Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Analysis of arginine-rich peptides from the HIV Tat protein reveals unusual features of RNA-protein recognition

Barbara J. Calnan/ Sara Biancalana/ Derek Hudson,^ and Alan D. Frankel^'^ 'whitehead Institute for Biomedical Research, Cambridge, Massachusetts 02142 USA; ^MilliGen/Biosearch, Novate, California 94949 USA

Arginine-rich sequences are found in many RNA-binding proteins and have been proposed to mediate specific RNA recognition. Fragments of the HIV-1 Tat protein that contain the arginine-rich region of Tat bind specifically to a 3-nucleotide bulge in TAR RNA. To determine the amino acid requirements for specific RNA recognition, we synthesized a series of mutant Tat peptides spanning this domain (YGRKKRRQRRRP) and measured their affinity and specificity for TAR RNA. Several corresponding mutations were introduced into the full-length Tat protein, and trans-activation activity was measured. Systematic substitution of arginine residues with alanines or lysines suggested that overall charge density is important but did not point to any specific residues as being essential for binding. A glutamine-to-alanine substitution had no effect on binding. Remarkably, peptides with scrambled or reversed sequences showed the same affinity and specificity for TAR RNA as the wild-type peptide. Trans-activation activity of the mutant Tat proteins correlated with RNA binding. Arginine-rich peptides from SIV Tat and from HIV-1 Rev, which can functionally substitute for the basic region of HIV-1 Tat, also bound specifically to TAR. Circular dichroism spectra suggest that the arginine-rich region of Tat is unstructured in the absence of RNA, becomes partially or fully structured upon binding, and induces a conformational change in the RNA. These results suggest that arginine-rich RNA-binding domains have considerable sequence flexibility, reminiscent of acidic domains found in transcriptional activators, and that RNA structure may provide much of the specificity for the interaction.

[Key Words: HIV Tat; viral trans-activator-, RNA-binding protein; TAR RNA; arginine-rich motif; peptide structure] Received October 31, 1990; revised version accepted December 4, 1990.

The Tat protein from human immunodeficiency virus recent in vitro trans-activation experiments (Marciniak (HIV) is a potent viral trans-activator (Sodroski et al. et al. 1990a). 1985a) that is essential for viral replication (Dayton et al. Trans-activation by Tat is dependent on a region near 1986; Fisher et al. 1986). Tat increases the rate of tran­ the start of transcription in the viral LTR called the scription from the HIV long terminal repeat (LTR) trans-acting responsive (TAR) element (Rosen et al. (Cullen 1986; Peterlin et al. 1986; Wright et al. 1986; 1985). TAR RNA forms a stable stem-loop structure Hauber et al. 1987; Muesing et al. 1987; Rice and (Muesing et al. 1987), and maintaining this structure is Mathews 1988; Laspia et al. 1989) and has also been pro­ important for the Tat response (Feng and Holland 1988; posed to increase translational efficiency (Cullen 1986; Hauber and Cullen 1988; Jakobovits et al. 1988; Feinberg et al. 1986; Rosen et al. 1986; Wright et al. Berkhout and Jeang 1989; Garcia et al. 1989; Selby et al. 1986; Muesing et al. 1987; Braddock et al. 1989; Roy et 1989; Roy et al. 1990c). Several experiments support the al. 1990b). Several experiments suggest that a major ef­ idea that TAR RNA, and not DNA, is essential for Tat fect of Tat is to increase the efficiency of transcriptional activation (Berkhout et al. 1989; Braddock et al. 1989). elongation (Kao et al. 1987; Laspia et al. 1989; Selby et al. TAR contains a 6-nucleotide loop and a 3-nucleotide py- 1989), and a recent study using intact HIV suggests that rimidine bulge that are essential for Tat activity. It ap­ this mechanism may account for the entire effect of Tat pears that cellular factors bind to the loop sequence within the virus (M.B. Feinberg, D. Baltimore, and A.D. (Gatignol et al. 1989; Gaynor et al. 1989; Marciniak et al. Frankel, in prep.). The elongation model is consistent 1990b) and that Tat binds to the bulge (Dingwall et al. with the observation that Tat acts on nascent RNA tran­ 1989; Muller et al. 1990; Roy et al. 1990a; Weeks et al. scripts (Berkhout et al. 1989) and is strongly supported by 1990). Tat is 86 amino acids long and contains a highly con­ ^Corresponding author. served cysteine-rich region (with 7 cysteines in 16 resi-

GENES & DEVELOPMENT 5:201-210 © 1991 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/91 $1.00 201 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Calnan et al.

+310 G+34 1 10 20 C A C G Met Glu Pro Val Asp Pro Arg Leu Glu Pro Trp Lys His Pro Gly Ser Gin Pro Lys Thr G C 21 30 40 A U Ala Cys Thr Asn Cys Tyr Cys Lys Lys Cys Cys Phe His Cys Gin Val Cys Phe lie Thr c" I Figure 1. Sequence of the HIV-1 Tat pro- o • J T- A D • T-l, U • i^ *^ JJ- 55 55-, ^° +23 AU tern and TAR site. The basic region or Tat, Lys Ala LOU Gly Ilo Ser |Tyr Cly Arg Lys Lys Arg Arg Gin Arg Arg Arg Pro| Pro Gin G C corresponding to the Tat 47-58 peptide, is A U C G highlighted. The numbering shown for the Gly Ser Gin Thr His Gin Val Ser Leu Ser Lys Gin Pro Thr Ser Gin Ser Arg Gly Asp +18;C' G^ TAR sequence is relative to the start of g^ gg G C transcription from the HIV LTR. pro rhr ciy Pro Lys GIU G C

