Two strains of SIVmac show differential transactivation mediated by sequences in the promoter

VIROLOGY 191, 559-568 (1992)

Two Strains of SIV,,, Show Differential Transactivation Mediated by Sequences in the Promoter

MARK G. ANDERSON AND JANICE E. CLEMENTS’

Department of Molecular Biology and Genetics, Johns Hopkins University School of Medicine, Baltimore, Maryland 2 12 18; and The Retrovirus Biology Laboratory, Division of Comparative Medicine, 720 Rutland Avenue, Traylor G-60, Baltimore, Maryland 2 1205

Received June 8, 7992; accepted August 10, 1992

Two infectious molecular clones of simian immunodeficiency virus, SIV,,,251 and SIV,,,239, have very different in vivo properties, SIV,,,239 being much more pathogenic than SIV,,,251. To assess whether the in viva differences between the two viruses would be reflected in transcriptional rates in vitro, transcriptional activity in the presence of the transactivation protein tat was analyzed by transient transfection assays in HUT-78 and U937 cells. Whereas the two promoters had similar basal activities (Anderson and Clements, 1991, J. Viral. 65,51-60) the promoter of SIV,,,239 was transactivated to a greater extent. Removal of sequences 5’to -225 and 3’to +18 maintained the basal activity, yet made the promoter unresponsive to tat. Addition of bases +19 to +149 reconstituted transactivation and decreased basal activity. Analysis of deletion mutants with reconstituted transactivation response region determined that differences between the two strains were maintained even when only the proximal sequences, -225 to +18 of the U3 and R region were placed upstream of the TAR sequences. This region contains four nucleotide differences and the potential Sp-1 -binding sites, where there are an additional 11 bases in SIV,,,239 that create a third potential Sp-1 site, compared to only 2 in SIV,,,251. Transactivation in this assay system was found to correlate better to RNA differences shortly after transfection (12 hr) than later (46 hr). o 1992 Academic press. IIIC.

INTRODUCTION

Simian immunodeficiency viruses are a family of pri- mate lentiviruses closely related to the human immunodeficiency viruses, HIV-1 and HIV-2 (Chakrabarti et al., 1987; Daniel et a/., 1988; Fukasawa et al., 1988; Hahn et al., 1987; Kanki et al., 1985). SIV,,,, isolated from captive rhesus macaques causes a disease in macaques similar to AIDS in humans (Daniel et a/., 1985, 1987; Letvin et al., 1983, 1985). SIV,,, infects T4 lym- phocytes and monocytes/macrophages in viva and causes cytopathic effects in infected T-cell lines in vitro (Daniel et a/., 1985; Kannagi et al., 1985; Ringler et al., 1989). These similarities in disease and cell tropism make SW,,, one of the best animal models with which to study the pathogenesis of AIDS. Closely related strains of SIV,,, have been isolated, and infectious molecular clones have been obtained (Kestler et al., 1988). One clone, SW,,,251 was isolated from a rhesus monkey with a malignant lymphoma (Daniel et a/., 1985). Virus passaged from this monkey led to isolation and subsequent cloning of a second strain, SIV,,,239 (Daniel et a/., 1985; Kestler et a/., 1988). ln vitro analysis of these SIV isolates, as well as that of the infectious molecular clones, showed ability to repli- cate to high levels in macaque T-cells and human T- cell lines (Naidu et a/., 1988). In viva, however, the vi-

’ To whom reprint requests should be addressed.

ruses exhibit dramatically different pathogenic properties. SIV,,,239-infected monkeys have persistent viremia accompanied by dissemination of virus into multiple tissues including spleen, lymph node, and bone marrow and usually have an acute disease course (Kestler et a/., 1988; Naidu et al., 1988; Sharma et a/., 1992). SIV,,,251-infected animals, conversely, have transient viremia, little virus in tissue, and a more chronic disease progression (Naidu et al., 1988; Sharma et a/., 1992). ln situ hybridization of bone marrow cells from SIV,,,239- and SIV,,,251 -infected animals demonstrated that far more viral RNA was present in those cells in the SIV,,,239-infected monkeys than those infected with SIV,,,25 1 (Sharma et a/., 1992; M. C. Zink, personal communication). The higher level of gene expression of SW,,,239 in bone marrow and tissues as compared to SIV,,,251 may be an important element in the initial infection and may lead to the more rapid disease progression caused by this virus.

The long terminal repeats (LTRs) of retroviruses contain the promoter elements required to initiate transcription and thus play a role in the level of viral gene expression and in the pathogenic potential of the virus (Chatis et al., 1984; Lenz et al., 1984; Tsichlis and Cof- fin 1980). Immediately upstream of the start site of transcription (+l) of SIV,,, are several protein binding regions. These include the TATA box, two or three Sp- 1 sites, one NF-kB site, and an additional upstream

559 0042.6822/92 $5.00 Copyright 0 1992 by Academic Press. Inc. All rights of reproductton in any form reserved.

560 ANDERSON AND CLEMENTS

protein-binding site at -155 to -137 (see Fig. 2). Ear- lier studies using deletion analysis of U3 sequences from both viral LTRs have shown that an additional upstream binding region is required in the SIV,,,251 promoter to maintain similar levels of expression as the SIV,,,239 promoter. However, the promoters have comparable levels of basal expression in the context of the complete intact viral LTR (Anderson and Clements, 1991). Therefore, basal transcription levels do not appear to contribute to the differences in replication between the two strains in vivo.

