Synergistic stimulation of HIV-1 rev-dependent export of unspliced mRNA to the cytoplasm by hnRNP A1

Synergistic Stimulation of HIV-1 Rev-dependent

cy virus

Article No. jmbi.1998.2473 available online at http://www.idealibrary.com on J. Mol. Biol. (1999) 285, 1951±1964

Export of Unspliced mRNA to the Cytoplasm byhnRNP A1

Isabel Najera, Marion Krieg and Jonathan Karn*

Medical Research Council The structural and accessory proteins of human immunode®cien
Laboratory of Molecular type 1 are expressed by unspliced or partially spliced mRNAs. Ef®cient Biology, Hills Road, CambridgeCB2 2QH, UK
*Corresponding author

Present address: M. Krieg, NeuroUniversitaÈ t Freiburg, Neuropatholo64, Freiburg 79106, Germany.

Abbreviations used: RRE, Rev resinstability element; hnRNP, heterogribonucleoprotein; NRS, negative reNLS, nuclear localisation sequence;signal; HIV, human immunode®cienacting repressor; CRM1, chromosommaintenance.

E-mail address of the [email protected]

0022-2836/99/051951±14 $30.00/0

transport of these mRNAs from the nucleus requires the binding of theviral nuclear transport protein Rev to an RNA stem-loop structure calledthe RRE (Rev response element). However, the RRE does not permit Revto stimulate the export of unspliced mRNAs from the ef®ciently splicedb-globin gene in the absence of additional cis-acting RNA regulatory sig-nals. The p17gag gene instability (INS) element contains RNA elementsthat can complement Rev activity. In the presence of the INS elementand the RRE, Rev permits up to 30 % of the total b-globin mRNA to beexported to the cytoplasm as unspliced mRNA. Here, we show that aminimal sequence of 30 nt derived from the 50 end of the p17 gag geneINS element (50 INS) is functional and permits the export to the cyto-plasm of 14 % of the total b-globin mRNA as unspliced pre-mRNA. Gelmobility shift assays and UV cross-linking experiments have shown thatheterogeneous nuclear ribonucleoprotein (hnRNP) A1 and a cellularRNA-binding protein of 50 kDa form a complex on the 50 INS. Mutantsin the 50 INS that prevent hnRNP A1 and 50 kDa protein binding areinactive in the transport assay. To con®rm that the hnRNP A1 complex isresponsible for INS activity, a synthetic high-af®nity binding site forhnRNP A1 was also analysed. When the high af®nity hnRNP A1 bindingsite was inserted into the b-globin reporter, Rev was able to increase thecytoplasmic levels of unspliced mRNAs to 14 %. In contrast, the mutanthnRNP A1 binding site, or binding sites for hnRNP C and L are unableto stimulate Rev-mediated RNA transport. We conclude that hnRNP A1is able to direct unspliced globin pre-mRNA into a nuclear compartmentwhere it is recognised by Rev and then transported to the cytoplasm.

# 1999 Academic Press

Keywords: splicing; commitment factors; mRNA export; hnRNP proteins

Introduction

zentrum dergie, Breisacherstrasse

ponse element; INS,eneous nucleargulatory sequence;NES, nuclear exportcy virus; CRS, cis-al region 1

ing author:

Retroviral replication requires the simultaneoussynthesis and export to the cytoplasm of theunspliced virion RNA and spliced subgenomicmRNAs. In order to express these distinct mRNAspecies, retroviruses have developed a number ofstrategies to regulate splicing rates and defeat cel-lular mechanisms designed to prevent the appear-ance of unspliced mRNA in the cytoplasm. In Roussarcoma virus, splicing is moderated by the use ofa suboptimal 30 branch point at the env gene spliceacceptor (Katz et al., 1988) as well as a negative

# 1999 Academic Press

mailto:[email protected]

http://www.idealibrary.com

regulatory sequence (NRS) that acts as a bindingsite for splicing factors, such as the SR proteins,

0

To achieve a balance between the expression ofthe fully spliced, singly spliced, and unspliced

1952 Control of RNA Export by Rev and hnRNP A1

and partially inhibits the recognition of the 5 splicedonor (Cook & McNally, 1998; McNally &Beemon, 1992). Specialised mechanisms are alsoneeded to ensure ef®cient transport of theunspliced mRNAs from the cellular nucleus.Viruses such as Mason-P®zer virus contain RNAelements that provide signals for nuclear export(Bray et al., 1994; Ernst et al., 1997).

Human immunode®ciency virus (HIV) producesa wider range of subgenomic mRNAs than simpleretroviruses such as Rous sarcoma virus or Mason-P®zer virus, and has evolved correspondinglymore elaborate post-transcriptional control mech-anisms to regulate RNA synthesis. Over 30 distinctmRNAs falling into three major classes are pro-duced during HIV infection: doubly spliced 1.8 kbRNAs encoding the tat, rev and nef regulatorygenes, singly spliced 4 kb RNAs for the vif, vpr,and vpu/env genes and, ®nally, the unspliced 9 kbvirion RNA which also acts as the mRNA for thegag/pol gene (Arrigo et al., 1990; Kim et al., 1989;Purcell & Martin, 1993; Schwartz et al., 1990).Ordered expression of these diverse populations ofmRNAs is controlled by the viral regulatory pro-tein Rev (Feinberg et al., 1986; Sodroski et al., 1986),which promotes the export of the 4 kb and 9 kbmRNAs from the nucleus (Malim et al., 1989).

Rev activity requires a 351 nt RNA elementcalled the Rev response element (RRE) which islocated in the env gene and is therefore present oneach of the HIV mRNAs encoding the viral struc-tural and accessory proteins (Malim et al., 1989;Mann et al., 1994; Rosen et al., 1988). The RRE actsas a speci®c binding site for Rev (Daly et al., 1989;Heaphy et al., 1990; Zapp & Green, 1989). Initially,a monomer of Rev binds to a speci®c high-af®nitysite within the RRE (Bartel et al., 1991; Heaphyet al., 1991). Subsequently, up to nine further mol-ecules bind to the RRE in a co-operative mannerthrough protein-protein and protein-RNA inter-actions (Charpentier et al., 1997; Heaphy et al.,1991; Kjems et al., 1991; Malim & Cullen, 1991;Mann et al., 1994; Zemmel et al., 1996).

The Rev protein carries a nuclear localisation sig-nal (NLS), an arginine-rich sequence which over-laps the RRE-binding domain, as well as an``activation'' domain which contains a leucine-richnuclear export signal (NES; BoÈhnlein et al., 1991;Malim et al., 1991). These two signals enable Rev toshuttle between the nucleus and the cytoplasm(Fischer et al., 1995; Meyer & Malim, 1994; Stauberet al., 1995). The binding of Rev to viral mRNA cre-ates an RNP ®lament with the NES displayed onthe surface (Henderson & Percipalle, 1997) andprovides a transient ``tag'' which directs unsplicedtranscripts carrying the RRE to a speci®c nuclearexport pathway involving the transport factorsexportin (CRM1p) and Ran GTPase (Fornerod et al.,1997; Neville et al., 1997; Stade et al., 1997; Wolffet al., 1997; Zolotukhin & Felber, 1997).

mRNAs, HIV uses a variety of different mechan-isms. Splicing rates are reduced by the virusthrough the use of sub-optimal splice acceptorsequences (Chang & Sharp, 1989; Dyhr-Mikkelsen& Kjems, 1995; McNally & Beemon, 1992; O'Reillyet al., 1995). In addition, at least two types of cis-acting RNA signals in addition to the RRE arerequired for Rev-dependent export of unsplicedmRNA to the cytoplasm. First, splicing rates arereduced by cis-acting inhibitory sequences locatedadjacent to the splice acceptor sequences of the®rst and second exons of the tat and rev genes(Amendt et al., 1994; Barksdale & Baker, 1995;Staffa & Cochrane, 1995). HIV-1 also encodes aseries of poorly characterised regulatory sequencescalled the instability (INS) or cis-acting repressor(CRS) sequences (Maldarelli et al., 1991; MikaeÂlianet al., 1996; Nasioulas et al., 1994; Rosen et al., 1988;Schwartz et al., 1992). One possible role for theseelements is to ensure that unspliced pre-mRNAsare not simply degraded in the nucleus prior toexport to the cytoplasm by sequestering the RNAin nuclear sub-compartments that are inaccessibleto the splicing machinery (Berthold & Maldarelli,1996; Chang & Sharp, 1989; MikaeÂlian et al., 1996).