dues) and a highly conserved basic region (with 2 lysines sociation constant of 6 nM for the complex. This binding and 6 arginines in 9 residues) (Arya et al. 1985; Sodroski constant was confirmed by titrating the peptide at sev­ et al. 1985b). The cysteine-rich region is essential for Tat eral RNA concentrations (Fig. 2). Clearly, at an RNA function (Garcia et al. 1988; Kubota et al. 1988; Sadaie et concentration above the K^ (10 nM), the fraction of bound al. 1988; Kuppuswamy et al. 1989; Ruben et al. 1989; RNA was greater than at a concentration below the K^ (2 Rice and Carlotti 1990) and mediates the formation of nM). At high RNA concentrations virtually all of the metal-linked dimers in vitro (Frankel et al. 1988a,b). The RNA was bound at a 1 : 1 peptide/TAR stoichiometry basic region is important for nuclear localization (Dang (data not shown), suggesting that one peptide binds per and Lee 1989; Endo et al. 1989; Hauber et al. 1989; TAR molecule. An unrelated arginine-rich peptide, prot­ Ruben et al. 1989; Siomi et al. 1990) and mediates spe­ amine, caused the RNA to precipitate in the wells of the cific binding to TAR RNA (Weeks et al. 1990). Here we gel and did not give a gel shift at any concentration show that the amino acid requirements for specific RNA tested (data not shown). The dissociation constant of the binding are surprisingly flexible, reminiscent of acidic Tat 47-58/TAR complex and the binding stoichiometry activation domains of transcription factors, and that the are similar to values reported by Weeks et al. (1990). arginine-rich RNA-binding domain of Tat is unstruc­ The specificity of Tat 47-58 for TAR was assessed by tured in solution and becomes partially or fully struc­ comparing binding to wild-type and mutant TAR RNAs. tured upon interaction with RNA. We discuss the impli­ Genetic experiments have shown that the 3-nucleotide cations for RNA recognition by arginine-rich domains. pyrimidine bulge in TAR is important for trans-activa­ tion (Berkhout et al. 1989; Roy et al. 1990b), and both purified Tat protein and Tat fragments bind specifically Results to this region (Roy et al. 1990a; Weeks et al. 1990). As A 12-amino-acid Tat peptide binds specifically to TAR shown in Figure 3, Tat 47-58 also shows specificity for RNA the TAR bulge, with at least 20-fold higher specificity for wild-type TAR than for mutant TAR RNAs that con­ The TAR site is an RNA stem-loop structure that is tained either a deletion of the 3-nucleotide bulge or a located just 3' to the start of viral transcription and is single-nucleotide substitution within the bulge. A 4-nu- necessary for Tat trans-activation (see Fig. 1). Recently, it has been shown that the HIV-1 Tat protein and frag­ ments of Tat containing the basic region bind specifi­ cally to TAR RNA (Dingwall et al. 1989; Muller et al. 0) 1990; Roy et al. 1990a; Weeks et al. 1990). We had ini­ t 47-58/TAR tially examined the binding of a set of synthetic peptides, 1 g 1 2 3 5 10 used previously to define the regions of Tat required for trans-activation (Frankel et al. 1989), and found that res­ idues 38-58 were sufficient for specific binding to a 57- 2nM nucleotide TAR RNA (data not shown). From these re­ sults and the results of Weeks et al. (1990), it was clear that RNA binding resided in the basic region. We syn­ thesized a peptide. Tat 47-58, that contained only the lOnM basic region of Tat (Fig. 1) and tested it for binding to a 31 nucleotide TAR RNA (Fig. 1). This top part of the TAR Figure 2. TAR RNA binding by the Tat 47-58 peptide. Com­ stem-loop has been shown by genetic analyses to be suf­ plexes were formed at the indicated peptide/TAR molar ratios ficient for the Tat response (Hauber and Cullen 1988; and analyzed on 10% polyacrylamide gels as described in Ma­ Jakobovits et al. 1988). A single band representing the terials and methods. Binding reactions were carried out at two TAR RNA concentrations (2 and 10 nM). The differences in the RNA-peptide complex was observed as the concentra­ fraction of RNA bound at these concentrations show that the tion of Tat 47-58 in the binding reaction was increased binding constant falls between these values. Unbovmd TAR (Fig. 2). By analyzing binding curves we calculated a dis­ RNA is also shown (no peptide).

202 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

RNA recognition by peptides

RNA binding of mutant peptides G G G G To define specific amino acids within the basic region op; that might be important for recognition of TAR, we syn­ thesized a series of mutant Tat 47-58 peptides and mea­ c sured their affinities and specificities for TAR (Table 1; Fig. 5). We focused primarily on arginine residues be­ cause HIV-1 Tat, simian immunodeficiency virus (SIV) 0 3 10 30 6:i0f 0 3 10 30 6:i0i 0 3 10 30 6i0r 0 3 10 30 60 Tat, and HIV-1 Rev all contain a basic region character­ ized by a cluster of arginine residues. It is known that SIV Tat can trans-activate the HIV-1 promoter (Viglianti and Mullins 1988) and that the basic region of Rev can •^^^iimi^-UK** Wf ^ functionally substitute for the Tat basic region (Endo et al. 1989; Subramanian et al. 1990). Figure 3. Binding of Tat 47-58 to TAR RNA mutants. TAR Single arginine residues were substituted with alanine RNAs were transcribed in vitro, and binding to Tat 47-58 was measured by gel shift analysis at the peptide/RNA ratios indi­ residues at positions 52, 53, 55, and 56, and each was cated. Mutants shown are a 4-nucleotide substitution in the found to reduce binding approximately twofold (Table 1; loop, which has little effect on binding, and a single-nucleotide Fig. 5). Substitution of glutamine 54 with alanine had no substitution and 3-nucleotide deletion of the bulge, which both effect. Substituting pairs of arginine with alanine resi­ decrease binding. Unbound TAR RNAs are shown in all cases (0 dues at positions 52 and 53 or positions 55 and 56 re­ peptide). duced binding by >20-fold. Binding was restored to the wild-type level by substituting arginine residues 55 and 56 with lysine instead of alanine residues, whereas sub­ stituting arginine residues 52 and 53 with lysine residues restored binding to a level twofold lower than wild type. cleotide substitution in the loop had a slight effect on Replacing any combination of two arginine residues at binding (Fig. 3). Changes in the loop have been shown positions 52, 53, 55, and 56 with alanine residues re­ previously to have little effect on the binding of full- duced binding by >20-fold. length Tat (Dingwall et al. 1989; Roy et al. 1990a) or of other Tat fragments (Weeks et al. 1990) to TAR. Com­ petition experiments using unlabeled wild-type and mu­ tant TAR RNAs provided further evidence for a specific Table 1. Dissociation constants of Tat 47-58 and variant peptides interaction (Fig. 4). Wild-type TAR RNA competed with the peptide-TAR complex at about a fivefold lower con­ K^ centration than did TAR with the 3-nucleotide bulge de­ (X 10-'M) letion. Competition experiments with the single-nucle­ 47-58 YGRKKRRQRRRP 6 otide bulge substitution or 4-nucleotide loop substitu­ ala52 YGRKKARQRRRP 10 tion confirmed the specificity. Completely unrelated ala53 YGRKKRAQRRRP 12 RNAs, such as tRNA, did not compete for Tat 47-58 ala54 YGRKKRRARRRP 5 binding even at a 200-fold excess (data not shown). Thus, ala55 YGRKKRRQARRP 12 Tat 47-58 specifically binds to the bulge in TAR RNA ala56 YGRKKRRQRARP 12 and shows specificity similar to the intact protein (Roy ala52ala53 YGRKKAAQRRRP >I00 et al. 1990a). A 9-amino acid peptide containing only the Iys52lys53 YGRKKKKQRRRP 13 basic residues (Tat 49-57) gave identical results (data not ala55ala56 YGRKKRRQAARP >100 shown). Iys551ys56 YGRKKRRQKKRP 4 ala52ala55 YGRKKARQARRP >100 ala52ala56 YGRKKARQRARP >100 ala53ala55 YGRKKRAQARRP >100 ala53ala56 YGRRRRAQRARP >I00 competitor Q 0 G G 38-58 FITKALGISYGRKKRRQRRRP 6 62-48 SGQPPRRRQRRKKRG 3 scrambled-1 YRKRRQRRGKRP 4 scrambled-2 YRKRGRQRRKRP 3 YEKSHRRRRTPKKAKA 6 3 1* '2 5 15 50 "2 5 15 50' SIV 76-91 Rev 34-50 TRQARRNRRRRWRERQR 2 Purified synthetic peptides were bound to wild-type TAR RNA Figure 4. Specific competition of Tat 47-58 binding by wild- (2 nM) at a minimum of five peptide concentrations. Binding type TAR RNA. Tat 47-58 was bound to 2 nM wild-type ^^P- constants were determined by calculating the concentration at labeled TAR RNA at a 3 : 1 peptide/TAR stoichiometry (lane which half the RNA was bound using the RNA gel shift assay. marked 47-58). Increasing amounts of unlabeled wild-type or 3 Experiments were repeated at least twice for each peptide, and nucleotide bulge deletion TAR RNAs (at the ratios indicated) the variation in binding constants between experiments was were added and analyzed on a 10% polyacrylamide gel. <20% in all cases.

