Mutational Inactivation of an Inhibitory Sequence ... - Journal of Virology

6 downloads 0 Views 1MB Size Report
STEFAN SCHWARTZ,1t MEL CAMPBELL,1 GEORGIOS NASIOULAS,2 JEANETTE HARRISON,' ...... sites of interaction for Rev, Tev, and Rex proteins within the.
Vol. 66, No. 12

JOURNAL OF VIROLOGY, Dec. 1992, p. 7176-7182 0022-538X/92V127176-07$02.00/0 Copyright © 1992, American Society for Microbiology

Mutational Inactivation of an Inhibitory Sequence in Human Immunodeficiency Virus Type 1 Results in RevIndependent gag Expression STEFAN SCHWARTZ,1t MEL CAMPBELL,1 GEORGIOS NASIOULAS,2 JEANETTE BARBARA K. FELBER,2 AND GEORGE N. PAVLAKIS1*

HARRISON,'

Human Retrovirus Section1 and Human Retrovirus Pathogenesis Group,2 National Cancer Institute-Frederick Cancer Research and Development Center, ABL-Basic Research Program, Frederick, Maryland 21702-1201 Received 22 June 1992/Accepted 4 September 1992

We have characterized an inhibitory RNA element in the human immunodeficiency virus type 1 (HIV-1) gag coding sequence that prevents gag expression. The inhibition exerted by this element could be overcome by the presence of the Rev-responsive element in cis and of Rev protein in trans. To understand the mechanism of function, we inactivated the inhibitory element by mutagenesis while maintaining an intact gag coding region. A constitutive high level of Rev-independentgag expression was achieved only after the introduction of 28 point mutations over a large region of 270 nucleotides within the gag coding region. To our knowledge, this is the first demonstration of inactivation of a negative RNA element within a coding region without alteration of the expressed protein. Elimination of the inhibitory element in the p178`9 region, named INS-1, offered the opportunity to detect a second inhibitory element in the gag-pol region. The presence of either INS element is sufficient to inhibit gag expression, demonstrating that multiple INS elements acting independently can inhibit HIV RNA expression. Expression of gag from Rous sarcoma virus, a retrovirus that does not require Rev-like regulatory proteins, revealed that the Rous sarcoma virus pl99g region does not contain inhibitory elements. These results demonstrate the presence of a strong inhibitory element acting at the level of mRNA and provide a general method for the removal of such elements from mRNA coding regions. The inhibitory element functions in the absence of any HIV-1 proteins, suggesting that cellular factors are responsible for this inhibition. The interaction of the Rev regulatory protein of human immunodeficiency virus type 1 (HIV-1) with its RNA target, named the Rev-responsive element (RRE), is necessary for expression of the viral structural proteins (for reviews, see references 4, 26, and 28). Rev acts by promoting the nuclear export and increasing the stability of the RRE-containing mRNAs. Recent results also indicate a role for Rev in the efficient polysome association of these mRNAs (1, 5). Since the RRE-containing HIV-1 mRNAs do not efficiently produce protein in the absence of Rev, it has been postulated that these mRNAs are defective and contain inhibitory sequences named INS, CRS, or IR (3, 8, 14, 21, 29, 33). The nature and function of these inhibitory sequences have not been characterized in detail. It has been postulated that inefficiently used splice sites may be necessary for Rev function (2); the presence of such splice sites may confer Rev dependence on HIV-1 mRNAs. Analysis of HIV-1 hybrid constructs led to the initial characterization of some inhibitory sequences in the gag and pol regions of HIV-1 (3, 21, 33). We have previously reported the identification of an inhibitory RNA element located in the coding region of the p179a9 matrix protein of HIV-1 (33). It was shown that this sequence acted in cis to inhibit HIV-1 tat expression after insertion into a tat cDNA. The inhibition could be overcome by Rev-RRE, demonstrating that this element plays a role in regulation by Rev. This element does not contain any known splice sites, indicating that sequences other than splice sites prevent expression in

the absence of Rev. Here we further analyze the region for inhibition by introducing multiple point mutations in the gag region. These mutations do not alter the amino acid sequence of the produced protein. We show that it is possible to eliminate INS by mutagenesis, resulting in constitutively high gag expression in the absence of Rev. These results demonstrate the presence of a strong inhibitory element acting at the level of mRNA and provide a general method for the removal of such elements from mRNA coding regions. necessary

MATERIALS AND METHODS Plasmid constructions. Plasmid pl7R has been described previously as pNL17R (33). Plasmid p17 was generated from pl7R by digestion with Asp 718 and religation. This procedure deleted the RRE and HIV-1 sequences spanning nucleotides 8021 to 8561 upstream of the 3' long terminal repeat (LTR). To generate mutants of pl7?a8, we subcloned the p176ag coding sequence into a modified pBluescript vector (Stratagene) and generated single-stranded uracil-containing DNA. Site-directed mutagenesis was performed as described before (19, 34). Clones containing the appropriate mutations were identified by DNA sequencing. Plasmid p37M1234 was derived from p17M1234 and contains in addition the p24&ag coding region. The stop codon introduced after p178ag was removed, and a stop codon was introduced after p24&a&. This plasmid produced p37gag protein. pl9(sd+) and pl9R(sd+) were derived from p17 by deleting the entire p179a9 coding region, 6 nucleotides upstream of the p17'a9 AUG, and 9 nucleotides of linker sequences 3' of the p179a9 stop codon. This fragment was

* Corresponding author. t Present address: Department of Virology, Karolinska Institute, Stockholm, Sweden.