The use of multiple cis-acting RNA signals byHIV to regulate mRNA synthesis has made it dif®-cult to study the Rev regulatory system using het-erologous reporter genes. For example, the simpleaddition of an RRE to the ef®ciently spliced b-glo-bin gene does not permit ef®cient Rev-dependentexport of unspliced pre-mRNA (Chang & Sharp,1989; MikaeÂlian et al., 1996). Recently, we havefound that cis-acting sequences from the p17gagINS element can be used to complement RRE func-tion in the b-globin gene reporter system(MikaeÂlian et al., 1996). The two elements act syner-gistically and permit the export of up to 30 % ofthe total b-globin transcripts as unspliced b-globinmRNA. We have now used this assay to identifycellular factors that interact with INS elements andcomplement Rev-dependent export activity.

Here, we show that the HIV-1 p17 gag INSelement carries a high af®nity binding site forhnRNP A1 and permits the formation of a complexcontaining hnRNP (heterogeneous nuclear ribonu-cleoprotein) A1 and a second RNA-binding proteinof 50 kDa. Genetic experiments using the b-globingene reporter demonstrate that hnRNP A1 is ableto co-operate with Rev to promote the export ofunspliced mRNA from the nucleus. Since hnRNPA1 is known to regulate distal 50 splice site selec-tion and it is known to be able to shuttle betweenthe nucleus and the cytoplasm (Burd & Dreyfuss,1994; Michael et al., 1995; PinÄ ol-Roma & Dreyfuss,1992), it seems likely that it contributes to theexport of unspliced b-globin by creating a pool ofunspliced pre-mRNA in the nucleus that is separ-ated from the splicing machinery and can be recog-nised by Rev.

Results

Figure 1. (a) Structure of rabbit b-globin gene reporterconstruct (pIM41) and position of probes used to detectspliced and unspliced mRNAs. The p17gag INSsequence corresponds to nucleotides 787 to 1150 of theNL4-3 subtype of HIV-1 provirus (Adachi et al., 1986).The RRE sequence corresponds to nucleotides 7701 to8051. The plasmid contains four additional base-pairs atthe ®lled-in BamH1 (Ba*) site in the second exon (indi-cated in white). The position of the 253 nt L12 riboprobeused in this study is indicated as well as the size of theRNA fragments resulting from protection with thespliced and unspliced b-globin RNA. Note that the L12probe sequence is complementary to the sequence foundin plasmids carrying the ®lled-in BamH1 site. The L12probe also carries a 30 extension of an unrelatedsequence (open bar). After hybridisation to theunspliced transcripts from pIM41, or related plasmids,and digestion with RNase T1, a 196 nt fragment is pro-duced, whereas the spliced RNA produces a 72 nt frag-ment. (b) RNase protection assay. HeLa cells were co-transfected with 500 ng of reporter plasmids in theabsence (ÿ) or presence (�) of 50 ng of the Rev-expres-sing plasmid pF31. The plasmids analysed included thewild-type p17gag INS element (pIM41; Complete, WT)or the inactive mutant INS element (pIM93; Complete,MUT) and a series of deletions of the INS element(pMK48, F1, nt 817 to 893; pMK50, F2, nt 984 to 1036;pMK53, F3, nt 1038 to 1103; and pMK33, F4, nt 1107 to1149). Cytoplasmic RNA was isolated 48 hours aftertransfection and assayed by RNase protection exper-iments using the L12 probe. The protected fragmentswere fractionated on a 6 % polyacrylamide gel contain-ing 7 M urea. The positions of the protected fragmentscorresponding to the unspliced and spliced b-globinmRNAs are indicated to the left of the gel.

Control of RNA Export by Rev and hnRNP A1 1953

Co-operation between INS elements and RRE

Using a reporter plasmid which carried the rab-bit b-globin gene, MikaeÂlian et al. (1996) showedthat both the RRE and the INS element of thep17gag gene of HIV-1 are needed before Rev canstimulate high levels of unspliced b-globin mRNAexport from the nucleus. The structure of the repor-ter construct used in these experiments, pIM41(MikaeÂlian et al., 1996), is shown in Figure 1(a). InpIM41, the b-globin gene (containing three exons)was expressed under the control of the CMVimmediate early promoter. The INS and RREsequences were inserted into the third exon andthe complete reporter gene carried the SV40 polya-denylation signal at its 30 end.

The most accurate method to measure Rev-dependent transport of unspliced mRNA is to com-pare the ratio of unspliced to spliced RNA in thecytoplasm. In the experiments described below,spliced and unspliced RNAs were detected in thecytoplasm by RNase T1 protection assays using asingle 253 nt probe (L12) that spans the boundariesof the second exon and intron of the b-globin gene.The probe detects the spliced RNA as a 72 nt frag-ment that corresponds to part of the second exonand the unspliced RNA as an overlapping 196 ntfragment that extends into the second intron(Figure 1(a)). Since a single probe is able to detectboth RNA species, the ratio of spliced to unsplicedmRNAs provides a measurement of Rev activitythat is independent of the variations in absoluteexpression levels that arise in different transfectionexperiments.

As shown in Figure 1(b) (lane 1), after transfec-tion of cells by pIM41 only 1.9 % of the b-globinmRNA appearing in the cytoplasmic fraction isunspliced in the absence of Rev. Co-transfection ofpIM41 with the Rev-expressing plasmid pF31(Mann et al., 1994) resulted in a dramatic increaseof unspliced b-globin mRNA in the cytoplasmicfraction (Figure 1(b), lane 2). In this experiment,29.5 % of the total b-globin mRNA found in thecytoplasm was unspliced, a level of unsplicedmRNA that is close to the mean values obtained inreplicate experiments (Table 1).

Control experiments using the pIM93 plasmidcarrying a fully functional RRE and an inactiveINS element (MikaeÂlian et al., 1996) demonstratethat there are strict sequence requirements for INSfunction (Figure 1(b), lanes 3 and 4). In the absenceof Rev only 1.9 % of the cytoplasmic mRNA isunspliced. Addition of Rev produces a smallincrease in the transport of unspliced mRNA, withonly 5.8 % of the total cytoplasmic RNA unspliced.Previous experiments have also shown that theINS element alone is unable to promote the exportof unspliced mRNA from the nucleus (MikaeÂlianet al., 1996). Thus, the RRE and the INS element actsynergistically to permit Rev-dependent export ofunspliced b-globin mRNA from the nucleus.

A minimal INS element HnRNP A1 and a 50 kDa nuclear proteinrecognise the 50 INS RNA

Table 1. Rev-dependent export of unspliced b-globin mRNA requires a functional INS site

Unspliced cytoplasmic RNA (%)a

Plasmid INS elementb Minus Rev Plus Rev

pIM41 p17 gag INS 2.20 � 0.96 (4) 29.53 � 10.24 (4)pIN27 50 INS 1.48 � 0.45 (4) 14.25 � 6.80 (4)pMK86 hnRNP A1 0.83 � 0.23 (5) 13.56 � 3.98 (5)pMK87 hnRNP A1* 0.89 � 0.42 (5) 3.11 � 1.50 (5)

a The per cent of unspliced mRNA in cytoplasm was determined from quantitative analysis of the RNase protection assay gels bydensitometry. Values are averages with standard deviations. The number of replicates is shown in parentheses.

b Sequences for the 50 INS and hnRNP A1 binding sites are given in Table 2. hnRNP A1* indicates the mutant binding sitesequence.


In order to determine which regions of thep17gag INS element are responsible for its activity,we analysed a series of deletions of the 363 ntp17gag INS element from HIV-1NL4-3. The p17 INSelement was initially divided into four fragments:F1 (corresponding to nucleotides 817 to 893 of theproviral DNA sequence), F2 (984 to 1036), F3 (1038to 1103) and F4 (1107 to 1149). The 50 fragment, F1(Figure 1(b), lanes 5 and 6) was able to co-operatewith Rev and increase levels of unspliced mRNAin the cytoplasm to 25.6 %, a level similar to thatobtained using the full-length INS element.Strongly reduced levels of cytoplasmic unsplicedRNA were observed using the F2 (Figure 1(b), lane8; 3.6 %) or F3 (Figure 1(b), lane 10; 6.1 %) INS frag-ments in the presence of Rev. F4 (Figure 1(b), lanes11 and 12) showed signi®cant INS activity and wasable to co-operate with Rev to permit the levels ofunspliced RNA appearing in the cytoplasm to riseto 18.9 %.