GENES & DEVELOPMENT 203 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Calnan et al.

leled the measured affinities (data not shown). Thus, it appears that most or all of the binding specificity for OOOOOOE EcM<0'*lrtC0IMlrtwif5PJ«MMMt~~C^ TAR resides within the basic region of Tat.

Tians-activation activity of basic region mutants To determine whether RNA binding in vitro is relevant to trans-activation in vivo, we used the mammalian ex­ pression vector pSV2tat72 (Frankel and Pabo 1988) to construct a set of plasmids encoding mutant Tat pro­ teins. Single alanine substitutions were introduced at Figure 5. RNA binding and native gel of Tat 47-58 and variant arginine 53, glutamine 54, and arginine 55, and double peptides. [Top] RNA gel shift of each peptide at a 5 : 1 peptide/ alanine or double lysine substitutions were introduced at TAR molar ratio (2 nM RNA). The extent of the gel shift is arginine 52/arginine 53 and arginine 55/arginine 56. We proportional to the peptide molecular weight. [Bottom) Native 20% polyacrylamide gel (pH 4.5) of peptides (2 |j,g each). A single also made a mutant containing the scrambled-2 se­ band is observed for each peptide, and each migrates according quence. Plasmids were transfected into HeLa cells that to its molecular weight. contained an HIV-1 LTR chloramphenicol acetyl trans­ ferase (CAT) reporter (HL3TI cells; Felber and Pavlakis 1988), and trans-activation activity was measured. Sin­ gle arginine to alanine substitutions at positions 53 and These results suggested that the overall charge density 55 reduced Tat activity somewhat (Fig. 6A), which cor­ of Tat 47-58 is important but did not point to specific related with a twofold decrease in peptide RNA binding residues in the peptide as being essential for binding. To (Table 1). The double arginine to alanine substitutions examine sequence specificity more closely, v^^e synthe­ showed much weaker trans-activation activity, with sized another peptide in w^hich the sequence of the basic substitution at positions 52 and 53 almost completely region was reversed (synthesized from the carboxyl to eliminating activity (Fig. 6A). These substitutions also amino terminus), and two in which the amino acid se­ reduced RNA binding markedly (Table 1). When arginine quence was scrambled. Remarkably, all three peptides residues at positions 52/53 or 55/56 were replaced with specifically bound TAR with a higher affinity than the lysine instead of alanine residues, trans-activation activ­ wild-type peptide (Table 1; Fig. 5). It should be appreci­ ity was restored (Fig. 6A), as was RNA binding (Table 1). ated that it is not possible to change every residue in a Interestingly, substituting lysine residues at positions 52 single scrambled peptide because this region is so argin- and 53 did not quite restore trans-activation or RNA ine-rich. For example, the scrambled-1 peptide contains binding to wild-type levels, whereas substituting lysine a 6-amino acid homology to the wild-type peptide residues at positions 55 and 56 completely restored both (KRRQRR). This sequence is not present in the scram­ activities. The scrambled-2 protein showed wild-type bled-! peptide (Table 1). trans-activation (Fig. 6A), which correlated with the We then asked whether an arginine-rich peptide from wild-type affinity of the scrambled-2 peptide (Table 1). SIV Tat, which trans-activates the HIV-1 promoter, Substitution of glutamine 54 with alanine had no effect could bind to TAR and found that it bound with affinity on trans-activation (Fig. 6A) or RNA binding (Table I). and specificity equal to that of Tat 47-58 (Table 1; Fig. To assess Tat activity quantitatively, each plasmid 5). A similar arginine-rich region from Rev, which can was transfected at several concentrations, and trans-ac­ functionally substitute for the HIV-1 Tat basic region tivation activity was determined (Fig. 6B,C). Again, (Endo et al. 1989; Subramanian et al. 1990), also bound to trans-activation paralleled RNA binding. Immunopre- TAR RNA with wild-type affinity and specificity (Table cipitation of '^S-labeled extracts from COS-1-transfected I; Fig. 5). Beyond the fact that the SIV Tat and Rev pep­ cells showed that the level of expression of each protein tides are arginine-rich, they do not share significant se­ was similar (data not shown). Trans-activation in the quence homology with the HIV-1 Tat basic region (Ta­ COS-1 cells (measured by cotransfecting Tat plasmids ble I). with an HIV-CAT reporter) was the same as in the HeLa cells (data not shown). These results demonstrate a Figure 5 shows the binding of all peptides, at the same strong correlation between in vitro RNA binding by the concentration, to compare relative affinities and migra­ arginine-rich Tat peptides and trans-activation activity tion. Each peptide-TAR complex resolved in the gel ac­ of the full-length protein in vivo. cording to its relative peptide molecular weight, and the unbound peptides resolved by weight on native poly­ acrylamide gels (Fig. 5). It should be noted that Tat 38- 58, which contains additional residues between the cys- Circular dichroism of Tat 47-58 and TAR RNA teine-rich and basic regions of Tat, binds with an affinity As a preliminary examination of the structure of the ba­ equal to that of Tat 47-58. In addition, we measured the sic region of Tat, we measured circular dichroism (CD) relative specificities of all peptides by comparing their spectra of the Tat 47-58 peptide. The CD spectrum of binding to the 3-nucleotide bulge deletion and to wild- the peptide (Fig. 7A) clearly indicates a random coil type TAR RNA. In all cases, binding specificity paral­ structure (for basis spectra, see Creighton 1984). We also