7176

VOL. 66, 1992

replaced by a polymerase chain reaction (PCR)-amplified DNA fragment containing eight nucleotides upstream of the Rous sarcoma virus (RSV) gag AUG and the RSV p199a9 coding sequence, immediately followed by an introduced translational stop codon. The RSV gag fragment was derived from the infectious RSV proviral clone SR-A (39), kindly provided by S. Hughes. pl9R(sd+) contains the same RRE fragment as pl7R. The splice donor of RSV gag was mutated by site-directed mutagenesis, changing GTG (Val) to ATA (Ile), generating plasmids p19 and pl9R, respectively. Plasmid p55 contains the entire p55gag coding region, while plasmid pcgag also containspol sequences to the BalI site (9). DNA transfection. HLtat cells (30) were transfected by the calcium coprecipitation technique (11) as described before (9), with 5 jig of gag-expressing plasmids in the absence or presence of 2 Fg of the Rev-expressing plasmid pL3crev (9). The total amount of DNA in transfection mixes was adjusted to 17 ,g per 0.5 ml of precipitate per 60-mm plate by using pUC19 carrier DNA. Cells were harvested 20 h after transfection, and cell extracts were subjected to electrophoresis on denaturing 12.5% polyacrylamide gels and analyzed by immunoblotting with either serum from an HIV-i-infected patient (Scripps) or a rabbit anti-p199&9 serum, kindly provided by V. Vogt. pRSV-luciferase (6), which contains the firefly luciferase gene linked to the RSV LTR promoter, was used as an internal standard to control for transfection efficiency, and luciferase activity was quantitated as described before (37). RNA preparation. HLtat cells were transfected as described above and harvested 20 h posttransfection. Twenty micrograms of cytoplasmic RNA (9, 12) was subjected to Northern (RNA blot) analysis as described before (14). The filters were hybridized to a nick-translated PCR-amplified DNA fragment spanning ,nucleotides 8304 to 9008 in the HIV-1 3' LTR or to a 1-actin fragment (5). Nucleotide sequence accession number. The nucleotide sequence in Fig. 4 has been deposited in GenBank under accession number L04602. RESULTS Coding sequence of the HIV-1 matrix protein pl7W contains an inhibitory element that prevents Gag production in the absence of Rev. To study the inhibitory element in pl7?a8, we constructed a p178a9 expression plasmid (p17, Fig. 1). The p17'a9 sequence was engineered to contain a translational stop codon immediately after the protease cleavage site, and thus it produced only p176a9. The major 5' splice site of HIV-1 upstream of the gag AUG has been deleted from this vector (9). To investigate whether plasmid p17 could produce p17'a9 in the absence of Rev and the RRE, we transfected p17 into HLtat cells (30). These cells constitutively produce HIV-1 Tat protein, which is necessary for transactivation of the HIV-1 LTR promoter. Plasmid p17 was transfected in the absence or presence of Rev, and the production of pl7'ag was analyzed by Western immunoblotting. Very low levels of pl?7'g protein were produced (Fig. 2A). The presence of Rev did not increase gag expression, as expected, since this mRNA did not contain the RRE. We next constructed a plasmid that contained both the p17'a9 coding sequence and the RRE (pl7R, Fig. 1). Like p17, this plasmid produced very low levels of p17'a& in the absence of Rev. High levels of p17'a9 were produced only in the presence of Rev (Fig. 2A). These experiments suggested that an inhibitory element was located in the p17'a9 coding

INACTIVATION OF INHIBITORY HIV-1 SEQUENCE

7177

A gag p24

p7

Vpr

vpt

p17 p24 p7vpvu

R

GOp

pl7gag

(HIV-1)

)Hl]

p17

pl7gag (HIV-1) )-Em RRE

p1

7R

p19gag (RSV)