In subsequent experiments we were able tofurther reduce the length of the active INS frag-ment. The minimal active fragment we obtainedwas a 30-mer (nt 825 to 854) located at the 50 endof F1 (Figure 2(a)). As shown in Figure 2(c) andTable 1, plasmids carrying the 50 INS sequencewere able to respond to Rev and showed14.3(�6.8)% unspliced cytoplasmic mRNA. How-ever, in the absence of Rev, the levels of unsplicedcytoplasmic mRNA were negligible (1.5(�0.5)%).

It is important to note that the presence of theINS sequence produced only a slight increase inthe levels of unspliced mRNA appearing in thenucleus (MikaeÂlian et al., 1996). This suggests thatthe addition of the INS element to the reporterplasmid does not result in a large accumulation ofunspliced mRNA due to a block to splicing. Asshown in Figure 2(c), the addition of Rev does notincrease the levels of unspliced mRNA in thenucleus more than twofold, demonstrating that thelarge increase in unspliced mRNA in the cytoplasmis due to Rev's ability to stimulate RNA exportfrom the nucleus.

Since Rev does not bind to the p17gag INSelement, its contribution to mRNA export is pre-sumably mediated by cellular factors that regulateeither splicing or RNA transport. Strong candidatesfor this role are found amongst the hnRNP pro-teins which are known to bind to pre-mRNAs andto regulate splicing (Dreyfuss et al., 1993). A num-ber of hnRNP proteins, including hnRNP I, hnRNPA1, hnRNP C and PABP-1, have been shown pre-viously to bind to sequences found in HIV-1 andHTLV-1 (Afonina et al., 1997; Black et al., 1995,1996). However, none of these candidate proteinshave been shown to be responsible for INS func-tion using genetic assays.

In order to identify proteins which could bind tothe 50 INS sequence, we ®rst performed gel mobi-lity shift assays using nuclear and cytoplasmicextracts. To minimise non-speci®c binding, thesereactions were carried out in the presence of excesstRNA and heparin. As shown in Figure 3(a), the 50INS RNA readily formed a complex of slowermobility than the free RNA in the presence ofnuclear extract. Titration experiments showed com-plete RNA-protein binding in the presence ofsaturating concentrations of nuclear extract. In con-trast, no signi®cant RNA-protein complex for-mation was observed when cytoplasmic extracts orpuri®ed SR proteins were tested (data not shown).

To identify the approximate size of the cellularproteins binding to the 50 INS, uridine-50-[a-35S]thiophosphate-labelled 50 INS RNA wasincubated with nuclear extract from HeLa cells.The binding reactions were irradiated with UVlight, treated with RNase T1, and the labelled pro-teins analysed by SDS-PAGE gels. Figure 3(c)shows that two nuclear proteins of apparent mol-ecular masses of 34 kDa and 50 kDa form cross-links with the 50 INS RNA.

Burd & Dreyfuss (1994) have de®ned a high-af®-nity binding site for hnRNP A1 using SELEX, aselection/ampli®cation method starting from poolsof random sequence RNA. UV-induced cross-link-ing experiments showed that two proteins, of mol-

ecular masses of approximately 34 kDa and50 kDa are able to bind to the high-af®nity binding

to be able to bind speci®cally to the 50 INS elementand form a RNP complex together with the 50 kDa

Figure 2. Mutations in the minimal 50 INS that do not bind nuclear proteins are inactive. (a) Wild-type 50 INSsequence and the sequences of two mutations that inhibit HnRNP A1 binding. (b) Gel mobility shift assay. Bindingreactions (10 ml) contained 10,000 cpm of [a-32P]UTP RNA probe and either 0 (ÿ) or 1 ml (�) of HeLa nuclear extract(�30 mg/ml) in binding buffer (10 mM Tris-HCl (pH 7.5), 50 mM KCl, 2 mM DTT, 3 mM MgCl2, 5 % glycerol),50 mg/ml heparin and 40 units of RNasin. After incubation at 4 �C for 15 minutes, the samples were fractionated atroom temperature using non-denaturing 8 % polyacrylamide gels. (c) RNase protection assay. HeLa cells were co-transfected with 500 ng of reporter plasmids in the absence (ÿ) or presence (�) of 50 ng of the Rev-expressing plas-mid pF31. In the plasmids analysed the INS element was replaced by consensus sequences (WT) or mutant variants(M1 and M2) of the binding sites for hnRNP A1 in the 50 INS sequence (pIN30 and pIN31). RNase protection assayswere performed on cytoplasmic (left) and nuclear (right) extracts.


site for hnRNP A1. The 34 kDa protein was ident-i®ed as hnRNP A1, whereas the 50 kDa proteinhas not been cloned or characterised in detail(Burd & Dreyfuss, 1994).

Since the proteins identi®ed by Burd & Dreyfuss(1994) had similar molecular sizes to the proteinsbinding to the INS element, we decided to testwhether the 34 kDa protein could be hnRNP A1.The monoclonal antibody (4B10) raised againsthnRNP A1 was used to detect ``supershift'' com-plexes formed between 32P-labelled 50 INS RNAand nuclear proteins. As shown in Figure 3(b), themAb against hnRNP A1 recognised the complexesformed with the 50 INS RNA. Control experimentsshowed that these complexes were not recognisedby mAb raised against hnRNP C (4F4) or hnRNP L(4D11; data not shown). Thus, hnRNP A1 appears

protein.

Mutations in the 50 INS that fail to bind hnRNPA1 and 50 kDa are inactive

The consensus hnRNP A1 binding site contains aUAGGGA/U motif (Burd & Dreyfuss, 1994;Mayeda et al., 1998). Since this sequence closelyresembles the AUGGGAA sequence present in the50 INS sequence (Figure 2(a)), we decided to seewhether mutations that remove the GGG sequence(mutant 1, GGG! CCC), or permit internal pair-ing with the GGG sequence (mutant 2,AAAA! CCCC) are able inhibit protein bindingand INS activity. As shown in Figure 2(b), RNAfragments containing either of these mutationswere unable to bind hnRNP A1 present in the

nuclear extract. Furthermore, when tested foractivity in the b-globin reporter system, both

the INS element and co-operate with Rev to pro-mote the export of unspliced mRNAs. The control

Figure 3. HnRNP A1 and a 50 kDa protein recognise the 50 INS element. (a) Gel mobility shift assay of nuclear pro-tein binding to the 50 INS. Binding reactions (10 ml) contained 10,000 cpm of [a-32P]UTP RNA probe and between 0and 1 ml of HeLa nuclear extract (�30 mg/ml) in binding buffer (10 mM Tris-HCl (pH 7.5), 50 mM KCl, 2 mM DTT,3 mM MgCl2, 5 % glycerol), 50 mg/ml heparin and 40 units of RNasin. After incubation at 4 �C for 15 minutes, thesamples were fractionated at room temperature using non-denaturing 8 % polyacrylamide gels. (b) Antibody super-shift demonstrating HnRNP A1 binding to the 50 INS. Binding reactions were performed in the absence (ÿ) or pre-sence (�) of 1 ml (1:10) of anti-hnRNP A1 mAb 4B10 (gift from Dr G. Dreyfuss). (c) UV-cross-linking of nuclearproteins to the 50 INS. Binding reactions (15 ml) were similar to those described above and contained 106 cpm of uri-dine-50-[a-35S]thiophosphate-labelled RNA probe with 0.1 to 0.5 ml of HeLa nuclear extract. After the incubation onice, reactions were exposed to UV light at a distance of 4 cm from a 254 nm source at 4 �C for 20 minutes, the RNAdigested with a combination of RNase A and RNase T1, and the labelled proteins fractionated on SDS/12.5 % poly-acrylamide gels containing 7 M urea. Note that two proteins of 50 kDa and 34 kDa become cross-linked to the RNA.


mutations were inactive (Figure 2(c)).