204 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

RNA recognition by peptides

B

f2 ^ ^ <^ -Q

uo c\j c\i Lf) LO tz _co _c6 iU22 iIf3) lwo ^^ Lo 2 CO (Tj (0 fO -^ CO

• • • •

plasmid (ng) plasmid (ng) Figure 6. Tra/js-activation by Tat basic region mutants. [A] CAT assay of HeLa cells (containing an HIV-LTR CAT reporter) transfected with 25 ng of each Tat plasmid indicated, adjusted to 1 [ig of total DNA with nonspecific plasmid DNA. Acetylated (ac) and unacetylated (cm) [''*C|chloramphenicol was separated by thin layer chromatography. pSV2tat72 encodes wild-type Tat (residues 1-72) expressed from the SV40 early promoter (Frankel and Pabo 1988). {B,C) HeLa cells were transfected with 1-500 ng of each plasmid and adjusted to 1 jjig of total DNA, and CAT activity was quantitated as described (Frankel et al. 1989). In A, some points are beyond the linear range of the CAT assay; however, for quantition {B and C), assays were repeated with an appropriate amount of extract. [B] (•) Wild type, K^ 6 nM; (•) scrambled-2, K^ 3 nM; (D) ala54, K^ 5 HM; (O) ala53, K^ 12 nM; (A) ala55, K^ 12 nM. (C) (•) Wild type, K^ 6 nM; (A) Iys55lys56, K^ 4 nM; (a) Iys52lys53, K^ 13 nM; (A) ala55ala56, K^ >100 nM; (•) ala52ala53, K^ >100 nM. (Dissociation constants are from Table 1.)

measured spectra of TAR RNA and of the peptide-TAR dissociation constant. Tat 47-58/tRNA complexes were complex at a 1 : 1 stoichiometry (Fig. 7B). The spectrum also prepared but were found to be almost completely of TAR alone is similar to that of other RNAs (Aboul-ela insoluble. Interestingly, the Tat 47-58/wild-type TAR et al. 1988; Daly et al. 1990). The TAR spectrum was complexes also became insoluble when the peptide- subtracted from the spectrum of the complex to give a RNA ratio was >1 : I, suggesting that nonspecific bind­ difference spectrum (Fig. 7C) that contains contributions ing occurs at higher stoichiometrics. These solubility both from the peptide and from conformational changes properties are similar to those seen with Rev-Rev re­ in TAR that might have occurred upon peptide binding. sponsive element (RRE) and Rev/tRNA complexes (Daly The difference spectrum has a minimum near 260 nm, et al. 1990). Consistent with nonspecific binding is that which is almost certainly due to the RNA, as there is no protamine/TAR complexes are insoluble. From the CD known peptide conformation that contributes to the el- and solubility data we can tentatively conclude that Tat lipticity at this wavelength. There is a second minimum 47-58 is unstructured by itself, that it becomes partially near 216 nm, which probably consists of contributions or fully structured when bound to TAR, and that changes from both the peptide and a change in RNA conforma­ in the RNA conformation occur upon binding. Because tion. It is not possible to determine how much of the some of these properties are also seen with the bulge signal is due to peptide and how much to RNA, so we deletion RNA, at least some changes probably result cannot accurately calculate a peptide secondary struc­ from nonspecific interaction. More detailed studies us­ ture. However, it seems likely that the peptide contrib­ ing nuclear magnetic resonance (NMR) and x-ray crys­ utes significantly to the signal, because most peptide tallography will be needed to fully determine the nature conformations show a CD signal between 215 and 225 of these changes. nm and would cause the observed red shift of the 216 nm minimum (compared to the 212 nm minimum in TAR alone). Furthermore, the 216 nm minimum is much Discussion larger than might be expected from changes in RNA We have shown that a 12-amino-acid arginine-rich pep­ alone (compared to the 260 nm minimum). tide from the HIV-1 Tat protein, residues 47-58, binds To assess the specificity of these changes, we mea­ specifically to TAR RNA but that the precise amino acid sured the CD spectra of Tat 47-58 bound to TAR con­ sequence necessary for specific RNA recognition is sur­ taining a deletion of the 3-nucleotide bulge. The differ­ prisingly flexible; a peptide synthesized in the opposite ence spectrum of the complex was similar to that seen orientation, from the carboxyl to amino terminus, and with wild-type TAR RNA, although the intensities of peptides with scrambled sequences bind to TAR with the minima were reduced by —50% (data not shown). We affinity equal to the wild-type peptide. The overall assume that this reduction was not due to incomplete charge appears to be essential for RNA binding because peptide binding because the RNA and peptide concen­ replacing single arginine with alanine residues reduces trations were at least one order of magnitude above the binding by twofold, whereas substituting pairs of argin-

GENES & DEVELOPMENT 205 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Calnan et al.