HI

P19

p19gag (RSV) -LhIiI

p1

)NIL

p17M1234

)HLIE

p37M1234

OR

RRE

pl7gag (HIV-1)

p37gag (HIV-1) p24

FIG. 1. (A) Structure of the HIV-1 genome. Boxes indicate the different viral genes. (B) Structure of the gag expression plasmids. Plasmid p17 contains the complete HIV-1 5' LTR and sequences up to the BssHII restriction site at nucleotide (nt) 257. (The nucleotide numbering refers to the revised nucleotide sequence of the HIV-1 molecular clone pHXB2 [24].) This sequence is followed by the p179a9 coding sequence, spanning nt 336 to 731 (represented as an open box), immediately followed by a translational stop codon and a linker sequence. Adjacent to the linker is the HIV-1 3' LTR from nt 8561 to the last nucleotide of the U5 region. Plasmid pl7R contains in addition to the 330-nt StyI fragment encompassing the RRE (37) (represented as a stippled box) 3' to the p179W9 coding sequence. The RRE is followed by HIV-1 sequences from nt 8021 to the last nucleotide of the U5 region of the 3' LTR. Plasmids p19 and pl9R were generated by replacing the HIV-1 p17'a9 coding sequence in plasmid p17 with the RSV p191a9 coding sequence (represented as a black box). Plasmid pl7M1234 is identical to p17 except for the presence of 28 silent nucleotide substitutions within the gag coding region, indicated by XXX. Plasmid p37M1234 was derived from pl7M1234 and contains in addition the p249'9 coding region. Wavy lines represent plasmid sequences. sequence and that the effects of this element were corrected by the Rev-RRE interaction. Expression experiments with various eucaryotic vectors have indicated that several other retroviruses do not contain such inhibitory sequences within their coding sequences (see, for example, references 23 and 40). To verify these results with our expression vectors, we replaced the p179a9 gene of HIV-1 in plasmid p17 with the coding sequence for the p19Ia9 matrix protein of RSV strain SR-A. The resulting plasmid, p19 (Fig. 1), was identical to plasmid p17 except for the gag coding sequence. The production of p19&a8 protein from plasmid p19 was analyzed by Western immunoblotting, which revealed that this plasmid produced high levels of p19ra9 (Fig. 2B). The presence of the RRE in cis (plasmid pl9R, Fig. 1B) and of Rev protein in trans did not further increase gag expression. These experiments demonstrated that the pl9Wag coding sequence of RSV, in contrast to p17&a of HIV-1, could be efficiently expressed in this vector, indicating that the gag region of RSV did not contain any inhibitory elements. The p19 coding region contains the 5' splice site used to

7178

J. VIROL.

SCHWARTZ ET AL. fl-:

C.

N

A

A N N

,NA "A9

9

r

A.,',

N

1F-7C

2

Re

Z--

1.

r-

0-

+ 1 Rev

:wp

a ,S

-I

e

iR

-

A

q'-,\

t.

B

N~ *

cN~c ~ N~

m*

pl7gag

4 l

*

actir

FIG. 3. Cytoplasmic mRNA analysis. HLtat cells were transfected with the indicated plasmids in the absence or presence of Rev. Twenty micrograms of cytoplasmic RNA prepared from the transfected cells was analyzed by Northern blotting. The filter was hybridized with the HIV-specific probe (top) and reprobed with the ,-actin probe (bottom).

Q

*, *

- p1 9 gag

FIG. 2. gag expression from the different vectors. (A) HLtat cells were transfected with plasmid p17, pl7R, or p17M1234 in the absence (-) or presence (+) of Rev. The transfected cells were analyzed by immunoblotting with serum from an HIV-1-infected patient. (B) Plasmids pl9(sd+), pl9R(sd+), p19, and pl9R were transfected into HLtat cells in the absence or presence of Rev, as indicated. The transfected cells were analyzed by immunoblotting with a rabbit anti-RSV p19"ag serum. HIV-1 (A) or RSV (not shown) virion proteins served as markers in the same gels. The positions of p179a9 and pl9gag are indicated.

generate the RSV env mRNA, which is located downstream of the gag AUG. This 5' splice site has been mutated in the expression vectors shown in Fig. 2B from GTG to ATA. Comparisons of gag expression from the two vectors, one containing the splice site and the other not, revealed that the presence or absence of this splice site did not significantly affect p199a9 expression (Fig. 2B). On the other hand, the HIV-1 p17 expression plasmid did not contain any known splice sites and yet was not expressed in the absence of Rev (Fig. 2A). These results further indicate that sequences other than inefficiently used splice sites are responsible for inhibition of HIV gag expression. Inhibitory element acts in cis to decrease the levels of mRNA and protein in the absence of Rev. We next asked whether the inhibitory effect of the pl79a9 coding sequence was also detectable at the mRNA level. Northern blot analysis of cytoplasmic RNA extracted from HLtat cells transfected with p17 or pl7R demonstrated that they produced low mRNA levels (Fig. 3). An increase in p17R mRNA levels was observed in the presence of Rev. Similar levels of expression of a cotransfected RSV-luciferase gene were measured. As an additional control, the levels of actin mRNA were also measured (Fig. 3, bottom). These results established that the inhibitory element in gag also affects the mRNA levels and are in agreement with our previous findings (33). Quantitations of the mRNA and protein levels produced by p17R in the absence and presence of Rev were