Regulation of viral gene expression occurs via the tat protein, the trans-activation gene product found in HIV-l, HIV-2, and SIV. ln vitro, lentiviral tat proteins have been shown to greatly increase gene expression from their respective LTRs (Arya et a/., 1987; Sodroski et al., 1985) and to be essential for replication in the HIV-1 virus (Dayton et al., 1986; Fisher et al., 1986). The mechanism through which tat increases expression has been studied extensively in HIV-l (Cullen, 1990; Rosen and Pavlakis, 1990) but as yet it is not completely understood. In HIV-l, tat has been demonstrated to exhibit its effects at several levels including increases in transcription initiation, elongation, and translation efficiency (Braddock et al., 1989; Feinberg et al., 1991; Hauber and Cullen, 1988; Laspia et al., 1989; Selby et a/., 1989; SenGupta et a/., 1990; Wright et al., 1986). Increasing evidence exists that HIV-1 tat works through the HIV-l mRNA leader sequences which are predicted by computer modeling to form a stem-bulge-loop structure (Berkhout and Jeang, 1989; Cordingley et a/., 1990; Muesing et al., 1987). These sequences, designated the TAR region, have been shown to function by positioning tat near the promoter elements most proximal to the start site of transcription (Berkhout et a/., 1990; Berkhout and Jeang, 1992; Selby and Peterlin, 1990; Southgate et al., 1990). Computer analysis of the mRNA leader sequences of HIV-2 and SIV reveal extensive secondary structure (Arya, 1988). HIV-2 has been shown to have two functional TAR elements (Fenrick et a/., 1989). Transactivation by tat has been demonstrated in the SIV system (Arya et a/., 1987; Arya, 1988) and 3’ dele- tions up to +lOO maintained transactivation (Viglianti and Mullins, 1988). These studies were designed to investigate whether the sequence differences between the two viral LTRs led to differences in transactivation. Here, it is reported that the SIV,,,239 full-length LTR is consistently transactivated to a slightly higher level (1.6- to 2.0-fold) than the SIV,,,251 LTR by the SIV tat protein in HUT-78 and U937 cells. HUT-78 cells are permissive for the growth of both viruses (Naidu et al., 1988). Constructs in which sequences 3’to +18 as well as 5’to -225 were deleted were no longertransac-

tivated by tat. Further, the addition of bases +19 to +149 of SIV,,,239 to these constructs of both strains decreased the basal activity of the LTRs independent of orientation and reconstituted transactivation in an orientation dependent manner. As with the complete viral LTRs, the dl dr constructs again showed that the SIV,,,239 U3 region was more responsive to the tat protein than SIV,,,251 U3 region. Thus, the differences in the levels of transactivation between the two strains can be attributed to the sequences proximal to the promoter (-225 to +18). This region encompasses the Sp-1 -binding sites, where SIV,,,239 has 1 1 more bases than SIV,,,251, creating three potential Sp-1 binding sites, one more than SIV,,,251. Transactiva- tion was highest at 48 hr when CAT activity was mea- sured; however, increases in steady-state levels of RNA could not account for this transactivation. A time course analysis of transactivation showed that the RNA levels peaked prior to 48 hr after transfection while the CAT enzyme activity increases steadily during the 48 hr after transfection.

MATERIALS AND METHODS

Cell lines

The U937 cell line was obtained from T. Folks. HUT- 78 and U937 were maintained according to standard culture technique using RPMI medium with 10% fetal bovine serum (GIBCO) and 2 mlLI glutamine (GIBCO).

LTR CAT plasmids

Construction of LTR-CAT constructs has been described elsewhere (Anderson and Clements, 1991). The SIV-LTR DNA full-length (fl) and deletion (dl and d2) constructs which are upstream of the reporter gene CAT are outlined in Fig. 1. These are the 239dl:CAT, 239d2:CAT, 251dl:CAT, and 251d2:CAT constructs. The dl A constructs (239dlA:CAT and 25 1 d 1 A:CAT) are the d 1 constructs with a mutation in the protein-binding sequence from -155 to -139, changing this region from GAAGGCTAACCGCAAGA to TCGTTAACCAACAGCTG. This mutation abrogates binding of cellular proteins in DNase I protection assays (Anderson and Clements, 1991). Insertion of bases +19 to +149, sequences containing the direct repeats, was accomplished as follows. First, using the polymerase chain reaction (or PCR) (Saiki et a/., 1988) with standard conditions and primers homologous to the 5’ and 3’ ends of the sequences between +19 and +149, this region was synthesized. Digestion of the dl :CAT, d2:CAT, and dl A:CAT constructs of both strains with BarnHI and subsequent removal of the 5’ overhang with mung bean nuclease (Pharmacia) left

DIFFERENTIAL TRANSACTIVATION MEDIATED BY PROMOTER SEQUENCES 561

blunt ends in the linearized plasmid vector. The PCR- amplified fragment containing the direct repeat (dr) of 239 was ligated into the individual vectors with T4 DNA ligase and transformed into Escherichia co/i DH5 cells. Individual clones were sequenced using the dideoxy method of Sanger et al. (1977) to determine position, orientation, and correct sequence. These constructs were designated as follows: strain of SIV, size of construct, direct repeat (dr):CAT, i.e., 239dl dr:CAT. 239dl dr-:CAT is identical to 239dl dr:CAT, except that the direct repeat regions is in the opposite orientation. An additional G residue in the LTR-CAT vectors between LTR sequences + 18 and + 19 resulted from this cloning procedure. Another control plasmid was generated with both d2 constructs by digesting the LTR- CATvector with HindIll, filling the ends and cloning the fragment as described above. The HindIll site is 30 bp down stream of the BamHl site located at the end of the deletion constructs. This has the net effect of plac- ing the polylinker DNA in between the direct repeat region and the U3 sequences. As with all dr constructs, only the 239 dr region was used.

Computer analysis of RNA leader structure

Two-dimensional RNA structures were generated using the algorithms of the HIBIO macDNA ProSlS pro- gram of Hitachi.

Transfections and CAT assays

HUT-78 and U937 cell lines were transfected by the DEAE-dextran dimethyl sulfoxide shock method (Lo- pata et a/., 1984). Replicates of 1 O7 cells transfected with 5 pg of the respective LTR-CAT construct and either 5 pg of salmon sperm DNA or 5 pg of pl 1 (a gift from Dr. Flossie Wong-Staal), an Okyama-Berg expression vector containing a cDNA for SIV-tat under the control of the SV40 promoter (Colombini et al., 1989). Titration of DNA determined these DNA conditions to be optimal. Cells were washed and incubated for 46-48 hr in RPMI medium containing 1 Oq/o fetal bovine serum. The cells were lysed and cell extracts were assayed for CAT activity as previously described (Hess et a/., 1989). Extracts were incubated for 43 hr in the presence of [‘4C]chloramphenicol (DuPont, NEN Re- search Products) and acetyl coenzyme A. The acety- lated form of chloramphenicol was separated from the unacetylated form by thin-layer chromatography and percentage conversion was determined by scintillation counting of the excised spots. Background activity of the pUC-CAT vector with no promoter was subtracted. The CAT assay was linear for the incubation time and protein concentration used for both HUT-78 and U937 cell extracts.