supporting other studies suggesting that Tat binding to TAR is important for Tat function (Dingwall et al. 1989; MMMlM*M«i««^a Muller et al. 1990; Roy et al. 1990a; Weeks et al. 1990). This is supported further by the demonstration that RNA binding is important for Tat-mediated trans-acti­ vation of transcription in vitro (Marciniak et al. 1990a). oS>' -20000- Southgate et al. (1990) have shown indirectly that RNA binding is important for Tat function by making a Tat- ^ -30000 Rev fusion protein and replacing TAR with the RRE to provide an RNA-binding site for the fusion protein. Only -40000 220 240 260 the fusion protein, and not Tat or Rev alone, could trans- X (nm) activate. Similar results were observed when TAR was replaced with an MS2 coat protein-binding site or an API-binding site (Berkhout et al. 1990; Selby and Peter- B lin 1990). These experiments also demonstrate that al­ though Tat binds directly to RNA, the specific Tat-TAR interaction is not essential for trans-activation. •O 50000 Our results and those of Weeks et al. (1990) suggest that the arginine-rich region of Tat is an independent RNA-binding domain and that Tat contains independent Vo RNA-binding and trans-activation domains. These do­ 0 mains appear to be structurally distinct because pro­ teases readily cleave Tat just amino-terminal to the basic 240 260 region (Frankel et al. 1988a). Arginine-rich motifs similar to Tat are present in a number of RNA-binding proteins, including antiterminators, Gag proteins, ribosomal pro­ teins, and HIV Rev (Lazinski et al. 1989). Our results are consistent with the proposal that these are specific RNA-binding domains (Lazinski et al. 1989) and suggest that although these domains have significantly different primary amino acid sequences, they may recognize sim­ ilar RNA structures. This is supported by the findings g" -30000- that an arginine-rich peptide from the HIV Rev protein binds specifically to TAR (Table 1; Fig. 5) and can func­ 0 tionally substitute for the basic region of Tat (Endo et al. 1989; Subramanian et al. 1990), and that the sequence of

200 220 240 260 280 300 the Tat basic region can be completely rearranged and X (nm) still bind and function normally. The similarity of argi­ nine-rich RNA recognition between proteins is further Figure 7. Circular dichroism of Tat 47-58 and TAR RNA. [A] underscored by circular dichroism experiments, suggest­ Spectrum of Tat 47-58 alone. [B] Spectra of TAR RNA alone (•) ing that both the Tat basic region and Rev cause similar and of the Tat 47-58/TAR RNA complex at a 1 : 1 stoichiom- etry (O). (C) Difference spectrum of the spectra shown in B. All conformational changes in the RNA structure upon values of mean molar ellipticity were normalized to the peptide binding (Fig. 7; Daly et al. 1990). Such changes in RNA concentration (3 JJLM) to allow direct comparison of the spectra. structure might influence the binding of cellular pro­ teins to other sites on the RNA. The basic region of Tat has been shown to be a nuclear localization signal and also targets Tat to the nucleolus (Dang and Lee 1989; Endo et al. 1989; Hauber et al. 1989; ine with alanine residues reduces binding by >20-fold. Ruben et al. 1989; Siomi et al. 1990; Subramanian et al. When pairs of arginine residues are replaced with lysine 1990). Nucleolar localization correlates well with trans- instead of alanine residues, RNA binding is restored to activation activity. A reasonable interpretation is that near wild-type levels. Not every position in the sequence nucleolar localization is a consequence of RNA binding; is equally important because substitution of arginine res­ expression of the high levels of Tat needed for immuno­ idues at positions 52 and 53 has a slightly greater effect fluorescence may allow binding to nonspecific RNAs than substitution of arginines at positions 55 and 56. The containing TAR-like bulges. Such bulges might be specific sequence requirements are subtle, however, be­ present in ribosomal RNAs that are highly structured cause the sequence of the basic region can be scrambled (Stem et al. 1989) and are synthesized in the nucleolus. and still give wild-type binding. Interestingly, several ribosomal proteins contain argin­ Our results show a strong correlation between peptide ine-rich regions similar to Tat, and these regions have RNA-binding affinities and trans-activation activity, been proposed to mediate specific RNA binding (Lazin-