performed by scanning densitometry of appropriate serial dilutions of the samples; the difference was greater at the level of protein (60- to 100-fold) than at the level of mRNA (less than 10-fold). This result is compatible with previous findings of effects of Rev on mRNA localization and polysomal loading of both gag and env mRNAs (1, 5, 8, 9, 22). Northern blot analysis of the mRNAs produced by the RSV gag expression plasmids revealed that p19 produced high mRNA levels (data not shown). This confirmed that the pl9gag coding sequence of RSV does not contain inhibitory elements. Together, these results established that gag expression in HIV-1 is fundamentally different from that in RSV. The HIV-1 p179a9 coding sequence contains a strong inhibitory element, while the RSV p199a9 coding sequence does not. Inhibitory element can be inactivated by multiple nucleotide substitutions that do not alter the amino acid composition of p179a9 protein. To investigate the exact nature of the inhibitory element in HIV-1 gag, we performed site-directed mutagenesis of the p179a9 coding sequence with four different oligonucleotides, as indicated in Fig. 4. Each oligonucleotide introduced several point mutations over an area of 19 to 22 nucleotides. These mutations did not affect the amino acid sequence of the pl7?ag protein because they introduced silent codon changes. First, all four oligonucleotides were used simultaneously in mutagenesis with a single-stranded DNA template, as described before (19, 34). This allowed the simultaneous introduction of many point mutations over a large region of 270 nucleotides in vector p17. A mutant containing all four oligonucleotides, as determined by sequence analysis, was isolated and named p17M1234. Compared with p17, this plasmid contained a total of 28 point mutations, distributed primarily in regions with high AU content.

The phenotype of the mutant was assessed by transfections into HLtat cells and subsequent analysis of p175a5 expression by immunoblotting. Interestingly, p17M1234 produced high levels of p1785a protein, similar to those produced by pl7R in the presence of Rev (Fig. 2A). This result demonstrated that the inhibitory signals in pl7gag mRNA had been inactivated in plasmid p17M1234. The presence of Rev protein did not increase expression from p17M1234.

VOL. 66, 1992

INACTIVATION OF INHIBITORY HIV-1 SEQUENCE

7179

336

atg ggt gcg aga gcg tca gta tta agc ggg gga gaa tta gat cga tgg gaa aaa att cgg

Ml

396

aa aaa tat aaa tta aaa cat ata gta tgg G G C G C G C C

tta agg cca gqg

aaa gag

456

cta gaa cga ttc gca gtt aat cct ggc ctg tta gaa aca tca gaa ggc tgt aga caa ata

M2

516

ctg gga cag cta caa cca tcc ctt cag aca qqa tca gaa gaa ctt aga tca tta tat aat C C C C G G

M3 576 aca qta qca acc ctc tat tct qtq cat caa age ata gag ata aaa gac acc aag qgaa gct C C C G C G

M4

636

tta gac aag ata gag gaa gag caa aac aaa aat aag aaa aaa qca cag caa qca qca gct G TCC G G G C 696

gac aca q a cac agc aat cag gtc agc caa aat tac

FIG. 4. Nucleotide sequence of the p179a9 region. The locations of the four oligonucleotides (Ml through M4) used to generate all mutants are underlined. The nucleotide substitutions introduced by each mutagenesis oligonucleotide are indicated below the coding sequence. Numbering starts from nucleotide + 1 of the viral mRNA.

Thus, p179a9 expression from the mutant p17M1234 displayed the same general properties as expression of the p199a9 of RSV, that is, a high constitutive level of Revindependent gag expression. Northern blot analysis revealed that the cytoplasmic levels of mRNA produced by p17M1234 were higher than those produced by p17 (Fig. 3). To further examine the nature and exact location of the minimal inhibitory element, we mutated the p179a9 coding sequence in plasmid p17 with only one of the four mutated oligonucleotides at a time. This procedure resulted in four mutant plasmids, named p17M1, p17M2, p17M3, and p17M4 according to the oligonucleotide each plasmid contains. None of these mutants produced significantly higher levels of p17'a9 protein than plasmid p17 (Fig. 5), indicating that the inhibitory element was not affected. We next mutated the p17 coding sequence with two oligonucleotides at a time. The resulting mutants were named p17M12, p17M13, p17M14, p17M23, p17M24, and p17M34. Protein production from these mutants was minimally increased compared with that from p17, and it was considerably lower than that from p17M1234 (Fig. 5). In addition, a triple oligonucleotide mutant, p17M123, also failed to express high levels of p179a9 (data not shown). These findings suggest that multiple inhibitory signals are present in the coding sequence of pl7?a8. Alternatively, a single inhibitory element may span a large region whose inactivation requires mutagenesis with more than two oligonucleotides. This possibility is consistent with our previous data, suggesting that a 218-nucleotide inhibitory element in the p179a9 coding sequence is required for