High-affinity hnRNP A1 binding sites haveINS activity

We next tested whether the high-af®nity bindingsite for hnRNP A1 (Figure 4(a)) could substitute for

experiment shown in Figure 4(b) con®rms thathnRNP A1 found in nuclear extracts is able to bindwith high af®nity to the arti®cial binding site, butnot to the mutant binding site. UV-induced cross-linking experiments (Figure 4(c)) using the wild-type RNA also con®rmed the speci®c binding of the34 kDa (hnRNP A1) and 50 kDa proteins to this

sequence (Burd & Dreyfuss, 1994). As expected,control experiments using the 4B10 antibody also

As expected, the mutant hnRNP A1 binding sitewas unable to co-operate with Rev and the levels

Figure 4. The hnRNP A1 high-af®nity site has INS activity. (a) Consensus high-af®nity binding site for hnRNP A1(wild-type) and an inactive variant sequence (mutant). (b) Gel mobility shift assay of nuclear protein binding to thehnRNP A1 binding site. Binding reactions were performed as described in Materials and Methods and the legend toFigure 3, using between 0 and 1 ml of HeLa cell nuclear extract and the consensus hnRNP A1 binding site (wild-typeor mutant). (c) UV-cross-linking of nuclear proteins to the hnRNP A1 binding site. Cross-linking reactions were per-formed as described in Materials and Methods and the legend to Figure 3, using between 0.1 and 0.5 ml of HeLa cellnuclear extract. Note that the two proteins of 50 kDa and 34 kDa that become cross-linked to the hnRNP A1 bindingsite RNA are of identical molecular mass to the proteins that cross-link to the 50 INS RNA. (d) RNase protectionassay. HeLa cells were co-transfected with 500 ng of reporter plasmids in the absence (ÿ) or presence (�) of 50 ng ofthe Rev-expressing plasmid pF31. In the plasmids analysed, the INS element was replaced by wild-type consensussequence (WT) or mutant variant (MUT) of the binding site for hnRNP A1 (pIN86 and pIN87). RNase protectionassays were performed on cytoplasmic (left) and nuclear (right) extracts.


demonstrated that the wild-type hnRNP A1 bind-ing site RNA was also bound by the hnRNP A1 pre-sent in the nuclear extracts (data not shown).

The wild-type and mutant high-af®nity bindingsites for hnRNP A1 were cloned in place of the INSelement in the reporter plasmid and tested for INSactivity in HeLa cells. Figure 4(d) and Table 1show the results of RNase protection experimentsdemonstrating that the INS element can bereplaced by the hnRNP A1 binding site. ThehnRNP A1 element permitted an average of13.6(�4.0)% unspliced b-globin mRNA to be trans-ported to the cytoplasm in the presence of Rev.

of unspliced cytoplasmic RNA appearing in cellsco-transfected with this construct and the Rev-expressing plasmid were considerably reduced(3.1(�1.5)%).

To test whether the ability of the hnRNP A1 pro-tein to co-operate with Rev is speci®c, similarexperiments were performed using the high-af®-nity binding site for hnRNP C (GoÈrlach et al., 1994)and a region of the herpes simplex virus thymidinekinase which is known to bind hnRNP L with highspeci®city (Liu & Mertz, 1995). The antisensesequence to the hnRNP C binding site and asequence containing several nucleotide changes

known to reduce hnRNP L binding were used ascontrols. As shown in Figure 5, we did not observe

thesis of the trans-activator protein, Tat, and a pro-gressive shift towards the production of the 4 kb,

Figure 5. HnRNP C and hnRNP L are unable to pro-mote Rev-dependent export of unspliced mRNA fromthe nucleus. HeLa cells were co-transfected with 500 ngof reporter plasmids in the absence (ÿ) or presence (�)of 50 ng of the Rev-expressing plasmid pF31. The plas-mids analysed carried the wild-type p17gag INS element(INS; pIM41, WT) or the inactive mutant INS element(pIM93; MUT) and a series of reporters in which theINS element was replaced by the wild-type consensussequence (WT) or mutant variant (MUT) of the bindingsite for hnRNP A1 (pIN86 and pIN87), hnRNP C (pIN38and pIN39) and hnRNP L (pIN48 and pIN49). Thesequences for these elements are given in Table 2. Cyto-plasmic RNA was isolated 48 hours after transfectionand assayed by RNase protection experiments using theL12 probe. The protected fragments were fractionatedon a 6 % polyacrylamide gel containing 7 M urea. Thepositions of the protected fragments corresponding tothe unspliced and spliced b-globin mRNAs are indicatedon the left of the gel.


a signi®cant Rev response when the INS elementwas substituted for the hnRNP C (5.3 % unsplicedmRNA, Figure 5, lane 10) or hnRNP L bindingsites (1.7 % unspliced mRNA, Figure 5, lane 14).

Discussion

Control of HIV mRNA processing and transportby cis-acting RNA elements

Immediately after infection, HIV mRNA syn-thesis is restricted to short, multiply spliced 1.8 kbmRNAs encoding the regulatory proteins. As theinfection proceeds, there is a dramatic increase inthe overall level of transcription due to the syn-

partially spliced mRNAs and 9 kb virion RNA(Kim et al., 1989). The major mechanism used byHIV to regulate its complex pattern of geneexpression is the control of the nuclear export ofthe subgenomic mRNAs by the viral regulatoryprotein Rev. Rev binds directly to its RNA recog-nition element, the Rev response element (Bartelet al., 1991; Daly et al., 1989; Heaphy et al., 1990,1991; Kjems et al., 1991; Malim & Cullen, 1991;Mann et al., 1994; Zapp & Green, 1989; Zemmelet al., 1996), which is located in the env gene andpresent on all the transcripts encoding HIV's struc-tural and accessory genes (Malim et al., 1989; Mannet al., 1994; Rosen et al., 1988). The nuclear exportsignal of Rev remains exposed on the surface ofthe RNP complex and this permits transport ofmRNAs bound by Rev from the nucleus (Fischeret al., 1995; Henderson & Percipalle, 1997; Malimet al., 1991; Meyer & Malim, 1994; Stauber et al.,1995; Wolff et al., 1997).

The export pathway used by Rev is related tothe transport pathway for snRNA (small nuclearRNA) and tRNAs (Fischer et al., 1995). A numberof proteins that are able to distinguish betweenwild-type and mutant nuclear export signals havebeen identi®ed using the yeast two-hybrid selec-tion assay (Bogerd et al., 1995; Fritz et al., 1995;Stutz et al., 1995). This group of proteins includesthe Rab (Rip) protein which is related to thenucleophilins (Bogerd et al., 1995; Fritz et al., 1995;Stutz et al., 1995). However, puri®ed Rab (Rip) isunable to bind Rev (Henderson & Percipalle, 1997).The strongest candidate to be the cellular transportfactor that recognises the Rev NES is the CRM1(chromosomal region 1 maintenance) protein.CRM1 shows homology to importin b-like trans-port factors and has recently been reported to bindthe Rev NES motif together with Ran GTPase(Fornerod et al., 1997; Neville et al., 1997; Ossareh-Nazari et al., 1997; Stade et al., 1997; Zolotukhin &Felber, 1997).

Although the Rev/RRE system is necessary forthe export of unspliced mRNAs, careful regulationof splicing rates is also needed in order for HIV toexpress simultaneously all its mRNAs. HIVachieves this through the use of inef®cient 30 spliceacceptor sites function (Amendt et al., 1994;Barksdale & Baker, 1995; Staffa & Cochrane, 1995)and through the activity of cis-acting repressivesequences or instability elements which are locatedthroughout the genome (Maldarelli et al., 1991;MikaeÂlian et al., 1996; Nasioulas et al., 1994; Rosenet al., 1988; Schwartz et al., 1992).

Previous studies from our laboratory haveshown that the instability sequence within theHIV-1 p17gag gene can act as a genetic signal thatpermits ef®cient export of unspliced mRNAs byRev (MikaeÂlian et al., 1996). The experimentsdescribed here demonstrate that the splicing factorhnRNP A1 can co-operate with Rev and permit the

ef®cient export of unspliced mRNAs from the nor-mally ef®ciently spliced b-globin mRNA.

Thus, the b-globin gene is spliced and transportedto the cytoplasm via the normal mechanism.