RNase protection assay

Duplicates of 3 x 1 O7 HUT-78 cells were transfected with 15 pg of LTR-CAT plasmid and either 15 pg of salmon sperm DNA or 15 pg of the pl 1 tat plasmid. Approximately lo7 cells were harvested for subsequent CAT assays, and the remaining cells were harvested for RNA at 46 to 48 hr, according to previously described methods (Chirgwin et a/., 1979) with modifi- cations (Hess et al., 1989). RNase protection assays were done as previously described (Veillette et al., 1987). The CAT RNA probe contained the first 252 bases of the CAT gene. Bases 149 to 255 of the human beta-actin gene (Ponte et a/., 1984) were cloned into the HindIll and Accl sites pGem-4Z and 32P-labeled antisense RNA was made as previously described (An- derson and Clements, 1991). The EcoRI restriction site was used to linearize the template, and T7 polymerase was used to transcribe 153 bp of antisense RNA. The 46 bases of RNA that is complimentary to polylinker DNA are digested in the RNase assay. The assay was repeated three times on the 239 constructs and yielded similar results. A representative X-ray film of one experiment was used for quantification by laser densitometry to determine band intensities. CAT RNA levels were normalized to beta-actin RNA concentra- tions. Time course analyses were carried out as described above, except each was scaled up four times. One-fourth of the cells from both of two independent transfection were harvested at various time points and RNA and protein activity were determined as described above.

RESULTS

Tat transactivation of LTR-CAT constructs

These studies were undertaken to examine whether differences in the U3 regions of the SIV,,,239 and SIV,,,251 LTR affected the level of viral transactivation. The tat plasmid pl 1 was used in these studies (Colombini et al., 1989). The amino acid sequence of tat protein encoded in the pl 1 plasmid is identical to the sequence of the SIV,,,251 tat and differs by only five amino acids from the SIV,,,239 tat. The five amino acid changes are outside of the functional domains identified for the HIV-l tat protein that include the cys- teine-rich region and the basic region (Hauber et al., 1989; Ruben et a/., 1989). The fl and dl, d2, and dl A constructs of the LTRs for 239 and 251 (Fig. 1; Ander- son and Clements, 1991) were cotransfected with tat- DNA or control DNA, into HUT-78 and U937 cells as described under Materials and Methods. In HUT-78 cells, both the 239 and 251 fl LTR constructs were induced to high levels of activity when cotransfected


-496 ; I I I I I I 1

B i

I I

fl '

dl

d2

d?z

FIG. 1, Diagram of SW-LTR sequences analyzed. (A) Oligonucleotides used for the PCR amplification of the SIV-LTR sequences and their positions within the SIV,,,251 LTR. The fl construct is from -469 to +310, dl is from -225 to +18, d2 is from -123 to +18, and the direct repeats (dr) is from +19 to +149. (B) Sequences amplified from both SIV,,251 and SIV,,,239 as shown and subcloned upstream of the bacterial CAT reporter gene of a pUC-CAT vector. Fragment labeled dr, which contains the direct repeats that confer far responsiveness was cloned immediately downstream of the dl , d2, and dl A constructs. The dr fragment was also positioned in the reverse orientation downstream of the dl constructs (dldr-:CAT) and also 30 bp downstream in the d2 constructs (d2hinddr:CAT). Sizes in base pairs are indicated on the right.

with tat when compared to cotransfection with control DNA, as analyzed by CAT activity (Table 1). No transactivation was seen with the 239dl:CAT, 251dl:CAT, 239d2:CAT, 251d2:CAT, 239dlA:CAT, and 251dlA: CAT constructs, which contain no direct repeats (data not shown). However, the 239fl:CAT constructs were consistently 1.6 to 2.0 times more responsive in both HUT-78 cells and U937 cells to the tat protein than 251 fl:CAT was responsive (Table 1 and data not shown). This was directly attributed to differences in the levels of activity in the presence of tat, since basal levels of activity of the 239 and 251 LTR constructs were identical.

Construction of tat responsive deletion mutants

It has been shown that HIV-l tat transactivation is mediated by the positioning of tat near the proximal promoter elements of the HIV-1 LTR (Berkhout et a/., 1990; Berkhout and Jeang, 1992; Selby and Peterlin, 1990; Southgate et a/., 1990). In order to assess the importance of the U3 regions of both SIV LTRs and in particular, the additional 11 base pairs (that creates a third Sp-1 site in 239), and the four nucleotide differences in 239, the region containing the direct repeats (+19 to +149) from the 239 molecular clone were

placed downstream of the dl , d2, and dl A constructs of 239 and 251 (Fig. 1). The dlA constructs are the same as dl except for a mutation in the upstream protein-binding sequence that eliminates protein binding to that sequence in HUT-78 and U937 cells (Fig. 2; Anderson and Clements, 199 1). The dr from 239 was used in all constructs to control for small differences between the viral sequences (Fig. 2). Thus, comparison of these constructs assayed only differences in the U3 regions. The predicted structure and stability (relative to free energy) of the leader sequences was determined as described under Materials and Methods. The 239 leader sequence as well as the dr region that was added to the deletion constructs were predicted to fold into the double stem and loop structure (data not shown) previously determined for SIV,,,251 (Arya, 1988). The direct repeat region in the opposite orientation (dr constructs) and 30 bp downstream (d2 hind constructs) was not predicted to produce the same structure.

Analysis of tat responsive deletion mutants

CAT assays were done in HUT-78 cells and data from a representative experiment in which transfections were done in triplicate are shown in Table 1.


