the 26 bp downstream of the EcoRI site are deleted (SAg(ABgl/Hin)D26), there remains correct ini- tiation from the two most upstream start sites (Fig. 7), both of which are upstream of the EcoRI cleavage site, although the relative efficiency of these two sites are changed. However, the three downstream initiation sites are no longer used. Conversely, when the 54 bp upstream of the EcoRI site are replaced with an irrelevant sequence (SAg(ABgl/Hin)R54), transcripts from the two downstream initiation sites remain, whereas there is no transcription from the most upstream site (Fig. 7). The remaining two transcripts are also present but appear to be initiated 1 bp downstream from the wild-type sites. This shift may have resulted from the shorter-than- wild-type length of the stuffer fragment. In toto, these data suggest that the region upstream of the EcoRI site around the most upstream initiation site acts as one initiation region,

HUH-7 cv- 1 FIG. 5 . Transcription from clustered point mutants of the S

promoter. The indicated HHV constructs with deletions (see Fig. 1) or clustered point mutations (see Fig. 3 H ) in the S promoter were transfected in to HUH-7 hepatoma cells (kit p o n d ) or CV-1 kidney cells (right p o n d ) and the S transcripts analyzed with primer exten- sion. Similar results were obtained with three independent transfec- tions.

whereas the sequences surrounding the EcoRI site c0nstitut.e a second initiation region.

DISCUSSION

Since the HBV S promoter has the unusual feature of being embedded in two overlapping open reading frames, it seemed likely that it would have a simplified structure. Our results presented here pinpoint a single upstream promoter element with a CCAAT motif that is necessary to up-regulate this promoter; as may be expected, this motif is conserved in all HBV strains (Fig. 8). No other upstream promoter elements seem to be necessary for S transcription, since sequences either upstream or downstream of the CCAAT sequence can be deleted with no deleterious effect on the efficiency of transcription (Figs. 2 and 7). Specifically, in contrast to most TATA-less promoters (I), no GC-rich element is present. Also unlike some TATA-less promoters (4, 32), there is no requirement for elements downstream of the initiation sites, since we have shown previously that the HRV sequences downstream of the EcoRI site can be deleted and replaced with the chloramphenicol acetyltransferase gene (18).

In addition, there is a negative-acting element further up- stream, in the region of an NF-I binding site. The physiolog- ical significance of this negative element is unknown, al- though Paonessa et al. (28) also found negative regulation of transcription in liver cells by an NF-I site. Also unclear is the role of cellular factors that bind to footprinted regions I and 111, which appear to have no function in the S promoter. We believe it likely that these factors may regulate one or more of the other HBV promoters. Specifically, Bulla and Siddiqui (33) have postulated the existence of an element in this region that down-regulates the upstream pres1 promoter. We are performing experiments to test this h-ypothesis.

2" Hepatitis R Virus S Promoter Elements

"* 1

N z c .- c

FIG. 6. T r a n s l a t i o n f r o m c l u s t e r e d p o i n t m u t a n t s o f t h e S p r o m o t e r . ' I ' h e indicated HH\' constructs. with or without clustered point mutations in the S promoter (see Fig. 314 for sequence), were transfected into HUH-7 cells. and the media harvested after 2 days I'or radioimmunoassay of secreted surface proteins. The results rep- resent mean ? S.D. of three independent experiments.

Our results are in agreement with the deletional analyses of Cattaneo et ai. (8) and Raney et al. (IO), but diverge from that of deMedina et al. (9), who were not able to detect any up-regulation by the CCAAT sequence. This difference can- not be due to strain differences, since the first two groups used strain ayw, whereas we and the last group both used strain adw. Furthermore, the CCAAT motif is conserved among all HRV strains examined (Fig. 8). More likely, the difference is due to the viral enhancer 11, which overlaps the 3' end of the X gene (18) and is absent from the constructs used by deMedina et al. (9). We have shown previously that enhancer I1 up-regulates the S promoter by -10-fold (IS), and our preliminary data indicate that this effect is dependent on the CCAAT sequence.'

At present it is unknown which member, if any, of the five known families of CCAAT-binding factors is responsible for up-regulating the S promoter. It is possible that many differ- ent factors can bind this sequence and regulate the S promoter in different cell types. With the availability of cloned CCAAT- binding factors (23-31), we may be able to resolve this ques- tion in the future with co-transfection and/or in vitro t ran- scription studies.