HnRNP A1 can co-operate synergistically withRev to promote export of unspliced mRNA

There has been a considerable effort to identifythe cellular proteins that can recognise the HIVINS element and co-operate with Rev to promoteRNA transport. The best candidates for cellularproteins that can recognise the HIV INS elementand co-operate with Rev to promote RNA trans-port are found amongst the ubiquitous nuclearRNA binding proteins, such as the hnRNP proteins(Dreyfuss et al., 1993) and the SR proteins (Zahleret al., 1993). For example, Olsen et al. (1992)described the interaction of hnRNP C, a nuclearprotein which is involved in splicing and containsa nuclear retention signal (Nakielny & Dreyfuss,1996), and a 270 bp segment of the pol gene knownto repress gene expression. HnRNP A1 and hnRNPI have also been reported to bind to the INSelements within pol and env of HIV-1 as well as toan INS element within HTLV-II (Black et al., 1995,1996). Finally, the INS element from the p17gagcoding region has been recently shown to preferen-tially bind to poly(A)-binding protein 1 (PABP1),which is believed to increase mRNA stabilitythrough the inhibition of mRNA decapping by thepoly(A) tail or by in¯uencing the rate of polyade-nylation (Afonina et al., 1997). However, unam-biguous genetic data linking the binding of theseproteins to HIV mRNA expression has not beenreported.

Here, we have de®ned a 30 nt region within thep17gag gene, the 50 INS element, which is the mini-mal element that is able to co-operate with Rev topromote the nuclear export of unspliced b-globinmRNAs. Addition of the 50 INS element to reportergenes carrying the RRE permits up to 14 % of theb-globin mRNA transported from the nucleus inthe presence of Rev to be unspliced.

Several lines of evidence demonstrate thathnRNP A1 is responsible for INS activity. First, gelmobility shift and UV-cross-linking experimentshave shown that hnRNP A1 and a 50 kDa nuclearprotein are able to bind to the 50 INS with highaf®nity. Mutations in the INS element that preventhnRNP A1 binding are also inactive in promotingRev-dependent nuclear export. Finally, geneticexperiments have con®rmed that the consensushigh-af®nity binding site for hnRNP A1 can func-tion as an INS element and can cooperate synergis-tically with Rev to stimulate nuclear export. Incontrast, hnRNP C and hnRNP L are unable tostimulate Rev-mediated transport of unsplicedmRNAs.

A working hypothesis for the ability of hnRNPA1 to synergistically activate Rev-dependentexport from the nucleus is shown in Figure 6. Inthe reporter plasmids carrying only the RREelement, there is no signi®cant increase in theretention of unspliced mRNAs in the nucleus.

Addition of high-af®nity sites for hnRNP A1 in cisdecreases the rate of splicing of the b-globinmRNA. The partially spliced mRNAs can then beseparated from the splicing machinery, recognisedby Rev, and transported to the cytoplasm.

It is important to note that although hnRNP A1appears to reduce the rate of splicing, it does notimpose a strong block to splicing that results in theaccumulation of high levels of unspliced mRNA inthe nucleus and reduced levels of spliced mRNAin the cytoplasm. In contrast, when mutations thatprevent recognition of the intron are introducedinto the 50 splice donor or the 30 splice acceptorsequences of the b-globin gene, a high level ofunspliced mRNA accumulates in the nucleus. Thispool of unspliced mRNA can be ef®ciently trans-ported to the cytoplasm in the absence of Rev(MikaeÂlian et al., 1996).

The ability of hnRNP A1 to regulate splice siteselection and to shuttle between the nucleus andthe cytoplasm is consistent with its synergisticeffect on Rev-mediated pre-mRNA export from thenucleus. HnRNP A1 is known to play an importantrole in pre-mRNA processing, through regulationof 50 splice site selection (CaÂceres et al., 1994;Mayeda & Krainer, 1992). At high concentrationsof hnRNP A1 the use of distal 50 splice sites isfavoured (Burd & Dreyfuss, 1994; CaÂceres et al.,1994; Mayeda & Krainer, 1992). HnRNP A1 is alsoable to shuttle between the nucleus and the cyto-plasm in association with RNA (Izaurralde et al.,1997; Michael et al., 1995). The hnRNP A1 transportsignal has been mapped to 38 amino acid residuesin the carboxyl terminus of the protein and acts asboth a nuclear localisation signal and a nuclearexport signal. Once in the cytoplasm, the HnRNPA1 protein, like Rev, is removed from the RNAand re-imported to the nucleus. Thus, hnRNP A1could complement Rev activity both by affectingsplicing rates and by participating directly in theexport of the unspliced mRNA.

We have observed that in addition to hnRNPA1, a 50 kDa nuclear protein is able to bind to the50 INS element as well as to the high-af®nity sitefor hnRNP A1. The 50 kDa protein appears tohave a binding speci®city that closely mimics thatof hnRNP A1, and we have been unable to dis-cover sequences that can discriminate between thetwo proteins. Burd & Dreyfuss (1994) have alsoobserved that a 50 kDa protein found in nuclearextracts binds speci®cally to the high-af®nity A1binding site. They suggested that this protein alsoacts as a splicing factor, since the addition ofhnRNP A1 binding sites to an adenovirus-derivedmajor late promoter model pre-mRNA inhibitssplicing. Control experiments have shown that the50 kDa protein is not a member of the SR proteinfamily (data not shown), but we have not charac-terised this protein further.

In conclusion, the major control elements regu-lating the post-transcriptional processing of the

HIV mRNAs are the cis-acting RNA signals, theINS and RRE elements. The experiments described

likely that it is used to reduce splicing rates suf®-ciently for the unspliced and partially spliced viral

Figure 6. Model for role of hnRNP A1 in the Rev response. Under normal conditions, the b-globin pre-mRNAenters the splicing compartment, is spliced and exported by the normal Rev-independent export pathway. Rev isunable to increase the export of unspliced mRNA under these circumstances, either because splicing of b-globin ishighly ef®cient, or because the b-globin pre-mRNA is sequestered by the splicing machinery. In contrast, whenhnRNP A1 is introduced in cis (by addition of speci®c high-af®nity binding sites) partially spliced b-globin mRNAbecomes accessible to Rev. As a result, Rev is able to stimulate the nuclear export of unspliced pre-mRNAs carryingRREs.


here demonstrate that the INS element in thep17gag gene is able to recruit hnRNP A1. Althoughthe precise mechanism by which hnRNP A1 is ableto participate in the Rev-mediated export ofmRNA from the nucleus is not yet known, it seems

mRNAs to accumulate in the nucleus. RNP com-plexes carrying the viral mRNA, hnRNP A1 andRev are then ef®ciently exported to the cytoplasmthrough the synergistic activities of these twoimportant export proteins.

Materials and Methods

Preparation of RNA transcripts

resuspended in HBZ buffer (10 mM Hepes (pH 7.9),0.75 mM spermidine, 0.15 mM spermine, 10 mM ZnSO4,

Table 2. Sequences of the 50 INS element and hnRNP binding sites

Plasmid Elementa Sequence (50!30)b

pIN27 50 INS gaattcTTAGATCGATGGGAAAAAATTCGGTTAAGgctagc

pIN30 50 INS* gaattcTTAGATCGATCCCAAAAAATTCGGTTAAGgctagc

pIN31 50 INS* gaattcTTAGATCGATGGGACCCCAATTCGGTTAAGgctagc

pIN86 hnRNP A1 gaattcTATGATAGGGACTTAGGGTGgctagc

pIN87 hnRNP A1* gaattcTATGATACCCACTTACCCTGgctagc

pIN38 hnRNP C1 gaattcTATGATTTTTTTTTTgctagc

pIN39 hnRNP C1* gaattcTATGAAAAAAAAAAAgctagc

pIN48 hnRNP gaattcTATGATCGCGAACATCTACACCACACAACACCGCCTCGAgctagc

pIN49 hnRNP L* gaattcTATGATCGCGAACTAGTAGATCTAGAAACACCGCCTCGAgctagc

a Asterisk indicates mutant sequence.b Altered residues found in the mutant sequences are underlined. The sequences of the ¯anking EcoRI (50) and NheI (30) sites are

shown in lower case.


The DNA templates used in this study were based ona pUC19 derivative carrying the T3 promoter. An EcoRIsite was inserted at position �3 for subsequent cloningof DNA oligonucleotides carrying INS sequences andhnRNP protein binding sites (Table 2).