TABLE 1

ACTIVITIESOF LTR-CATCONSTRUCTSIN HUT-78 CELLS

CAT constructa CAT activity

(standard deviation)b Fold induction

239:d 1 1 .OO (0.26) 239:d2 0.310 (0.028) 239:dl A 0.577 (0.075) 239:fl 0.256 (0.19) 239:fl + tat 78.5 (2.3) 307 239:dldr 0.1 16 (0.070) 239:dl dr +tat 14.8 (3.4) 128 239:d2dr 0.0890 (0.028) 239:d2dr +tat 10.7 (1.7) 120 239:dl Adr 0.150 (0.018) 239:dl Adr +tat 17.5 (2.4) 116 239:d2hinddr 1.192 (0.25) 239:d2hinddr+tat 21.1 (4.2) 17.7 239:dldr- 0.262 (0.11) 239:dldr- + tat 0.166 (0.081) 0.63

251:dl 1.69 (0.62) 251:d2 0.151 (0.028) 251:dlA 0.376 (0.18) 251 :fl 0.273 (0.089) 25 1 :fl +tat 49.5 (7.9) 181 251:dldr 0.207 (0.060) 251:dldr +tat 14.59 (2.1) 70.5 25 1 :d2dr <0.068 - 251 :d2dr .rtat 5.87 (0.48) >86.8 251:dlAr r 10.068 - 251:dlb Ir +tat 5.95 (0.50) >87.9 251 :d2r inddr 0.254 (0.043) 25 1 :d2’ linddr+tat 2.33 (0.51) 9.2 251:dldr- <0.068 - 251:dldr- +tat 10.068 - -

a Replicates of 1 O7 cells were cotransfected with 5 rg of LTR-CAT DNA and either 5 rg of control DNA or 5 pg of the raat expression plasmid where indicated. Cells were harvested at 46 hours and assayed as described in Materials and Methods.

b Values are from a representative experiment. Repeats of experiments yielded similar results. In this experiment, sets of triplicates were averaged, background was subtracted, values were normalized for protein concentration and then expressed relative to the 239dl :CAT construct (with standard deviation in parenthesis). All results were clearly within the linear range, except the co.068 esti- mates.

When the direct repeat region is placed in the deletion constructs, the basal level of CAT activity declines four- to ninefold. This is orientation-independent since insertion of the direct repeats in the opposite orientation leads to a similar decline (Fig. 3, Table 1). Transactivation was reconstituted in 239dl dr:CAT, 239d2dr:CAT, and 239dl Adr:CAT to approximately equal levels. Similarly, 251dldr:CAT, 251d2dr:CAT, and 251dlAdr:CAT were also transactivated, but to a lesser degree than the 239 constructs, implying that the differences between transactivation of the two LTRs is due to differences in these

sequences. The low basal activity of 251d2:CAT and 25 1 A:CAT and even lower basal activity when the direct repeats were added resulted in CAT activity near the bottom of the linear range of the assay. The exact activity, and hence, exact transactivation levels, were therefore not possible to determine. Since some of the replicates were at or near the bottom of the linear range, an estima- tion was made. However, the absolute CAT activity in the presence of tat was higher in 239d2dr:CAT and 239dlAdr:CAT than in 251d2dr:CAT, and 251dlAdr: CAT, respectively.

When testing for position dependence of the direct repeats, a surprising result was found. Instead of de- creasing basal activity, the insertion of the direct repeats 30 bp downstream in the polylinker lead to a twofold (251 d2hinddr:CAT) to fourfold (239d2hinddr: CAT) increase in basal activity (Fig. 3 and Table 1). While absolute CAT activity levels were high, transactivation was severely diminished. Whether this was due to the much higher basal levels of activity or to a decreased responsiveness to the tat protein is unclear.

Analysis of steady-state CAT RNA levels

The steady state mRNA levels of the CAT gene were assessed by RNase protection assay using a riboprobe complementary to the first 242 bases of the CAT gene as described under Materials and Methods. While the CAT activity or protein level was increased dramatically

-225 TATGAGGCAT ATGTTAGATA CCCAGAAGAG lTrGGAAGCA AGTCAGGCCl

GTCAGAGGAA GAGGTTAGAA -C CTTCTTAACA NF-KB

%XTGACAA GA:GGAAACT CGCTGAeACA GC&&XTT TCfACAAGGG

- ATATCACTGC All-TCGCTCT GTAn

CAGA GAGC CCTGGGAGGT TCTCTCCAGC ACi'jAGCAGGT AGAGCCTGGG d ma reDen*

GTTCCCTGC TAGACTCTCA CCAGCAC G GCCiGTGCTG GGCAGAGTGt c +149,

CTCCACGCTT GCTTGCTTAA AGACCTC,

FIG. 2. Comparison of the SIV,,,251 and SIV,,,239. DNA sequences surrounding the start of transcription (R or +l). The 251 sequence is shown. Nucleotide differences found in 239 are shown in bold, above the 251 sequence, and * is the site of the ll-bp insertion TACTGGGGAGG which creates a third, nonconsensus, Sp-1 site found in 239. Potential protein-binding sequences for NF- KB. Sp-1, and the TATA box are underlined. Actual protein-binding regions are denoted by an oval above the sequences. The stippled oval indicates the protected region that partially overlaps the most 3’ Sp-1 site, which is protected only in the 239 LTR. The direct repeats that form two stem loops are boxed and the dr region (see text) is included within the vertical slashed lines (+19 to +149).


A 23941 constructs

n

251d2 Constructs 1

d2 d2dr dZdrd2hlnd dfhind +h, l ta,

239d2 Constructs

-I-

/ d2 d2dr d2dr d2hlnddlhlnd

+t.t 4.t

D r 251dl Constructs

o/ dl dldr dldr dldr- dldr-

+,a, rt.,

E 23911 and 251fl Constructs 60

FIG. 3. CAT activity of LTR-CAT constructs. Data from Table 1 is depicted graphically with error bars indicating standard deviation, (A) Data from (A) the 239dl:constructs, (B) the 239d2 constructs, (C) the 25tdl constructs, (D) the 251d2 constructs, and (E) the 239fl and 251fl constructs. Values printed on the bar, followed by an X, indicate fold transactivation determined by dividing the activity of the dr containing construct in the presence rat by the activity of the same dr containing construct in the absence of fat. Values above the bars represent the activity of dr containing plasmid relative to identical plasmid without the dr region, i.e., in A 239dl dr:CAT is 12% as active as 239dl :CAT and transactivated by 126fold in the presence of fat.

by tat, the steady-state mRNA levels were increased only two- to fourfold or less in the 239 constructs. This small difference was observed in three independent experiments (data not shown). Only in the 239d2hind- dr:CAT construct was the difference in CAT activity comparable to the difference in RNA levels.