,1+N + I.111 9 Q 9 9 9 q 9 Q " 9 It

,'

Ift

I&

9I9

,

pl7gag

FIG. 5. gag expression by different mutants. HLtat cells were transfected with the plasmids indicated above the lanes. Plasmid p17R was transfected in the absence (-) or presence (+) of Rev; the other plasmids were analyzed in the absence of Rev. p179a9 production was assessed by immunoblotting as described in the legend to Fig. 2.

strong inhibition of gag expression. Further deletions of this sequence resulted in gradual loss of inhibition (33). The inhibitory element may coincide with a specific secondary structure on the mRNA. We are currently investigating whether a specific structure is important for the function of the inhibitory element. Demonstration of additional INS elements within the gag-pol region. To examine the effect of removal of the inhibitory sequence (INS) on the expression of the complete gag gene, we introduced the 28 silent nucleotide substitutions of pl7gag described above into a vector expressing p379ag protein, which is the precursor of pl7&ag and p240ag, or in the complete p55gag gene. The resulting plasmids, p37M1234, p55M1234, and pcgagMl234 (Fig. 6), were transfected into HLtat cells. pcgagMl234 contains a portion of the pol gene in addition to the p55fag gene. The expressed proteins were quantitated by a p249ag antigen capture assay and compared with the levels of Gag produced by their wild-type counterparts (p37, p55, and pcgag, Fig. 6). This comparison showed that p37M1234 produced high levels of Gag, indicating that no major INS elements are contained within the p249ag coding region (Fig. 6). In contrast, low levels of Gag expression from p55M1234 and pcgagMl234 were observed. These results indicate the presence of an additional INS in the gag-pol region, as was suggested earlier from deletion experiments (33). Therefore, multiple independent INS elements are found within the HIV genome.

DISCUSSION In this report, we demonstrate that mutational inactivation of an inhibitory sequence in the p17'a coding sequence of HIV-1 results in Rev-independent gag expression. Our experiments demonstrate that the Rev-RRE interaction corrects preexisting defects on the viral structural mRNAs caused by specific RNA regions (INS) which inhibit expression. In the absence of INS, HIV-1 expression would be similar to that of simpler retroviruses and would not require Rev. Thus, INS elements are necessary components of Rev regulation. Sequence comparisons suggest that the INS element identified here is conserved in all HIV-1 isolates, although this has not been verified experimentally. The majority (25 of 28) of the mutated nucleotides in gag are conserved in all HIV-1 isolates, while 22 of 28 nucleotides are also conserved in HIV-2 (24). Functionally similar ele-

7180

J. VIROL.

SCHWARTZ ET AL. gag

LE}I1m

-{I

p37

[Ell}

-LII]

p55

1

EII I- 51

P p17 p24 p7

147

43 51

-L

cgag 0

EI

{] p37M1234

Ixxxl

L[L ElxxlI LELI-XX

WM

-EL]

p55M1234

-EL]

cgagMl234

10000

I41424

p24gag (pg/mi) FIG. 6. Quantitation of gag expression by an antigen capture assay. HLtat cells were transfected with the indicated plasmids, and on the following day, cell extracts were analyzed for gag expression with a commercial antigen capture assay for p249a9 (Coulter). Shaded areas indicate the approximate locations of the INS elements in the gag region. The amount of p249ag produced is shown to the right of each bar.

ments may exist in cellular RNAs with short half-lives. Recent reports suggest the presence of sequences affecting

mRNA levels within the coding regions of cellular mRNAs (36). The successful elimination of coding-region elements described for the first time here may provide a general method for the elimination of such elements from coding regions without alteration of the protein produced. Since functionally similar cis-acting regulatory sequences may be present in cellular mRNAs, further analysis of these HIV regulatory elements should lead to a better understanding of viral expression and of the posttranscriptional mechanisms of gene regulation. The accumulated experimental evidence suggests that multiple INS or CRS elements exist within HIV coding regions (3, 21, 33). Analysis of the multiple INS elements has revealed that they may inhibit expression efficiently even when found alone in an mRNA. This represents a significant problem in designing experiments to demonstrate their elimination, since subsets of these INS elements in various combinations in general have severe effects on the overall levels of mRNA and protein expression. The strategy we chose to demonstrate the function of such elements has been (i) identification of the minimal INS elements with appropriate indicator vectors and (ii) elimination of the INS by mutagenesis. Elimination of all regions with INS activity is necessary for constitutive Rev-independent expression, since elimination of only some INS elements does not increase expression significantly (see Fig. 6). Several lines of evidence indicate that all lentiviruses and other complex retroviruses, such as the human T-cell lymphotropic virus group, contain similar INS regulatory elements. We have identified strong INS elements in the gag region of human T-cell lymphotropic virus type I and simian immunodeficiency virus (unpublished data). This suggests that these elements are found in all complex retroviruses and may be responsible for some of the biological characteristics of these retroviruses. Since INS elements inhibit expression, it must be concluded that their presence is advantageous to the virus, because otherwise they would be rapidly eliminated by mutations. Thus, the presence of inhibitory sequences is a property specific to Rev-encoding complex