206 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

RNA lecognition by peptides ski et al. 1989). This interpretation is consistent with the proteins. Several recent studies have shown that the ba­ behavior of a trans-dominant Tat mutant that localizes sic DNA-binding domain adopts an a-helical structure to the nucleus but not to the nucleolus (Pearson et al. upon specific DNA binding (O'Neil et al. 1990; Talanian 1990). This protein is truncated at glutamine 54 and et al. 1990; Weiss 1990). Thus, it appears that interaction therefore is unlikely to bind RNA but presumably can of protein domains with other proteins, RNA, or DNA still interact with cellular factors or other Tat molecules can all induce conformational changes required for spe­ in the nucleus to produce the trans-dominant pheno- cific recognition. type. Furthermore, our ala55/ala56 mutant is more ac­ tive than our ala52/ala53 mutant (Fig. 6A,C), consistent with the idea that basic residues at positions 52 and 53 Materials and methods are involved in nuclear localization in addition to RNA binding. Peptide synthesis, purification, and analysis How do arginine-rich regions interact specifically with Syntheses were performed using Fmoc chemistry on a Milli- RNA? The observations that Tat 47-58 is unstructured Gen/Biosearch model 9600 peptide synthesizer with a peptide when not bound to RNA and that the sequence require­ amide linker-norleucine-4-methylbenzyhydrylamine (Pal- ments for RNA binding are highly flexible suggest that Nle-MBHAj polystyrene resin (MilliGen/Biosearch, 0.5 gram). the RNA may provide much of the information for rec­ The benzotriazolyloxytris (dimethylamino)phosphonium hexa- ognition. Indeed, mutagenesis experiments show that fluorophosphate/l-hydroxybenzotriazole (BOP/HOBt) and N- methylmorpholine/DMF coupling method was used with cou­ many mutations abolish TAR function (Hauber and pling times of 1-4 hr. Protecting groups were 2,2,5,7,8-penta- CuUen 1988; Jakobovits et al. 1988; Feng and Holland methyl chroman-6-sulfonyl (Arg), t-butyloxycarbonyl (Lys), 1988; Garcia et al. 1989; Selby et al. 1989; Berkhout and trityl (Gin, Asn, His), t-butyl (Ser, Thr, Tyr), and t-butyl ester Jeang 1989; Roy et al. 1990c), whereas here we show that (Glu). All peptides were synthesized as their carboxy-terminal the basic region of Tat can be completely rearranged and amides. After syntheses were completed, protecting groups still retain its function. It seems likely that the TAR were removed and the peptide chains were cleaved from the bulge is a highly structured element that interacts spe­ resin with trifluoroacetic acid/thioanisole/ethanedithiol/anis- cifically with arginine residues. Other RNA hairpins and ole [90 : 5 : 3 : 2 (vol/vol)] for 2 hr. The mixtures were filtered bulges are known to form very stable, ordered tertiary into cold anhydrous diethyl ether, and the precipitated peptides structures (Bhattacharyya et al. 1990; Cheong et al. 1990; were filtered, washed with additional ether, and dried. Puglisi et al. 1990), and in the few well-studied examples Amino acid composition was determined by hydrolysis in 6 M of RNA-protein recognition, RNA structure is clearly an HCl containing 0.5% phenol at 110°C and analysis on an LKB 4151 Alpha Plus analyzer. Peptides were purified on a C4 re­ important component of the interaction (Romaniuk et verse-phase HPLC column (Vydac) using an acetonitrile gradi­ al. 1987; Rouid et al. 1989; Stem et al. 1989). "Specific" ent of 0.2%/min in 0.1% trifluoroacetic acid. Peptide absorp­ contacts between Tat and TAR are likely to result from tion spectra were recorded, and concentrations of most peptides ionic interactions between arginine amino groups and were determined by tyrosine absorbance at 278 nm (e = 1420 RNA phosphates arranged in a specific tertiary confor­ M ' cm" '; Creighton 1984). The concentration of the Rev pep­ mation rather than from base-specific hydrogen bonds or tide was determined by trytophan absorbance at 278 nm other interactions such as those seen in DNA-binding (e = 5600 M " 'cm~ '; Creighton 1984) and the concentration of proteins (Jordan and Pabo 1988). The ionic nature of the Tat 62—48 by peptide absorbance at 229 nm using Tat 47-58 as interaction is supported by arginine to lysine substitu­ a standard. Purity and concentrations were confirmed by native tions at positions 52, 53, 55, and 56, which still show gel electrophoresis (20% polyacrylamide in 30 mM sodium ac­ specific RNA binding and trans-activation. The only etate at pH 4.5), and peptides were visualized by Coomassie blue staining. The mass of several peptides was confirmed by fast nonbasic amino acid in this region, glutamine 54, atom bombardment (FAB) mass spectrometry (University of seemed a good candidate for a base-specific contact be­ California, Berkeley) as described previously (Frankel et al. cause it is known that glutamine can form specific hy­ 1989). drogen bonds with bases in the major groove of DNA (Jordan and Pabo 1988); however, substituting alanine for this glutamine did not affect binding or trans-activation. RNA synthesis and purification The sequence flexibility shown by these arginine-rich TAR RNAs were transcribed in vitro by T7 RNA polymerase domains is reminiscent of acidic trans-activating do­ from synthetic oligonucleotide templates containing a 17-base mains of transcription factors (Ptashne 1988). Trans-ac­ double-stranded T7 promoter and single-stranded TAR se­ tivating domains have been proposed to be unstructured, quences coding for sense TAR RNAs. All RNAs contained GG "negative noodles" (Sigler 1988), which become struc­ at their 5' end, which increases the efficiency of transcription tured upon interaction with other molecules, presum­ (Milligan and Uhlenbeck 1989) and CC at the 3' end to base-pair ably components of the transcription apparatus. Our cir­ with the Gs. Wild-type TAR RNA contained sequences -I-18 to cular dichroism results suggest that the Tat peptide is -1-44 of the HIV LTR TAR site (see Fig. 1). Mutant TARs also contained this sequence but with substitutions or deletions as unstructured when not bound to RNA and becomes par­ highlighted in Figure 3. All RNAs were purified on 10% poly- tially or fully structured upon binding, suggesting that acrylamide/8 M urea gels, eluted from gels in 0.5 M ammonium arginine-rich RNA-binding domains may interact with acetate, 10 mM magnesium acetate, 1 mM EDTA, and 0.1% SDS, RNA in an analogous manner. "Induced" structure is extracted twice with phenol, and ethanol-precipitated. Purified also seen in the DNA-binding domains of leucine zipper RNA was resuspended in sterile deionized water. The concen-

GENES &, DEVELOPMENT 207 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Calnan et al.