Recent experiments with HIV-l tat transactivation showed that the RNA levels peaked at 4 hr after fusion of HIV tat and HIV LTR-CAT-expressing cell lines (Drysdale and Pavlakis, 1991). To investigate the possi- bility that RNA peaked long before CAT protein levels with SIV-LTR-CAT constructs, a time course analysis was carried out. A representative experiment is shown (Fig. 4). A parallel time course analysis, as well as a

separate experiment, yielded similar results. The RNA levels in cells transiently transfected with 239fl:CAT and tat rose during the first 12 hr and was 10 times that of cells transfected with 239fl:CAT and control DNA at that time, as determined by laser densitometry. How- ever, by 48 hr, the RNA had dropped off to levels al- most equal to basal. During the same time, there was a linear rise in CAT activity in cells cotransfected with tat and in transactivation (Fig. 4B and 4C). No RNA was detected in cells transfected with the pUC-CATvector without SIV-LTR sequences (data not shown). The differences in RNA levels at the earlier time points are more reflective of the level of transactivation than the RNA levels at the 46-hr time point.


-tat

-CAT

6 12 22 46 6 12 22 46

-A&in

CAT Activity

12 -

10 -

E 7l ;

: 8 v -tat

- +tat

;

8 t

0

0 10 20 30 40 50

TIME (hours)

tinepoint 6hr 12hr 22hr 46hr fold incr. 0 25 95 240

C Transactivation

3o01

0 10 20 30 40

time

FIG. 4. Time course analysis of RNA levels and CAT protein activity. (A) Antisense RNA (500 cpm) of the first 252 bases of the CAT gene and 153 bases of anti-sense RNA of the beta-actin clone are shown on the left. The beta-actin probe contains 46 noncomplimentary bases and therefore is cleaved to 107 bases in the RNase protection assay. HUT-78 cells (12 X 10’) were transfected with 60 pg of 239fl:CAT and either 60 Fg of control DNA (-fat) or 60 Pg of the rat plasmid (+tad. At 6, 12, 22, and 46 hr post-transfection 2 X 1 O7 cells were harvested and RNA was extracted. Twenty micrograms of total cellular RNA was hybridized to 100,000 cpm of the labeled CAT antisense RNA and 2500 cpm of the labeled beta-actin antisense RNA before digestion with RNase A for 45 min at 37”. Reactions were extracted precipitated and run out on a 6% acrylamide-urea gel and exposed to X-ray film using intensifying screens to visualize the bands. Cells (1 07) were also harvested at 6, 12, 22, and 46 hr and protein extracts were prepared for CAT assays. (6) Plot of the CAT activity for time points in A in cells cotransfected with control DNA (-tat) or the fat plasmid (+taf). CAT assays were carried out as described under Materials and Methods except that extracts were incubated for 5 hr. The amount of transactivation is below each time point. (C) Transactivation levels are plotted against time.

DISCUSSION

In this study, we have found that although the promoters of SW,,,239 and SW,,,251 had comparable levels of basal activity, the SW,,,239 promoter was induced to a greater extent (1.6- to 2.0-fold) by the tat gene product in both HUT-78 and U937 cell lines. These differences were seen both in absolute CAT activity as well as in the level of tat induction. This paral- lels the higher levels of virus and viral RNA found in macaques experimentally infected with cloned SIV,,,239 compared to those infected with SW,,,251 (Sharma et al., 1992; M. C. Zink, personal communication). Previous experiments have shown that deletion

of sequences 5’to -225 in conjunction with deletion of sequences 3’to + 18 (d 1) led to increases in basal activity (Anderson and Clements, 1991). Addition of the direct repeat region (+ 19 to + 149) decreased basal activity, implying that the SIV-TAR region down-regulates expression from the promoter.

Previous examination of sequences in the U3 region showed that deletion of bases -225 to -123 had a more deleterious effect on the 251d2:CAT construct than the 239d2:CAT construct. This suggested that the promoter elements (NF-KB, Sp-1, and TATA box region) closest to the start site of transcription were less active in 251 than in 239. Thus, the upstream U3 sequences were more important for 251. The reconsti-


tution of tat responsiveness was achieved with the deletion mutants and the region just 3’ to the start site (+ 19 to +149). The differences seen in transactivation between 239dldr:CAT and 251dldr:CAT (1.8-fold) were similar to differences seen in the full-length LTR. In addition, the 239d2dr:CAT and 239dlSA:CAT data and approximation of the 251d2dr:CAT and 25 1 d 1 A:CAT data indicate that deletion of sequences between -225 and -123 and elimination of the protein binding sequences therein have little effect on transactivation. This is in agreement with the HIV-l model where position of tat and interaction of tat with promoter elements most proximal to the TATA box medi- ate transactivation (Berkhout et a/., 1990; Berkhout and Jeang, 1992; Selby and Peterlin, 1990; Southgate et a/., 1990). These data indicate that the most proximal promoter sequences which are more active in 239 than in 251 are important for transactivation and contain the elements that cause the differential transactivation of the two strains. It seems most probable that the difference between 239 and 251 is due to the 1 l- bp sequences in the 239 LTR which create an additional Sp-l-binding domain. DNase I protection analysis demonstrated that there is a protein-binding region just downstream from this 1 1 -bp insertion and it overlaps the 3’-most Sp-1 site in 239; however, no protein binding site was detected in this region of 25 1 (Ander- son and Clements, 1991). Thus, a major difference between these LTRs appear to be this 1 1-bp insertion and protein binding region. In addition, the four nucleotide differences could contribute to the differential transactivation.