In contrast to the upstream sequence requirements, the cis- elements necessary for correct initiation of the S transcripts are complex. There is no evidence for a TATA-like element, bu t a t least two initiation elements are present, each of which can act on its own. We have noted that one region (well conserved among all HBV strains) contains a sequence with a 6/8 nucleotide match to the concensus initiation sequence (asterisk, Fig. 8) , recently identified by Bucher (34) in a computer-assisted analysis of 303 eucaryotic promoters. The other region contains two motifs in opposite orientations (", Fig. 8) similar to a concensus sequence (HIP1 site) for several transcriptional initiating sites (ii), but these motifs are less well conserved among HBV strains (Fig. 8). This latter region

D. X. Zhou and T. S. H. Yen. unpuhlished ohservations.

A 3 1 2 3 MC.L&TTCCTCCTCCTGCCTCCACC2bY?JCGGCAGTCAGGAAGGCA

- I

GCCTACTCCCATCTCTCCACCTCTMGAGAGAC4GTCATCCTCAGGC or

""""""""""I It. 0

CATGCAG?GEAATTCCACTGGCTTGCACCAAGCTCTGCAGGATC 3 5

0 .

"11

B

.L

c

z m a

- m

v 0,

m Q

*

* *

* *

*

* D *

* *

E

FIG. 7. T r a n s c r i p t i o n f r o m 3' d e l e t i o n a n d r e p l a c e m e n t m u - tants o f t h e S promoter.A, sequence o fS promoter Irom the Hincll to NarnHI sites. The deleted region in SAg(lHgI/Hin)D85 is ocer- lined; the deleted region in SAg(SRgl/Hin)D26 is underlined with a solid line; the replaced region in SAg(lRgI/Hin)R54 is undcrlinod with a d d ~ d [in(,. The CCAAT motif is shown in outline Iype, whereas the I:'coRI site is in hold type. *, normal start sites; #, start sites in SAg(lRgl/Hin)DZB; 0, start sites in SAg(lHgl/Hin)R54 (see data in part 13). R, the indicated HRV constructs were transfected into HUH-7 cells and S transcripts analyzed with primer extension. The symhols for the start sites correspond to those in part A. Similar results were ohtained with three independent transfections. Recause of the deletion in SAg(lRgl/Hin)I126, the transcripts in this mutant are 26 h s e s smaller than the corresponding wild-type transcripts.

is also contained within a long palindrome, the central core of which is conserved among HBV strains (Fig. 8). Further experiments are needed to determine the functional impor- tance of these sequence motifs and to identify the cellular factor(s) that bind to them.

For most cellular TATA-less promoters, the significance of multiple initiation sites is unclear. However, in the case of the S promoter, it is known that the different transcripts code for two forms of the surface protein, since the various start sites span the initiating AUG codon for the middle surface protein (see Fig. 8; reviewed in Ref. 11). Interestingly, the single (largest) transcript that can code for the middle surface protein is initiated from the upstream initiator region, whereas the other transcripts that code only for the small surface protein are initiated from the downstream region with

Hepatitis B Virus S 3 1 2 3 ~ ~ ~ c n a t t c c t [ : C T C C T G C C T C C a C C M T C ( ; ~ C A G T ~ A G G ~ t i q C A ~ i ~ ~ ' ~ A C t ~ C C a T C T C l C C A ~ ~ C T C

TCAG7:'TT' I . ~ . ~ ~ R ~ A C A ~ ~ ~ . C R T C C T C A G G ~ ~ ~ ~ A G T ~ G A ~ ~ ~ - T C C A C ~ ~ C ~ ~ . T ~ ~ A C C A A ~ C ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

ATTTC. . . .GCCrl'

TGGC. . . . . . . . . GAAATO 35

FIG. 8. Summary of cis-elements of the S promoter con- served in different HBV strains. The entire S promoter, from the HincII to the BamHI sites, is shown, with the normal start sites indicated by bold type of the initiating nucleotides. Nucleotides con- served among six different strains of HBV (adw, adw2, adr, adrcg, ayw, X51970), representing all three groups of human-derived HBV defined by Orito et ai. (35), are shown in capital letters. The CCAAT sequence is shown in outline type; the initiating codon (shown as the DNA equivalent, ATG) for the middle surface protein is underlined; the palindrome around the four downstream start sites is highlighted with head-to-head arrows. *, the concensus initiation sequence of Rucher (34) juxtaposed to a similar sequence in the S promoter; O ,

the concensus HIP l site (5) (ATTTCN,..,,,GCCA; shown in both orientations), juxtaposed to similar sequences in the S promoter.

HIP l homologies (Fig. 8). Therefore, if different cellular factors are needed to initiate transcription at these two sites, there may be differential regulation of the relative amounts of these two protein products during the viral life cycle. In any case, in view of the clear physiological role for having multiple start sites, and in view of the simple structure of the upstream element, the S promoter may provide a unique model for studying non-TATA promoters with multiple ini- tiation elements.

Acknowledgments-We thank K. Leetung and A. Taraboulos for technical assistance and E. Fodor and J. Ou for critical reading of the manuscript.