RNA synthesis reactions contained between 0.5 and1 mg of template DNA in 40 mM Tris-HCl (pH 8.0),6 mM MgCl2, 10 mM DTT, 2 mM spermidine, 1 mMGTP, ATP and CTP, 50 mCi of [a-32P]UTP, 40 units ofRNasin (Promega) and 40 units of T3 RNA polymerase(Boehringer Mannheim). After incubation for two hoursat 37 �C, the reactions were stopped by extraction withphenol, and the RNA transcripts were fractionated on6 % (w/v) polyacrylamide gels containing TBE buffer(45 mM Tris base, 45 mM boric acid, 10 mM EDTA,pH 8.3) and 7 M urea. The labelled RNA was eluted at37 �C for 12 to 16 hours in 0.3 M sodium acetate, 10 mMTris-HCl (pH 7.4), 1 mM EDTA, 0.1 % (w/v) SDS andprecipitated with ethanol in the presence of 10 mg tRNA.

Constructs for transient transfection assays

The b-globin reporter plasmids were based on pIM41and its precursor pIM124 (MikaeÂlian et al., 1996). Theseplasmids contain the PvuII-EcoRI fragment of the rabbitb-globin gene under the CMV promoter/enhancersequences and ¯anked on its 30 end by a polylinkersequence and the SV40 polyadenylation signal. Thesecond exon of the b-globin gene also contains fouradditional base-pairs at the ``®lled-in'' BamHI site(MikaeÂlian et al., 1996). The sequences studied wereinserted between the EcoRI and NheI sites of the polylin-ker. The wild-type p17gag INS element and the INS frag-ments F1, F2, F3 and F4 were ampli®ed by PCR usingHIV-1NL4-3 as a template. The mutant control was ampli-®ed with the same oligonucleotides but using a mutanttemplate (p17M1234; kindly provided by Dr G. Pavlakis;Schwartz et al., 1992). The 50 INS sequence and bindingsites for the hnRNP proteins were introduced by cloningof the synthetic sequences (Table 2).

Preparation of HeLa nuclear andcytoplasmic extracts

HeLa cells were grown to a density of 3.8 � 105/ml inDMEM (Dulbecco) supplemented with antibiotics and10 % (v/v) foetal calf serum (Gibco). Cells were washedtwice in PBS, pelleted by centrifugation at 2000 rpm and

4 mM MgCl2, 2 mM DTT) containing 10 mM KCl. Afterswelling on ice for ®ve minutes, the cells were disruptedusing a Dounce homogeniser and the nuclei recoveredafter centrifugation through a 0.14 M sucrose cushion inHBZ � 10 mM KCl buffer at 5000 rpm for 30 seconds.The cytoplasmic fraction was saved and the nuclei wereextracted three times using HBZ � 360 mM KCl and pre-cipitated with ammonium sulphate as described (Rittneret al., 1995). The nuclear proteins were resuspended inHBZ � 100 mM KCl and the nuclear and cytoplasmicextracts were dialysed against HBZ � 100 mMKCl � 10 % (v/v) glycerol. The extracts were cleared bycentrifugation at 20,000 rpm for ten minutes and storedin aliquots in liquid nitrogen.

Gel mobility shift assays

Binding reactions (10 ml) contained 10,000 cpm of[a-32P]UTP RNA probe and variable amounts of HeLanuclear extract in binding buffer (10 mM Tris-HCl(pH 7.5), 50 mM KCl, 2 mM DTT, 3 mM MgCl2, 5 %(v/v) glycerol), 50 mg/ml heparin and 40 units ofRNasin. The RNA transcripts were refolded by heatingat 90 �C for three minutes and quick chilling on iceimmediately before use. After incubation at 4 �C for 15minutes 3 ml of gel loading buffer (50 % (v/v) glycerol,0.1 % (w/v) bromophenol blue, 0.1 % (v/v) xylene cya-nol FF) was added and the samples were fractionated atroom temperature using non-denaturing 8 % polyacryl-amide gels. Gels were dried, exposed to X-ray ®lm atÿ70 �C.

For the antibody supershift assays, RNA electrophor-etic mobility shift assays were carried out as describedabove in the absence or presence of 1 ml of mAb diluted1:10. The antibodies used were directed against hnRNPA1 (4B10), hnRNP C1/C2 (4F4) and hnRNP L (4D11)and were kindly provided by Dr G. Dreyfuss.

Ultraviolet cross-linking

Binding reactions (15 ml) were similar to thosedescribed above and contained 106 cpm of uridine-50-[a-35S]thiophosphate-labelled RNA probe with variableamounts of HeLa nuclear extract. After the incubation onice, reactions were exposed to UV light at a distance of4 cm from a 254 nm source at 4 �C for 20 minutes. TheRNA was digested by incubation with a mixture of 2 mgRNase A and 4 mg RNase T1 for 60 minutes at 27 �C andthe samples were run on SDS/12.5 % polyacrylamide

gels containing 7 M urea. After drying, the gels wereexposed to X-ray ®lm at ÿ70 �C.

the ®rst tat coding exon of human immunode®-ciency virus type 1. Mol. Cell Biol. 14, 3960-3970.


RNase protection assays

HeLa cells (ATCC CCL2) were plated at a concen-tration of 4.5 � 105 per 100 mm Petri dish the day beforetransfection by the calcium phosphate method (Gormanet al., 1983). Each dish received 400 ml precipitate con-taining 500 ng of CsCl puri®ed reporter plasmid in thepresence or absence of 50 ng of a Rev-expressing plas-mid, pF31, which carries a synthetic rev gene inserteddownstream of the CMV immediate early region promo-ter (Mann et al., 1994).

Cells were lysed 48 hours after transfection in 10 mMNaCl, 2 mM MgCl2, 10 mM Tris-HCl (pH 7.8), 5 mMDTT, 0.5 % (v/v) NP40 and the cytoplasmic extract wasseparated from the nuclei by low-speed centrifugation.The cytoplasmic extract was treated with proteinase K(50 mg/ml) in the presence of 0.2 % (w/v) SDS at 37 �Cfor 30 minutes. After extraction with phenol:chlorofor-m:isoamyl alcohol, the cytoplasmic RNA was recoveredby precipitation in ethanol. The RNA preparations werepuri®ed from contaminating DNA by digestion with 10units of RNase-free DNase I (HT Biotechnology) in thepresence of 40 units of RNasin (Promega).

pL12 DNA was linearized with EcoRI, to generate a253 nt RNA probe which overlaps the boundary of thesecond exon and intron of the b-globin and it is abledetect both unspliced and spliced mRNA in the RNaseT1 protection assay. T3 transcription reactions (20 ml)contained 400 mM Tris-HCl (pH 8.0), 60 mM MgCl2,100 mM DTT and 20 mM spermidine in the presence of1 mM GTP, ATP and CTP, 50 mCi of [a-32P]UTP, 40 unitsof RNasin (Promega) and 40 units of T3 RNA polymer-ase (Boehringer Mannheim). The labelled RNA was puri-®ed from a 6 % (w/v) polyacrylamide 7 M urea gel, asdescribed above.

Cytoplasmic RNA was hybridised to 150,000 to200,000 cpm of 32P-labelled L12 probe for 12 hours at37 �C in 20 ml hybridisation buffer (80 % formamide,400 mM NaCl, 40 mM Pipes pH 6.4, 1 mM EDTA). Reac-tions were diluted with 300 ml of RNase T1 buffer(300 mM NaCl, 10 mM Tris-HCl pH 7.4, 5 mM EDTA)and the excess probe and the unhybridised sequenceswere digested with 15 mg/ml of RNase T1 for 60 minutesat 30 �C. After extraction with phenol:chloroform:isoamylalcohol, the protected RNA fragments were analysed on6 % polyacrylamide gels containing 7 M urea. The gelswere dried, exposed to X-ray ®lm, and the ratio ofunspliced to spliced RNA measured by densitometry.

References

Adachi, A., Gendelman, H. E., Koenig, S., Folks, T.,Willey, R., Rabson, A. & Martin, M. A. (1986). Pro-duction of acquired immunode®ciency syndrome-associated retrovirus in human and non-humancells transfected with an infectious molecular clone.J. Virol. 59, 284-291.