While orientation of the direct repeat region had no effect on the decrease in basal activity, the transactivation was strictly dependent upon it. This has been demonstrated with HIV-1 (Muesing eta/., 1987). Somewhat surprising, however, was the finding that moving the direct repeat fragment 30 bp downstream actually leads to a four- and twofold increase in activity for the 239d2hinddr:CAT and 25ld2hinddr:CAT constructs, respectively. While absolute CAT activity was high in the d2hinddr:CAT constructs, transactivation was severely decreased. This result is far more dramatic than and in conflict with what has been shown for HIV-1 (Hauber and Cullen, 1988; Muesing et a/., 1987; Selby et a/., 1989) either indicating differences in assay conditions or reflecting different mechanisms involved in the SIV-TAR-tat system.

The fact that steady-state RNA levels at 46 hr after transfection do not reflect the differences seen at the protein level suggests several possibilities for transactivation mechanisms of SIV-tat. One is that SIV-tat ex- erts its effect at the level of translation as has been the interpretation of data gathered in the HIV-1 system

(Braddock et al., 1989; Hauber and Cullen, 1988; Sen- Gupta et al., 1990; Wright et al., 1986). Another is that under these assay conditions, the stability of the CAT protein is greater than that of the CAT RNA. Finally, recent findings in the HIV-l system have demonstrated a down-regulation of transcription due to tat at later time points (Drysdale and Pavlakis, 1991). This was due to decreases in transcription initiation. In these HIV-l experiments, the protein levels continued to increased over time, RNA levels decreased, and more utilization of CAT RNA was seen. An analogous situa- tion probably exists in the SIV system where the differences in CAT activity in the presence and absence of tat are more reflective of the differences in RNA seen early (12 hr) after transfection, than later (46 hr). Find- ings such as these suggest that assessment of the level at which tat is causing transactivation requires careful analysis of the kinetics of RNA synthesis and degradation. In contrast, the 239d2hinddr:CAT construct had steady-state RNA ratios comparable to protein ratios. Since basal levels were also extremely ele- vated, a novel mechanism for mRNA stabilization or translation may occur when the direct repeats are moved downstream. Further, the loss of transactivation could be due to high basal activity combined with an upper limit to transcription and/or translation. Fac- tors required for transcription and/or translation could be limiting or overwhelmed due to the high levels of activity. These studies of SIV transactivation suggest that the mechanism of action of SIV tat may differ slightly from HIV. In addition, the LTRs of different strains of SIV are transactivated to different levels.

ACKNOWLEDGMENTS

We thank Maryann Brooks for expert help in preparing the manu- script. We also thank Dr. Joanna Pyper for the actin-pGEM construct. This work was supported by grants from the National Institute of Health (Al27297 and Al28748).

REFERENCES

ANDERSON, M. G., and CLEMENTS, J. E. (1991). Comparison of the transcriptional activity of the LTRs of SIV,,,251 and SIV,,,239 in T-cell lines and macrophage cell lines. J. Viral. 65, 5 l-60.

ARYA, S. K.. BEAVER, B., JAGODZINSKI, L., ENSOLI, B., KANKI, P. J., AL- BERT, J., FENYO, E-M., BIBERFELD, G., ZAGURY, J. F., LAURE, F., ESSEX, M., NORRBY. E., WONG-STUL, F., and GALLO, R. C. (1987). New human and simian HIV-related retroviruses possess functional transactivator (tat) gene. Nature (London) 328, 548-550.

ARYA, S. K. (1988). Human and simian immunodeficiency retroviruses: Activation and differential transactivation of gene expression. AIDS Res. Hum. Retroviruses 4, 175-l 86.

BERKHOUT, B., GATIGNOL, A., RABSON, A., and JEANG, K. T. (1990). TAR-independent activation of the HV-1 LTR: Evidence that Tat requires specific regions of the promoter. Ce// 62, 757-767.

BERKHOUT, B., and JEANG, K. T. (1989). Transactivation of human immunodeficiency virus type 1 is sequence specific for both the sin-


gle-stranded bulge and loop of the trans-acting-responsive hair- pin: A quantitative analysis. J. Viral. 63, 5501-5504.

BERKHOUT, B., and JEANG, K. T. (1992). Functional roles for the TATA promoter and enhancers in basal and Tat-induced expression of the human immunodeficiency virus type I long terminal report. J. Viral. 66, 139-l 49.

BRADDOCK, M., CHAMBERS. A., WILSON, W.. ESNOUF, M. P., ADAMS, S. E., KINGSMAN, A. J., and KINGSMAN. S. M. (1989). HIV-1 Tat “acti- vates” presynthesized RNA in the nucleus. Cell 58, 269-279.

CHAKRABARTI. L., GUYADER, M., ALIZON, M., DANIEL, M. D., DESRO- SIERS. R. C., TIOLLAIS, P., and SONGIO, P. (1987). Sequence of simian immunodeficiency virus from macaque and its relationship to other human and simian retroviruses. Nature (London) 328, 543- 547.

CHATIS, P. A., HOLLAND, C. A., SILVER, J. E., FREDERICKSON, T. N., HOPKINS, N.. and HARTLEY, J. W. (1984). A 3’end fragment encom- passing the transcriptional enhancers of nondefective Friend virus confers erythroleukemogenicity on Moloney leukemia virus. /. Virol. 52, 248-254.

CHIRGWIN, S.. PFQYBYLA, A., MCDONALD, R., and RUT~ER, W. (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294-5299.

COLOMBINI, S., ARYA, S. K., REITZ, M. S., JAGODZINSKI, L., BEAVER, B., and WONG-STAAL, F. (1989). Structure of simian immunodeficiency virus regulatory genes. Proc. Nat/. Acad. Sci. USA 861, 48 13- 4817.

CORDINGLEY. M. G.. LAFEMINA. R. L.. CALLAHAN, P. L., CONDRA, J. H., SARDANA. V. V.. GRAHAM, D. J., NGUYEN, T. M., LEGROW, K., GOTLIB, L., SCHLABACH, A. J., and COLONNO, R. J. (1990). Sequence-specific interaction of Tat protein and Tat peptides with the transactivation-responsive sequence element of human immunodeficiency virus type 1 in vitro. Proc. Nat/. Acad. Sci. USA 87, 8985-8989.