trations of radiolabeled RNAs were determined from the spe­ The publication costs of this article were defrayed in part by cific activity of [^^P]CTP incorporated into the transcripts. Un­ payment of page charges. This article must therefore be hereby labeled RNA was quantitated by spectrophotometry. marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact. RNA-binding gel mobility shift assays Peptide and RNA were incubated together for 10 min on ice in References lO-jil binding reactions containing 10 mM Tris-HCl (pH 7.5), 70 mM NaCl, 0.2 mM EDTA, and 5% glycerol. To determine bind­ Aboul-ela, F., G. Varani, G.T. Walker, and I. Tinoco, Jr. 1988. ing constants, 2 nM radiolabeled wild-type TAR RNA was ti­ The TFIIIA recognition fragment d(GGATGGGGAG)- trated with peptide. At least five concentrations of a given pep­ d(CTCCCCATCC) is B-form in solution. Nucleic Acids Res. tide were used to determine the binding constant of the peptide. 16: 3559-3572. Peptide-RNA complexes were resolved on 10% polyacryl- Aldovini, A., C. Debouck, M.B. Feinberg, M. Rosenberg, S.K. amide, 0.5 x TBE gels that had been prerun for 1 hr and allowed Arya, and F. Wong-Staal. 1986. Synthesis of the complete to cool to 4°C. Gels were electrophoresed at 200 V for 3 hr at trans-activation gene product of human T-lymphotrophic vi­ 4°C, dried, and autoradiographed. A beta scanner was used to rus type III in Escherichia coli: Demonstration of immuno- quantitate peptide-bound RNA and free RNA. genicity in vivo and expression in vitro. Proc. Natl. Acad. Sci. 83: 6672-6676. Construction of mutant plasmids, CAT assays, and Arya, S.K., C. Guo, S.F. Josephs, and F. Wong-Staal. 1985. Trans- imm unoprecipitation activator gene of human T-lymphotrophic virus type III (HTLV-III). Science 229: 69-73. Oligonucleotide cassettes containing the basic region mutants Berkhout, B. and K.-T. Jeang. 1989. Trans-activation of human were synthesized and cloned into the tat gene of pSV2tat72 immunodeficiency virus type 1 is sequence specific for both (Frankel and Pabo 1988). Mutations were confirmed by dideox- the single-stranded bulge and loop of the trans- ynucleotide sequencing (Sequenase; U.S. Biochemicals). Vari­ acting-responsive hairpin: A quantitative analysis. /. Virol. ous amounts of plasmid were transfected into HeLa cells or 63: 5501-5504. COS-1 cells (50% confluent in 25-mm wells) by lipofection (Fei­ Berkhout, B., R.H. Silverman, and K.-T. Jeang. 1989. Tat trans- gner et al. 1987), cells were harvested 48 hr after transfection, activates the human immunodeficiency virus through a na­ and CAT activity was assayed and quantitated as described scent RNA target. Cell 59: 273-282. (Frankel et al. 1989). In all cases, plasmid concentrations were Berkhout, B., A. Gatignol, A.B. Rabson, and K.-T. Jeang. 1990. adjusted to 1 ^.g of total DNA with nonspecific pBR322 DNA. TAR-independent activation of the HIV-1 LTR: Evidence Cells were grown in Dulbecco's modified Eagle medium sup­ that Tat requires specific regions of the promoter. Cell plemented with 10% fetal bovine serum and penicillin/strepto­ 62: 757-767. mycin. For immunoprecipitations, COS-1 cells (50% confluent Bhattacharyya, A., A.I.H. Murchie, and D.M.J. Lilley. 1990. in 35-mm dishes) were transfected with the various plasmids for RNA bulges and the helical periodicity of double-stranded 48 hr, cells were incubated with serum-free medium containing RNA. Nature 343: 484-487. [''^S]cysteine for 4 hr, and Tat was immunoprecipitated with a Braddock, M., A. Chambers, W. Wilson, M.P. Esnouf, S.E. Ad­ rabbit polyclonal antibody (Aldovini et al. 1986) and analyzed as ams, A.J. Kingsman, and S.M. Kingsman. 1989. HIV-1 TAT described (Cullen 1987). "activates" presynthesized RNA in the nucleus. Cell 58: 269-279. Circular dichroism Cheong, C, G. Varani, and I. Tinoco, Jr. 1990. Solution struc­ ture of an unusually stable RNA hairpin, 5'GGAC(UUCG)- Circular dichroism spectra were measured using an Aviv model GUCC. Nature 346: 680-682. 60DS spectropolarimeter, equipped with a Hewlett-Packard Creighton, T.E. 1984. Proteins: Structures and molecular prin­ Peltier temperature controller. Spectra were recorded from 300 ciples. W.H. Freeman and Company, New York. to 200 nm and averaged over five scans, with a 10 sec averaging Cullen, B.R. 1986. Trans-activation of human immunodeficien­ time at each wavelength. Samples were prepared in 10 mM po­ cy virus occurs via a bimodal mechanism. Cell 46: 973-982. tassium phosphate buffer (pH 7.0) and 70 mM potassium fluo­ . 1987. Use of eukaryotic expression technology in the ride. Concentrations of peptide and TAR RNA were determined functional analysis of cloned genes. Methods Enzymol. by absorption spectra. Samples were prepared in a 1-cm path- 152:684-704. length cylindrical cuvette, and all spectra were recorded at 5°C. Daly, T.J., J.R. Rusche, T.E. Maione, and A.D. Frankel. 1990. Mean molar residue ellipticity was calculated using a molecular Circular dichroism studies of the HIV-1 Rev protein and its mass for Tat 47-58 of 1657 daltons, and all spectra, including specific RNA binding site. Biochemistry 29: 9791-9795. the TAR RNA, were normalized using these values to allow Dang, C.V. and W.M.F. Lee. 1989. Nuclear and nucleolar tar­ direct calculation of the difference spectrum. geting sequences of c-erb-A, c-myb, N-myc, p53, HSP 70, and HIV tat proteins. /. Biol. Chem. 264: 18019-18023. Acknowledgments Dayton, A. I., J.G. Sodroski, C.A. Rosen, W.C. Goh, and W.A. Haseltine. 1986. The trans-activator gene of the human T We thank David Mann for help with CAT assays, Chris Siebel cell lymphotropic virus type III is required for replication. and Don Rio for T7 RNA polymerase, and Don Rio, Peter Kim, Cell 44: 941-947. Carl Pabo, Stephen Fiarrison, Phillip Sharp, Alan Sachs, Douglas Dingwall, C, I. Emberg, M.J. Gait, S.M. Green, S. Heaphy, J. Vollrath, Mark Feinberg, and Robert Talanian for interesting Kam, A.D. Lowe, M.Singh, M.A. Skiimer, and R. Valerio. discussions and helpful comments. This work was supported by 1989. Human immunodeficiency virus 1 tat protein binds the Lucille P. Markey Charitable Trust and by National Insti­ trans-activation-responsive region (TAR) RNA in vitro. tutes of Fiealth National Institute of Allergy and Infectious Dis­ Proc. Natl. Acad. Sci. 86: 6925-6929. eases grant AI29135 (A.D.F.). Endo, S., S. Kubata, H. Siomi, A. Adachi, S. Oroszlan, M. Maki,