retroviruses and is required for regulation by the Rev protein. Therefore, the inhibitory sequences are important determinants of the expression levels of HIV-1 and probably of all complex retroviruses. The p179a9 coding sequence has a high content of A and U nucleotides, unlike the coding sequence of p196a9 of RSV (25, 33). Four regions with high AU content are present in the p179a9 coding sequence and have been implicated in the inhibition of gag expression in deletion mutagenesis experiments (33). Lentiviruses have a higher AU content than the mammalian genome. Regions of high AU content are found in the gag-pol and env regions, while the multiply spliced mRNAs have a lower AU content (25), supporting the possibility that the inhibitory elements are associated with mRNA regions with high AU content. It has been shown that a specific oligonucleotide sequence, AUUUA, found at the AU-rich 3' untranslated regions of some unstable mRNAs may confer RNA instability (35). Although this sequence is not present in the pl7gag sequence, it is found in many copies within the gag-pol region. The association of instability elements with AU-rich regions is not universal, since other HIV regions which show a strong inhibitory activity in our vectors are not AU rich. These observations suggest the presence of more than one type of inhibitory sequence. In addition to reducing the AU content, some of the mutations introduced in plasmid p17 changed rarely used codons to codons more favored for human cells. Although the use of rare codons could be an alternative explanation for poor HIV gag expression, this type of translational regulation is not favored by our results, since the presence of Rev corrects the defect in gag expression. In addition, we have reported that, in other vectors, translation through INS regions is not required for inhibition of expression (33). The observations that the inhibitory sequences act in the absence of any other viral proteins and that they can be inactivated by mutagenesis suggest that these elements may be targets for the binding of cellular factors that interact with the mRNA and inhibit posttranscriptional steps of gene expression. The interaction of HIV-1 mRNAs with such factors may cause nuclear retention, resulting in either further splicing or rapid degradation of the mRNAs. It has

VOL. 66, 1992

been proposed that components of the splicing machinery interact with splice sites in HIV-1 mRNAs and modulate mRNA expression (2, 3, 20). However, it is not likely that the inhibitory elements described here are functional 5' or 3' splice sites. Thorough mapping of HIV-1 splice sites performed by several laboratories with the reverse transcriptase-PCR technique failed to detect any splice sites within gag (10, 13, 27, 30-32). The suggestions that Rev may act by dissociating unspliced mRNA from the spliceosomes (2) or by inhibiting splicing (17) are not easily reconciled with the knowledge that all retroviruses produce structural proteins from mRNAs that contain unutilized splice sites. Splicing of all retroviral mRNAs, including HIV-1 mRNAs in the absence of Rev, is inefficient compared with the splicing of cellular mRNAs (16-18, 38). The majority of retroviruses do not produce Rev-like proteins, yet they efficiently express proteins from partially spliced mRNAs, suggesting that inhibition of expression by unutilized splice sites is not a general property of retroviruses. Experiments with constructs expressing mutated HIV-1 gag and env mRNAs lacking functional splice sites showed that only low levels of these mRNAs accumulated in the absence of Rev and that their expression was Rev dependent (8, 9, 22). This led to the conclusion that Rev acts independently of splicing (9, 22) and to the proposal that inhibitory elements other than splice sites are present on HIV-1 mRNAs (9, 14, 29). The inhibitory elements may be involved in the establishment of a state of low virus production in the host. We have proposed that HIV-1 has adapted a mode of expression mimicking cellular genes that are rapidly up- and downregulated, such as c-myc, c-fos, and granulocyte-macrophage colony-stimulating factor (7, 15, 35, 36, 41, 42). Like HIV-1 mRNAs, the mRNAs of these genes contain intragenic inhibitory elements, which in some cases coincide with areas of high AU content. These cellular mRNAs may be regulated by factors similar in function to those affecting HIV-1 expression. Therefore, detailed analysis of the inhibitory elements in HIV-1 may contribute to the understanding of gene regulation in general. HIV-1 and the other lentiviruses cause chronic active infections that are not cleared by the immune system. It will be interesting to examine the possibility that complete removal of the INS elements from the lentiviral genome would result in constitutive expression. This could prevent the virus from establishing a latent infection and escaping immune system surveillance. Our success in increasing expression of pl7?a8 by eliminating the INS indicates that, in the future, we could produce lentiviruses without any negative elements. Such lentiviruses may provide a novel approach to the production of attenuated vaccines. The identification of similar INS elements in simian immunodeficiency virus indicates that this virus may provide a convenient model with which to test this hypothesis. ACKNOWLEDGMENTS We are grateful to S. Hughes, V. Vogt, and M. Federspiel for plasmids and anti-p199a9 antiserum; M. Powers for oligonucleotide synthesis; S. Hughes, P. Johnson, D. D'Agostino, L. Solomin, and M. Saltarelli for discussions; and A. Arthur for editing. This research was sponsored by the National Cancer Institute, DHHS, under contract N01-CO-74101 with ABL. REFERENCES 1. Arrigo, S. J., and I. S. Y. Chen. 1991. Rev is necessary for translation but not cytoplasmic accumulation of HIV-1 vif, vpr, and env/vpu2 RNAs. Genes Dev. 5:808-819.