CULLEN, 6. R. (1990). The HIV-1 Tat protein: An RNA sequence-specific processivity factor? Cell 63, 655-657.

DANIEL, M. D., LEIVIN, N. L., KING, N. W., KANNAGI, M., SEHGAL, P. K., HUNT, R. D., KANKI, P. J., ESSEX, M., and DESROSIERS, R. C. (1985). Isolation of T-cell tropic HTLV-Ill-like retrovirus from macaques. Science 228, 1201-1204.

DANIEL, M. D., LETUN, N. L., SEHGAL, P. K.. HUNSMANN, G., SCHMIDT, D. K., KING, N. W., and DESROSIERS, D. C. (1987). Long-term persistent infection of macaque monkeys with simian immunodeficiency virus. J. Gen. Viral. 68, 3183-3 189.

DANIEL, M. D., LI, Y.. NAIDU, Y. M., DURDA, P. J.. SCHMIDT, D. K., TROUP, C. D., SILVA, D. P., MACKEY, J. J.. KESTLER, H. W., SEHGAL, P. K., KING, N. W., OHTA, Y., HAYAMI, M., and DESROSIERS, R. C. (1988). Simian immunodeficiency virus from African green monkeys. J. viral. 62, 4123-4128.

DAMON, A., SODROSKI, J., ROSEN, C. A., GOH, W., and HASELTINE, W. A. (1986). The transactivator gene of the human T cell lymphotropic virus type Ill is required for replication. Cell 44, 941-947.

DRYSDALE, C. M., and PAVALAKIS, G. N. (1991). Rapid activation and subsequent down-regulation of the human immunodeficiency virus type I promoter in the presence of Tat: Possible mechanisms contributing to latency. J. Viral. 65, 3044-3051,

FEINBERG, M. B., BALTIMORE, D., and FRANKEL, A. D. (1991). The role of tat in the human immunodeficiency virus life cycle indicates a primary effect on transcriptional elongation. Proc. Nat/. Acad. Sci. USA 88,4045-4049.

FENRICK, R., MALIM, M., HAUBER, J., LE, S. Y., MAIZEL, J., and CULLEN, B. R. (1989). Functional analysis of the tat transactivator of human immunodeficiency virus type Il. J. Viral. 63, 5006-5012.

FISHER, A. G., FEINBERG, S. F., JOSEPHS, S. F., HARPER, M. E., MAR- SELLE, L. M., REYES, G., GONDA, M. A., ALDOVINI, A., DEBOUCK, C., GALLO, R. C., and WONG-STAAL, F. (1986). The transactivator gene

of HTLV-III is essential for virus replication. Nature (London) 320, 367-371.

FUKASAWA, M., MIURA, T.. HASEGAWA, A., MORIKAWA, S., TSUJIMOTO, H., MIKI, K., KITAMURA, T., and HAYAMI, M. (1988). Sequence of simian immunodeficiency virus from African green monkeys, a new member of the HIV/SIV group. Nature (London) 333, 457- 461.

HAHN, B. H., KONG, L. I., LEE, S.. KUMAR, P., TAYLOR, M. E., ARYA, S. K.. and SHAW, G. M. (1987). Relation of HTLV-4 to simian and human immunodeficiency viruses. Nature (London) 330, 184- 186.

HAUBER, J.. and CULLEN, B. R. (1988). Mutational analysis of the trans-activation-response region of the human immunodeficiency virus type 1 long-terminal repeat. J. Viral. 62, 673-679.

HAUBER, J., MALIM, M., and CULLEN, B. R. (1989). Mutational analysis of the conserved basic domain of human immunodeficiency virus tat protein. J. Viral. 63, 1 181-l 187.

HESS, J. L., SMALL, J. A., and CLEMENTS, J. E. (1989). Sequences in the visna virus long terminal repeat that control transcriptional activity and respond to viral trans-activation: Involvement of AP-1 sites in basal activity and trans-activation. J. Viral. 63, 3001-3015.

KANKI, P. J., MACLANE, M. F., KING, N. W., LETVIN. N. L., HUNT, R. D., SEHGAL, P. K., DANIEL, M. D., DESROSIERS, R. C.. and ESSEX, M. (1985). Serologic identification and characterization of a macaque T-lymphotropic retrovirus closely related to HTLV-III. Science 228, 1199-1201.

KANNAGI, M., YETZ, J. M., and LETVIN, N. L. (1985). In vitro growth characteristics of simian T-lymphotropic virus type Ill. Proc. Nat/. Acad. Sci. USA 82,7053-7057.

KESTLER, H. W., LI, Y., NAIDU, Y. M., BUTLER, C. V.. OCHS, M. F., JAENEL, G., KING, N. W., DANIEL, M. D., and DESROSIERS. R. C. (1988). Comparison of simian immunodeficiency virus isolates. Narure 331, 6 19-62 1.

KESTLER, H., KODAMA, T., RIGLER, D., REGIER, D., SEHGAL, P., DANIEL, M. D., KING, N. W., and DESROSIERS, R. (1990). Induction of AIDS in rhesus monkeys by molecularly cloned simian immunodeficiency virus. Science 248, 1109-l 1 12.

LASPIA, M.. RICE, A., and MA-HEWS, M. (1989). HIV-l tat protein increases transcriptional initiation and stabilizes elongation. Ce// 59, 283-292.

LENZ, J., CELANDER, D., CRO\NTHER, R., PATARCA, R., PERKINS, D., and HASELTINE, W. A. (1984). Determination of the leukaemogenicity of a murine retrovirus by sequences within the long terminal repeat. Nature (London) 308, 467-470.

LENIN, N. L., DANIEL, M. D., SEHGAL, P. K., DESROSIERS, R. C., HUNT, R. D., WALDRON. L. M., MACKEY, J. J., SCHMIDT, D. K., CHALIFOUX, L. V., and KING, N. W. (1985). Induction of AIDS-like disease in macaque monkeys with T cell tropic retrovirus STLV-III. Science 230,71-73.