Afonina, E., Neumann, M. & Pavlakis, G. N. (1997). Pre-ferential binding of poly(A)-binding protein 1 to aninhibitory RNA element in the human immunode®-ciency virus type 1 gag mRNA. J. Biol. Chem. 272,2307-2311.

Amendt, B. A., Hesslein, D., Chang, L.-J. & Stoltzfus,C. M. (1994). Presence of negative and positive cis-acting RNA splicing elements within and ¯anking

Arrigo, S. J., Wietsman, S., Zack, J. A. & Chen, I. S. Y.(1990). Characterization and expression of novel sin-gly spliced RNA species of human immunode®-ciency virus type 1. J. Virol. 64, 4585-4588.

Barksdale, S. K. & Baker, C. C. (1995). The humanimmunode®ciency virus type 1 Rev protein and theRev-response element counteract the effect of aninhibitory 50 splice site in a 30 untranslated region.Mol. Cell Biol. 15, 2962-2971.

Bartel, D. P., Zapp, M. L., Green, M. R. & Szostak, J. W.(1991). HIV-1 Rev regulation involves recognition ofnon-Watson-Crick base pairs in viral RNA. Cell, 67,529-536.

Berthold, E. & Maldarelli, F. (1996). Cis-acting elementsin human immunode®ciency virus type 1 RNAsdirect viral transcripts to distinct intranuclearlocations. J. Virol. 70, 4667-4682.

Black, A. C., Luo, J., Watanabe, C., Chun, S., Bakker, A.,Fraser, J. K., Morgan, J. P. & Rosenblatt, J. D.(1995). Polypyrimidine tract-binding protein andheterogeneous nuclear ribonucleoprotein A1 bind tohuman T-cell leukemia virus type 2 RNA regulatoryelements. J. Virol. 69, 6852-6858.

Black, A. C., Luo, J., Chun, S., Bakker, A., Fraser, J. K. &Rosenblatt, J. D. (1996). Speci®c binding of polypyr-imidine tract binding protein and hnRNP A1 toHIV-1 CRS elements. Virus Genes, 12, 275-285.

Bogerd, H. P., Fridell, R. A., Madore, S. & Cullen, B. R.(1995). Identi®cation of a novel cellular co-factor forthe Rev/Rex class of retroviral regulatory proteins.Cell, 82, 485-494.

BoÈhnlein, E., Berger, J. & Hauber, J. (1991). Functionalmapping of the human immunode®ciency virustype 1 Rev RNA binding domain: new insights intothe domain structure of Rev and Rex. J. Virol. 65,7051-7055.

Bray, M., Prasad, S., Dubay, J. W., Hunter, E., Jeang,K.-T., Rekosh, D. & HammarskjoÈ ld, M.-L. (1994).A small element from the Mason-P®zer monkeyvirus genome makes human immunode®ciencyvirus type 1 expression and replication Rev-inde-pendent. Proc. Natl Acad. Sci. USA, 91, 1256-1260.

Burd, C. G. & Dreyfuss, G. (1994). RNA binding speci-®city of hnRNP A1: signi®cance of hnRNP A1 high-af®nity sites in pre-mRNA splicing. EMBO J. 13,1197-1204.

CaÂceres, J. F., Stamm, S., Helfman, D. M. & Krainer,A. R. (1994). Regulation of alternative splicingin vivo by overexpression of antagonistic splicingfactors. Science, 265, 1706-1709.

Chang, D. A. & Sharp, P. A. (1989). Regulation by HIVRev depends upon recognition of splice site. Cell,59, 789-795.

Charpentier, B., Stutz, F. & Rosbash, M. (1997).A dynamic in vivo view of the HIV-1 Rev-RREinteraction. J. Mol. Biol. 266, 950-962.

Cook, C. R. & McNally, M. T. (1998). SR protein andsnRNP requirements for assembly of the Rous Sar-coma Virus negative regulator of splicing complexin vitro. Virology, 242, 211-220.

Daly, T. J., Cook, K. S., Gary, G. S., Maione, T. E. &Rusche, J. R. (1989). Speci®c binding of HIV-1recombinant Rev protein to the Rev-responsiveelement in vitro. Nature, 342, 816-819.

Dreyfuss, G., Matunis, M. J., PinÄ ol-Roma, S. & Burd,C. G. (1993). hnRNP proteins and the biogenesis ofmRNA. Annu. Rev. Biochem. 62, 289-321.

Dyhr-Mikkelsen, H. & Kjems, J. (1995). Inef®cient spli-ceosome assembly and abnormal branch site selec-

Liu, X. & Mertz, J. E. (1995). HnRNP L binds a cis-actingRNA sequence element that enables intron-indepen-


tion in splicing of an HIV-1 transcript in vitro. J. Biol.Chem. 270, 24060-24066.

Ernst, R. K., Bray, M., Rekosh, D. & HammarskjoÈ ld,M.-L. (1997). A structured retroviral RNA regulat-ory element that mediates nucleocytoplasmic exportof intron-containing RNA. Mol. Cell Biol. 17, 135-144.

Feinberg, M. B., Jarrett, R. F., Aldovini, A., Gallo, R. C.& Wong-Staal, F. (1986). HTLV-III expression andproduction involve complex regulation at the levelsof splicing and translation of viral RNA. Cell, 46,807-817.

Fischer, U., Huber, J., Boelens, W. C., Mattaj, I. W. &LuÈ hrmann, R. (1995). The HIV-1 Rev activationdomain is a nuclear export signal that accesses anexport pathway used by speci®c cellular RNAs.Cell, 82, 475-483.

Fornerod, M., Ohno, M., Yoshida, M. & Mattaj, I. W.(1997). CRM1 is an export receptor for leucine-richnuclear export signals. Cell, 90, 1051-1060.

Fritz, C. C., Zapp, M. L. & Green, M. R. (1995). Ahuman nucleoporin-like protein that speci®callyinteracts with HIV Rev. Nature, 376, 530-533.

GoÈrlach, M., Burd, C. G. & Dreyfuss, G. (1994). Thedeterminants of RNA-binding speci®city of the het-erogeneous nuclear ribonucleoprotein C proteins.J. Biol. Chem. 269, 23074-23078.

Gorman, C. M., Padmanabliam, R. & Howard, B. H.(1983). High ef®ciency DNA-mediated transform-ation of primate cells. Science, 221, 551-553.

Heaphy, S., Dingwall, C., Ernberg, I., Gait, M. J., Green,S. M., Karn, J., Lowe, A. D., Singh, M. & Skinner,M. A. (1990). HIV-1 regulator of virion expression(Rev) protein binds to an RNA stem-loop structurelocated within the Rev-response element region.Cell, 60, 685-693.

Heaphy, S., Finch, J. T., Gait, M. J., Karn, J. & Singh, M.(1991). Human immunode®ciency virus type1 regu-lator of virion expression, Rev, forms nucleoprotein®laments after binding to a purine-rich ``bubble''located within the Rev-responsive region of viralRNA. Proc. Natl Acad. Sci. USA, 88, 7366-7370.

Henderson, B. R. & Percipalle, P. (1997). Interactionsbetween HIV Rev and nuclear import and exportfactors: the Rev nuclear localisation signal mediatesspeci®c binding to human importin-b. J. Mol. Biol.274, 693-707.

Izaurralde, E., Jarmolowski, A., Beisel, C., Mattaj, I. W.,Dreyfuss, G. & Fischer, U. (1997). A role for the M9transport signal of hnRNP A1 in mRNA nuclearexport. J. Cell Biol. 137, 27-35.

Katz, R. A., Kotler, M. & Skalka, A. M. (1988). Cis-actingintron mutations that affect the ef®ciency of avianretroviral RNA splicing: implication for mechanismsof control. J. Virol. 62, 2686-2695.

Kim, S., Byrn, R., Groopman, J. & Baltimore, D. (1989).Temporal aspects of DNA and RNA synthesisduring human immunode®ciency virus infection:evidence for differential gene expression. J. Virol.63, 3708-3713.

Kjems, J., Brown, M., Chang, D. D. & Sharp, P. A.(1991). Structural analysis of the interaction betweenthe human immunode®ciency virus Rev proteinand the Rev-response element. Proc. Natl Acad. Sci.USA, 88(2), 683-687.

dent gene expression. Genes Dev. 9, 1766-1780.Maldarelli, F., Martin, M. A. & Strebel, K. (1991). Identi-

®cation of post-transcriptionally active inhibitorsequences in human immunode®ciency virus type 1RNA: novel level of gene regulation. J. Virol. 65,5732-5743.