REFERENCES

1. Mitchell, P. J., and Tjian, R. (1989) Science 245, 371-378 2. Lewin, B. (1990) Cell 61, 1161-1164 3. Kadonaga, J. T., and Tjian, R. (1986) Proc. Natl. Acad. Sci.

4. Ayer, D. E., and Dynan, W. S. (1988) Mol. Cell. Biol. 8, 2021-

5. Means, A. L., and Farnham, P. J. (1990) Mol. Cell. Biol. 10,653-

6. Smale, S. T., and Baltimore, D. (1989) Cell 57, 103-113 7. Smale, S. T., Schmidt, M. C., Berk, A. J., and Baltimore, D.

8. Cattaneo, R., Will, H., Hernandez, N., and Schaller, H. (1983)

9. De-Medina, T., Faktor, O., and Shaul, Y. (1988) Mol. Cell. Biol.

U. S. A. 83, 5889-5893

2033

661

(1990) Proc. Natl. Acad. Sci. U. S. A. 87, 4509-4513

Nature 305, 336-338

8,2449-2455

Promoter Elements 23421

10. Raney, A. K., Milich, D. R., and McLachlan, A. (1989) J . Virol.

11. Ganem, D., and Varmus, H. E. (1987) Annu. Reu. Biochem. 56, 651-693

12. Valenzuela, P., Quiroga, M., Zaldivar, J., Gray, P., and Rutter, W. (1980) in Animal Virus Genetics (Fields, B., Jaenisch, R., and Fox, C., eds) pp. 57-70, Academic Press, NY

13. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

14. Kunkel, T. K. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 488-491 15. Nakabayashi, H., Taketa, K., Miyano, K., Yamane, T., and Sato,

J . (1982) Cancer Res. 42, 3858-3863 16. Gorman, C. (1985) in D N A Cloning (Glover, D. M., ed) Vol. 11,

pp. 143-190, IRL Press, Oxford 17. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter,

W. J . (1979) Biochemistry 18, 5294-5299 18. Zhou, D.-X., and Yen, T. S. B. (1990) J. Biol. Chem. 265,20731-

20734 19. Zhou, D.-X., and Yen, T. S. B. (1991) Mol. Cell. Bid. 11, 1353-

1359 20. Maxam, A. M., and Gilbert, W . (1977) Proc. Natl. Acud. Sci.

U. S. A. 74, 560-564 21. Osborn, L., Kunkel, S., and Nabel, G. (1989) Proc. Natl. Acad.

Sci. U. S. A . 86, 2336-2340 22. Fried, M., and Crothers, D. (1981) Nucleic Acids Res. 9, 6505-

6525 23. Chodosh, L. A., Baldwin, A. S., Carthew, R. W. , and Sharp, P. A.

(1988) Cell 53, 11-24 24. Didier, D. K., Schiffenbauer, J., Woulfe, S. L., Zacheis, M., and

Schwartz, B. D. (1988) Proc. Nutl. Acad. Sci. U. S. A. 85, 7322- 7326

25. Landschulz, W. H., Johnson, P. F., Adashi, E. Y., Graves, B. J., and McKnight, S. L. (1988) Genes & Deu. 2, 786-800

26. Lum, L. S., Sultzman, L. A., Kaufman, R. J., Linzer, D. I., and Wu, B. J. (1990) Mol. Cell. Biol. 10, 6709-6717

27. Ozer, J., Faber, M., Chalkley, R., and Sealy, L. (1990) J. Biol. Chem. 265, 22143-22152

28. Paonessa, G., Gounari, F., Frank, R., and Cortese, R. (1988) E M B O J. 7, 3115-3123

29. Santoro, C., Mermod, N., Andrews, P. C., and Tjian, R. (1988) Nature 334, 218-224

30. van Huijsduijnen, R., Li, X. Y., Black, D., Matthes, H., Benoist, C., and Mathis, D. (1990) E M B O J . 9 , 3119-3127

31. Vuorio, T., Maity, S. N., and de Crombrugghe, B. (1990) J . Bioi. Chem. 265, 22480-22486

32. Farnham, P. J., and Means, A. L. (1990) Mol. Cell. Biol. 10,

33. Bulla, G . A., and Siddiqui, A. (1989) Virology 170, 251-260 34. Bucher, P. (1990) J. Mol. Biol. 212, 563-578 35. Orito, E., Mizokami, M., h a , Y., Moriyama, E. N., Kameshima,

N., Yamamoto, M., and Gojobori, T. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 7059-7062

36. Ou, J. H., and Rutter, W. J. (1985) Proc. Natl. Acad. Sci. U. S. A . 82,83-87

63,3919-3925

1390-1398