LETVIN, N. L., EATON, K. A., ALDRICH, W. R., SEHGAL, P. K., BLAKE, B. I., SCHLOSSMAN, S. F., KING, N. W., and HUNT, R. D. (1983). Acquired immunodeficiency syndrome in a colony of macaque monkeys. Proc. Natl. Acad. Sci. USA 80, 27 18-2722.

LOPATA, M. A., CLEVELAND, D. W., and SOLLNER-WEBB, B. (1984). High level transient expression of a chloramphenicol acetyl-trans- ferase gene by DEAE-dextran mediated DNA transfection cou- pled with a dimethyl sulfoxide or glycerol shock treatment. Nucleic Acid Res. 12, 5707-57 17.

Lu, Y.. STENZEL. M., SODROSKI, J.. and HASELINE, W. A. (1989). Ef- fects of long terminal repeat mutations on human immunodeficiency virus type 1 replication. J. Viral. 63, 41 15-41 19.

MCENTEE, M. F., ZINK, M. C.. ANDERSON, M. G., FARZADEGAN, H., ADAMS, R. J., KENT. K.. STOTT, E. J., CLEMENTS, J. E., and NARAYAN,


0. (1992). Neutralizing antibodies modulate replication of SlVmac in primary macaque macrophages. J. Viral., in press,

MUESING, M., SMITH, D., and CAPON, D. (1987). Regulation of mRNA accumulation by a human immunodeficiency virus trans-activator protein. Cell 48, 691-701.

NAIDU, Y. M., KESTLER Ill, H. W., LI, Y., BUTLER, C. V., SILVA, D. P., SCHMIDT, D. K., TROUP, C. D., SEHGAL, P. K., SONIGO, P., DANIEL, M. D., and DESROSIERS, R. C. (1988). Characterization of infectious molecular clones of simian immunodeficiency virus (SIV,,,) and human immunodeficiency virus type 2: Persistent infection of rhesus monkeys with molecular cloned SIV,,,. J. Viral. 62, 4691- 4696.

PONTE, P., NG, S. Y., ENGEL, J., GUNNING, P., and KEDES, L. (1984). Evolutionary conservation in the untranslated regions of actin mRNAs: DNA sequence of a human beta-actin cDNA. Nucleic Acids Res. 12, 1687.

RINGLER. D. J.. WYAND, M. S., WALSH, D. G., MACKEY, J. J., CHALIFOUX, L. V., POPOVIC, M., MINASSIAN, A. A., SEHGAL, P. K., DANIEL, M. D., DESROSIERS, R. C., and KING, N. W. (1989). Cellular localization of simian immunodeficiency virus in lymphoid tissues. I. Immuno- chemistry and electron microscopy. Am. J. Patho/. 134, 373-383.

ROSEN, C. A., and PAVAUKIS, G. N. (1990). Tat and rev: Positive regu- lators of HIV gene expression. AIDS 4, 499-509.

RUBEN, S., PERKINS, A., PURCELL, R., JOUNG, K.. SIA, R., BURGHOFF, R., HASELTINE. W. A., and ROSEN, C. A. (1989). Structural and functional characterization of human immunodeficiency virus tat protein. J. Viral. 63, l-8.

SAIKI, R. K., GELFAND, D. H., STOFFEL, S., SCHARF, S. J., HIGUCHI, R., HORN, G. T., MULLIS, K. B., and ERLICH, H. A. (1988). Primer-di- rected enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239, 487-491.

SANGER, F., NICKLEN, S., and COULSON, R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. NaN. Acad. Sci. USA 74, 5463-5467.

SELBY, M., BAIN, E., Lucm, P. A., and PETERLIN, B. (1989). Structure

sequence and position of the stem-loop in TAR determine transcriptional elongation by tat through the HIV-1 long terminal repeat. Genes Dev. 3, 547-558.

SELBY, M. J., and PETERLIN, B. M. (1990). Transactivation by HIV-1 Tat via a heterologous RNA binding protein. Cell 62, 769-776.

SENGUPTA, D. N., BERKHOUT, B., GATIGNOL, A., ZHOU, A., and SILVER- MAN, R. H. (1990). Direct evidence for translational regulation by HIV-1 leader RNA and Tat protein. Proc. Nat/. Acad. Sci. USA 87, 7492-7496.

SHARMA, D. P., ANDERSON, M. G., ZINK, M. C., ADAMS, R. J., DONNEN- BERG, A. D., CLEMENTS, J. E., and NARAYAN, 0. (1992). Pathogene- sis of acute infection in Rhesus Macaques with a Lymphocyte-tropic strain of SIVmac. [submitted for publication]

SODROSKI, J., PATARCA, R., ROSEN, R., WONG-STAAL, F.. and HASEL- TINE. W. A. (1985). Location of the transactivation region on the genome of human T-cell lymphotropic virus type Ill. Science 229, 74-77.

SOUTHGATE, C., ZAPP, M. L., and GREEN, M. R. (1990). Activation of transcription by HIV-l Tat protein tethered to nascent RNA through another protein. Nature (London) 345, 640-642.

TSICHLIS, P. N., and COFFIN, J. M. (1980). Recombinants between endogenous and exogenous avian tumor viruses: Role of the C region and other portions of the genome in the control of replication and transformation. 1. Viral. 63, 1891-l 896.

VEILLETTE, A., Foss, F., SAUSVILLE, E., BOLEN, J. B., and ROSEN, N. (1987). Expression of the Ick tyrosine kinase gene in human colon carcinoma and other non-lymphoid tumor cell lines. Oncogene Res. 1, 357-374.

VIGLIANTI, G. A., and MULLINS, J. I. (1988). Functional comparison of transactivation by simian immunodeficiencyvirus from rhesus macaques and human immunodeficiency virus type 1. J. Viral. 62, 4523-4532.

WRIGHT, C. M., FELBER, B. K., PASKALIS, H., and PAVLAKIS, G. N. (1986). Expression and characterization of the trans-activator of HTLV-III/LAV virus. Science 234, 988-992.

Documents

Two strains of SIVmac show differential transactivation mediated by sequences in the promoter