INACTIVATION OF INHIBITORY HIV-1 SEQUENCE

7181

2. Chang, D. D., and P. A. Sharp. 1989. Regulation by HIV Rev depends upon recognition of splice sites. Cell 59:789-795. 3. Cochrane, A. W., K. S. Jones, S. Beidas, P. J. Dillon, A. M. Skalka, and C. A. Rosen. 1991. Identification and characterization of intragenic sequences which repress human immunodeficiency virus structural gene expression. J. Virol. 65:5305-5313. 4. Cullen, B., and W. Greene. 1989. Regulatory pathways governing HIV-1 replication. Cell 58:423-426. 5. D'Agostino, D. M., B. K. Felber, J. E. Harrison, and G. N. Paviakis. 1992. The Rev protein of human immunodeficiency virus type 1 promotes polysomal association and translation of gag mRNA. Mol. Cell. Biol. 12:1375-1386. 6. de Wet, J. R., K. V. Wood, M. DeLuca, D. R. Helinski, and S. Subramani. 1987. Firefly luciferase gene: structure and expression in mammalian cells. Mol. Cell. Biol. 7:725-737. 7. Drysdale, C. M., and G. N. Paviakis. 1991. Rapid activation and subsequent downregulation of the human immunodeficiency virus type 1 promoter in the presence of Tat: possible mechanisms contributing to latency. J. Virol. 65:3044-3051. 8. Emerman, M., R. Vazeux, and K. Peden. 1989. The rev gene product of the human immunodeficiency virus affects envelopespecific RNA localization. Cell 57:1155-1165. 9. Felber, B. K., M. Hadzopoulou-Cladaras, C. Cladaras, T. Copeland, and G. N. Paviakis. 1989. Rev protein of human immunodeficiency virus type 1 affects the stability and transport of the viral mRNA. Proc. Natl. Acad. Sci. USA 86:1495-1499. 10. Garrett, E. D., L. S. Tiley, and B. R. Cullen. 1991. Rev activates expression of the human immunodeficiency virus type 1 vif and vpr gene products. J. Virol. 65:1653-1657. 11. Graham, F. J., and A. J. Van der Eb. 1973. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52:456-460. 12. Greenberg, M. E., and E. Ziff. 1984. Stimulation of 3T3 cells induces transcription of the c-fos proto-oncogene. Nature (London) 311:433-437. 13. Guatelli, J. C., T. R. Gingeras, and D. D. Richman. 1990. Alternative splice acceptor utilization during human immunodeficiency virus type 1 infection of cultured cells. J. Virol. 64:4093-4098. 14. Hadzopoulou-Cladaras, M., B. K. Felber, C. Cladaras, A. Athanassopoulos, A. Tse, and G. N. Pavlakis. 1989. The rev (trs/art) protein of human immunodeficiency virus type 1 affects viral mRNA and protein expression via a cis-acting sequence in the env region. J. Virol. 63:1265-1274. 15. Jones, T., and M. Cole. 1987. Rapid cytoplasmic turnover of c-myc mRNA: requirement of the 3' untranslated sequences. Mol. Cell. Biol. 7:4513-4521. 16. Katz, R. A., and A. M. Skalka. 1990. Control of retroviral splicing through maintenance of suboptimal splicing signals. Mol. Cell. Biol. 10:696-704. 17. Kjems, J., A. D. Frankel, and P. A. Sharp. 1991. Specific regulation of mRNA splicing in vitro by a peptide from HIV-1 rev. Cell 67:169-178. 18. Krainer, A. R., G. C. Conway, and D. KozaL 1990. Purification and characterization of pre-mRNA splicing factor SF2 from HeLa cells. Genes Dev. 4:1158-1171. 19. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc. Natl. Acad. Sci. USA 82:488-492. 20. Lu, X., J. Heimer, D. Rekosh, and M.-L. Hammarskjold. 1990. Ul small nuclear RNA plays a direct role in the formation of a rev-regulated human immunodeficiency virus env mRNA that remains unspliced. Proc. Natl. Acad. Sci. USA 87:7598-7602. 21. Maldarelli, F., M. A. Martin, and K. Strebel. 1991. Identification of posttranscriptionally active inhibitory sequences in human immunodeficiency virus type 1 RNA: novel level of gene regulation. J. Virol. 65:5732-5743. 22. Malim, M. H., J. Hauber, S. Le, J. V. Maizel, and B. R. Cullen. 1989. The HIV-1 rev transactivator acts through a structured target sequence to activate nuclear export of unspliced viral mRNA. Nature (London) 338:254-257. 23. Morris-Vasios, C., J. P. Kochan, and A. M. Skalka. 1988. Avian sarcoma-leukosis virus pol-endo proteins expressed indepen-

7182

24.