Malim, M. H. & Cullen, B. R. (1991). HIV-1 structuralgene expression requires the binding of multipleRev monomers to the viral RRE: implications forHIV-1 latency. Cell, 65, 241-248.

Malim, M. H., Hauber, J., Le, S.-Y., Maizel, J. V. &Cullen, B. R. (1989). The HIV-1 Rev trans-activatoracts through a structured target sequence to activatenuclear export of unspliced viral mRNA. Nature,338, 254-257.

Malim, M. H., McCarn, D. F., Tiley, L. S. & Cullen, B. R.(1991). Mutational de®nition of the human immu-node®ciency virus type 1 Rev activation domain.J. Virol. 65, 4248-4254.

Mann, D. A., MikaeÂlian, I., Zemmel, R. W., Green, S. M.,Lowe, A. D., Kimura, T., Singh, M., Butler, P. J. G.,Gait, M. J. & Karn, J. (1994). A molecular rheostat:co-operative Rev binding to Stem I of the Rev-response element modulates human immunode®-ciency virus type-1 late gene expression. J. Mol. Biol.241, 193-207.

Mayeda, A. & Krainer, A. R. (1992). Regulation ofalternative pre-mRNA splicing by hnRNP A1 andsplicing factor SF2. Cell, 68, 365-375.

Mayeda, A., Munroe, S. H., Xu, R.-M. & Krainer, A. R.(1998). Distinct functions of the closely related tan-dem RNA-recognition motifs of hnRNP A1. RNA,4, 1111-1123.

McNally, M. T. & Beemon, K. (1992). Intronic sequencesand 30 splice sites control Rous Sarcoma Virus RNAsplicing. J. Virol. 66, 6-11.

Meyer, B. E. & Malim, M. H. (1994). The HIV-1 Revtrans-activator shuttles between the nucleus and thecytoplasm. Genes Dev. 8, 1538-1547.

Michael, W. M., Choi, M. & Dreyfuss, G. (1995).A nuclear export signal in hnRNP A1: a signalmediated, temperature-dependent nuclear proteinexport pathway. Cell, 83, 415-422.

MikaeÂlian, I., Krieg, M., Gait, M. J. & Karn, J. (1996).Interactions of INS (CRS) elements and the splicingmachinery regulate the production of Rev-respon-sive mRNAs. J. Mol. Biol. 257, 246-264.

Nakielny, S. & Dreyfuss, G. (1996). The hnRNP C pro-teins contain a nuclear retention sequence that canoverride nuclear export signals. J. Cell Biol. 134,1365-1373.

Nasioulas, G., Zolotukhin, A. S., Tabernero, C., Solomin,L., Cunningham, C. P., Pavlakis, G. N. & Felber,B. K. (1994). Elements distinct from human immu-node®ciency virus type 1 splice sites are responsiblefor the Rev dependence of env mRNA. J. Virol. 68,2986-2993.

Neville, M., Stutz, F., Lee, L., Davis, L. I. & Rosbach, M.(1997). The importin-beta family member Crm1pbridges the interaction between Rev and the nuclearpore complex during nuclear export. Curr. Biol. 7,767-775.

Olsen, H., Cochrane, A. & Rosen, C. (1992). Interactionof cellular factors with intragenic cis-acting repres-sive sequences within the HIV genome. Virology,191, 709-715.

O'Reilly, M. M., McNally, M. T. & Beemon, K. L. (1995).Two strong 50 splice sites and competing, subopti-

0

Stade, K., Ford, C. S., Guthrie, C. & Wies, K. (1997).Exportin 1 (Crm1p) is an essential nuclear export


mal 3 splice sites involved in alternative splicing ofhuman immunode®ciency virus type 1 RNA. Virol-ogy, 213, 373-385.

Ossareh-Nazari, B., Bachelerie, F. & Dargemont, C.(1997). Evidence for a role of CRM1 in signal-mediated nuclear protein export. Science, 278, 141-144.

PinÄol-Roma, S. & Dreyfuss, G. (1992). Shuttling of pre-mRNA binding proteins between nucleus and cyto-plasm. Nature, 355, 730-732.

Purcell, D. F. J. & Martin, M. A. (1993). Alternative spli-cing of human immunode®ciency virus type 1mRNA modulates viral protein expression, replica-tion and infectivity. J. Virol. 67, 6365-6378.

Rittner, K., Churcher, M. J., Gait, M. J. & Karn, J. (1995).The human immunode®ciency virus long terminalrepeat includes a specialised initiator element whichis required for Tat-responsive transcription. J. Mol.Biol. 248, 562-580.

Rosen, C. A., Terwilliger, E., Dayton, A. I., Sodrowski,J. G. & Haseltine, W. A. (1988). Intragenic cis-actingart-responsive sequences of the human immunode®-ciency virus. Proc. Natl Acad. Sci. USA, 85, 2071-2075.

Schwartz, S., Felber, B. K., Benko, D. M., Fenyo, E.-M. &Pavlakis, G. N. (1990). Cloning and functional anal-ysis of multiply spliced mRNA species of humanimmunode®ciency virus type 1. J. Virol. 64, 2519-2529.

Schwartz, S., Campbell, M., Nasioulas, G., Harrison, J.,Felber, B. K. & Pavlakis, G. N. (1992). Mutationalinactivation of an inhibitory sequence in humanimmunode®ciency virus type 1 results in Rev-inde-pendent gag expression. J. Virol. 66(12), 7176-7182.

Sodroski, J., Goh, W. C., Rosen, C. A., Dayton, A.,Terwilliger, E. & Haseltine, W. A. (1986). A secondpost-transcriptional activator gene required forHTLV-III replication. Nature, 321, 412-417.

factor. Cell, 90, 1041-1050.Staffa, A. & Cochrane, A. (1995). Identi®cation of posi-

tive and negative splicing regulatory elementswithin the terminal tat-rev exon of human immuno-de®ciency virus type 1. Mol. Cell Biol. 15, 4597-4605.

Stauber, R., Gaitanaris, G. A. & Pavlakis, G. N. (1995).Analysis of traf®cking of Rev and transdominantRev proteins in living cells using green ¯uorescentprotein fusions: transdominant Rev blocks theexport of Rev from the nucleus to the cytoplasm.Virology, 213, 439-449.

Stutz, F., Neville, M. & Rosbach, M. (1995). Identi®-cation of a novel nuclear pore-associated protein asa functional target of the HIV-1 Rev protein inyeast. Cell, 82, 495-506.

Wolff, B., Sanglier, J.-J. & Wang, Y. (1997). LeptomycinB is an inhibitor of nuclear export: inhibition ofnucleo-cytoplasmic translocation of the humanimmunode®ciency virus type 1 (HIV-1) Rev proteinand Rev-dependent mRNA. Chem. Biol. 4, 139-147.

Zahler, A. M., Neugelauer, K. M., Lane, W. S. & Roth,M. B. (1993). Distinct functions of SR proteins inalternative pre-mRNA splicing. Science, 260, 219-221.

Zapp, M. L. & Green, M. R. (1989). Sequence-speci®cbinding by the HIV-1 Rev protein. Nature, 342, 714-716.

Zemmel, R. W., Kelley, A. C., Karn, J. & Butler, P. J. G.(1996). Flexible regions of RNA structure facilitateco-operative Rev assembly on the Rev-responseelement. J. Mol. Biol. 258, 763-777.

Zolotukhin, A. S. & Felber, B. K. (1997). Mutations inthe nuclear export signal of human Ran-bindingprotein RanBP1 block the Rev-mediated post-tran-scriptional regulation of human immunode®ciencyvirus type 1. J. Biol. Chem. 272, 11356-11360.

Edited by A. R. Fersht

(Received 30 September 1998; received in revised form 4 December 1998; accepted 9 December 1998)

Synergistic stimulation of HIV-1 rev-dependent export of unspliced mRNA to the cytoplasm by hnRNP A1

Documents

Transcript of Synergistic stimulation of HIV-1 rev-dependent export of unspliced mRNA to the cytoplasm by hnRNP A1