25. 26. 27.

28. 29.

30.

31.

32.

SCHWARTZ ET AL.

dently in mammalian cells accumulate in the nucleus but can be directed to other cellular compartments. J. Virol. 62:349-353. Myers, G., B. Korber, J. A. Berzofsky, R. Smith, and G. N. Pavlakis. 1991. Human retroviruses and AIDS: a compilation and analysis of nucleic acid and amino acid sequences. Los Alamos National Laboratory, Los Alamos, N.Mex. Myers, G., and G. N. Paviakis. 1992. Evolutionary potential of complex retroviruses, p. 1-37. In J. Levy (ed.), The Retroviridae. Plenum Press, New York. Pavlakis, G. N., and B. K. Felber. 1990. Regulation of expression of human immunodeficiency virus. New Biol. 2:20-31. Robert-Guroff, M., M. Popovic, S. Gartner, P. Markham, R. C. Gallo, and M. S. Reitz. 1990. Structure and expression of tat-, rev-, and nef-specific transcripts of human immunodeficiency virus type 1 in infected lymphocytes and macrophages. J. Virol. 64:3391-3398. Rosen, C. A., and G. N. Paviakis. 1990. Tat and Rev: positive regulators of HIV gene expression. AIDS 4:499-509. Rosen, C. A., E. Terwilliger, A. Dayton, J. G. Sodroski, and W. A. Haseltine. 1988. Intragenic cis-acting art gene-responsive sequences of the human immunodeficiency virus. Proc. Natl. Acad. Sci. USA 85:2071-2075. Schwartz, S., B. K. Felber, D. M. Benko, E. M. Fenyo, and G. N. Pavlakis. 1990. Cloning and functional analysis of multiply spliced mRNA species of human immunodeficiency virus type 1. J. Virol. 64:2519-2529. Schwartz, S., B. K. Felber, E. M. Fenyo, and G. N. Pavlakis. 1990. Env and Vpu proteins of human immunodeficiency virus type 1 are produced from multiple bicistronic mRNAs. J. Virol. 64:5448-5456. Schwartz, S., B. K. Felber, and G. N. Paviakis. 1991. Expression of human immunodeficiency virus type-1 vif and vpr mRNAs is Rev-dependent and regulated by splicing. Virology 183:677-686.

J. VIROL.

33. Schwartz, S., B. K. Felber, and G. N. Pavlakis. 1992. Distinct RNA sequences in the gag region of human immunodeficiency virus type 1 decrease RNA stability and inhibit expression in the absence of Rev protein. J. Virol. 66:150-159. 34. Schwartz, S., B. K. Felber, and G. N. Paviakis. 1992. Mechanism of translation of monocistronic and multicistronic human immunodeficiency virus type 1 mRNAs. Mol. Cell. Biol. 12:207-219. 35. Shaw, G., and R. Kamen. 1986. A conserved AU sequence from the 3' untranslated region of GM-CSF mRNA mediates selective mRNA degradation. Cell 46:659-667. 36. Shyu, A.-B., J. Belasco, and M. Greenberg. 1991. Two distinct destabilizing elements in the c-fos message trigger deadenylation as a first step in rapid mRNA decay. Genes Dev. 5:221-231. 37. Solomin, L., B. K. Felber, and G. N. Paviakis. 1990. Different sites of interaction for Rev, Tev, and Rex proteins within the rev-responsive element of human immunodeficiency virus type 1. J. Virol. 64:6010-6017. 38. Stoltzfus, C. M., and S. J. Fogarty. 1989. Multiple regions in the Rous sarcoma virus src gene intron act in cis to affect the accumulation of unspliced RNA. J. Virol. 63:1669-1676. 39. Weiss, R., N. Teich, H. Varmus, and J. Coffin. 1985. RNA tumor viruses: molecular biology of tumor viruses. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 40. Wills, J. W., R. C. Craven, and J. A. Achacoso. 1989. Creation and expression of myristylated forms of Rous sarcoma virus Gag protein in mammalian cells. J. Virol. 63:4331-4343. 41. Wilson, T., and R. Treisman. 1988. Removal of polyA and consequent degradation of c-fos mRNA facilitated by 3' AUrich sequences. Nature (London) 336:396-399. 42. Wisdom, R., and W. Lee. 1991. The protein-coding region of c-myc mRNA contains a sequence that specifies rapid mRNA turnover and induction by protein synthesis inhibitors. Genes Dev. 5:232-243.