208 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

RNA recognition by peptides

and M. Hatanaka. 1989. A region of basic amino-acid cluster Cell. Biol. 8:2555-2561. in HlV-1 Tat protein is essential for trans-acting activity and Jordan, S.R. and CO. Pabo. 1988. Structure of the lambda com­ nucleolar localization. Virus Genes 3: 99-110. plex at 2.5A resolution: Details of the repressor-operator in­ Feinberg, M.B., R.F. Jarrett, A. Aldovini, R.C. Gallo, and F. teractions. Science 242: 893-899. Wong-Staal. 1986. HTLV-III expression and production in­ Kao, S.-Y., A.F. Caiman, P.A. Luciw, and B.M. Peterlin. 1987. volve complex regulation at the levels of splicing and trans­ Anti-termination of transcription within the long terminal lation of viral RNA. Cell 46: 807-817. repeat of HIV-1 by tat gene product. Nature 330: 489-493. Felber, B. K. and G.N. Pavlakis. 1988. A quantitative bioassay Kubota, S., S.-I. Endo, M. Maki, and M. Hatanaka. 1988. Role of for HIV-1 based on transactivation. Science 239: 184—187. the cysteine-rich region of HIV Tat protein on its trans-ac- Feigner, P.L., T.R. Gadek, M. Holm, R. Roman, H.W. Chan, M. tivational ability. Virus Genes 2: 113-118. Wenz, J.P. Northrop, G.M. Ringold, and M. Danielsen. 1987. Kuppuswamy, M., T. Subramanian, A. Srinivasan, and G. Chin- Lipofection: A highly efficient, lipid-mediated DNA-trans- nadurai. 1989. Multiple functional domains of Tat, the fection procedure. Proc. Natl. Acad. Sci. 84: 7413-7417. trans-activator of HIV-1, defined by mutational analysis. Feng, S. and E.G. Holland. 1988. HIV-1 tat trans-activation re­ Nucleic Acids Res. 17: 3551-3561. quires the loop sequence within tar. Nature 334: 165-167. Laspia, M.F., A.P. Rice, and M.B. Mathews. 1989. HIV-1 Tat Fisher, A. G., M.B. Feinberg, S.F. Josephs, M.E. Harper, L.M. protein increases transcriptional initiation and stabilizes Marselle, G. Reyes, M.A. Gonda, A. Aldovini, C. Debouck, elongation. Cell 59: 283-292. R.C. Gallo, and F. Wong-Staal. 1986. The trans-activator Lazinski, D., E. Grzadzielska, and A. Das. 1989. Sequence-spe­ gene of HTLV-III is essential for virus replication. Nature cific recognition of RNA hairpins by bacteriophage antiter- 320:367-371. minators requires a conserved arginine-rich motif. Cell Frankel, A. D. and CO. Pabo. 1988. Cellular uptake of the Tat 59:207-218. protein from human immunodeficiency virus. Cell Marciniak, R.A., B.J. Calnan, A.D. Frankel, and P.A. Sharp. 55: 1189-1194. 1990a. HIV-1 Tat protein trans-activates transcription in Frankel, A. D., D.S. Bredt, and CO. Pabo. 1988a. Tat protein vitro. Cell 63: 791-802. from human immunodeficiency virus forms a metal-linked Marciniak, R.A., M.A. Garcia-Blanco, and P.A. Sharp. 1990b. dimer. Science 240: 70-73, Identification and characterization of a HeLa nuclear protein Frankel, A. D., L. Chen, R.J. Cotter, and CO. Pabo. 1988b. that specifically binds to the trans-activation-response Dimerization of the tat protein from human immunodefi­ (TAR) element of human immunodeficiency virus. Proc. ciency virus: A cysteine-rich peptide mimics the normal Natl. Acad. Sci. 87: 3624-3628. metal-linked dimer interface. Proc. Natl. Acad. Sci. Milligan, J.F. and O.C Uhlenbeck. 1989. Synthesis of small 85: 6297-6300. RNAs using T7 RNA polymerase. Methods Enzymol. Frankel, A. D., S. Biancalana, and D. Hudson. 1989. Activity of 180:51-62. synthetic peptides from the Tat protein of human immuno­ Muesing, M.A., D.H. Smith, and D.J. Capon. 1987. Regulation deficiency virus type 1. Proc. Natl. Acad. Sci. 86: 7397- of mRNA accumulation by a human immunodeficiency vi­ 7401. rus trans-activator protein. Cell 48: 691-701. Garcia, J. A., D. Harrich, L. Pearson, R. Mitsuyasu, and R.B. Muller, W.E.G., T. Okamoto, P. Renter, D. Ugarkovic, and H.C Gaynor. 1988. Functional domains required for tat-induced Schroder. 1990. Functional characterization of Tat protein transcriptional activation of the HIV-1 long terminal repeat. from human immunodeficiency virus. /. Biol. Chem. EMBO f. 7:3143-3147. 265:3803-3808. Garcia, J.A., D. Harrich, E. Soultanakis, F. Wu, R. Mitsuyasa, O'Neil, K.T., R.H. Hoess, and W.F. DeGrado. 1990. Design of and R.B. Gaynor. 1989. Human immunodeficiency virus DNA-binding peptides based on the leucine zipper motif. type 1 LTR TATA and TAR region sequences required for Science 249: 774-778. transcriptional regulation. EMBO f. 8: 765-778. Pearson, L., J. Garcia, F. Wu, N. Modesti, J. Nelson, and R. Gatignol, A., A. Kumar, A. Rabson, and K.-T. Jeang. 1989. Iden­ Gaynor. 1990. A transdominant tat mutant that inhibits tat- tification of cellular proteins that bind to the human immu­ induced gene expression from the human immunodeficiency nodeficiency virus type 1 frans-activation-responsive TAR virus long terminal repeat. Proc. Natl. Acad. Sci. 87: 5079- element RNA. Proc. Natl. Acad. Sci. 86: 7828-7832. 5083. Gaynor, R., E. Soultanakis, M. Kuwabara, J. Garcia, and D.S. Peterlin, B.M., P.A. Luciw, P.J. Barr, and M.D. Walker. 1986. Sigman. 1989. Specific binding of a HeLa cell nuclear protein Elevated levels of mRNA can account for the trans-activa­ to RNA sequences in the human immunodeficiency virus tion of human immunodeficiency virus. Proc. Natl. Acad. transactivating region. Proc. Natl. Acad. Sci. 86: 4858-4862. Sci. 83: 9734-9738. Hauber, J. and B.R. CuUen. 1988. Mutational analysis of the Ptashne, M. 1988. How eukaryotic transcriptional activators trans-activation-responsive region of the human immunode­ work. Nature 335: 683-689. ficiency virus type 1 long terminal repeat. /. Virol. 62: 673- Puglisi, J.D., J.R. Wyatt, and I. Tinoco, Jr. 1990. Solution con­ 679. formation of an RNA hairpin loop. Biochemistry 29: 4215- Hauber, J., A. Perkins, E.P. Heimer, and B.R. CuUen. 1987. 4226. Trans-activation of human immunodeficiency virus gene ex­ Rice, A. P. and F. Carlotti. 1990. Mutational analysis of the pression is mediated by nuclear events. Proc. Natl. Acad. conserved cysteine-rich region of the human immunodefi­ Sci. 84: 6364-6368. ciency virus type 1 Tat protein. /. Virol. 64: 1864—1868. Hauber, J., M.H. Malim, and B.R. CuUen. 1989. Mutational Rice, A.P. and M.B. Mathews. 1988. Transcriptional but not analysis of the conserved basic domain of human immuno­ translational regulation of HIV-1 by the tat gene product. deficiency virus tat protein. /. Virol. 63: 1181-1187. Nature 332: 551-553. Jakobovits, A., D.H. Smith, E.B. Jakobovits, and D.J. Capon. Romaniuk, P.J., P. Lowary, H.-N. Wu, G. Stormo, and O.C. 1988. A discrete element 3' of human immunodeficiency Uhlenbeck. 1987. RNA binding site of R17 coat protein. Bio­ virus 1 (HIV-1) and HIV-2 mRNA initiation sites mediates chemistry 26:1563-1568. transcriptional activation by an HIV trans activator. Mol. Rosen, C.A., J.G. Sodroski, and W.A. Haseltine. 1985. The lo-

GENES & DEVELOPMENT 209 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Calnan et al.

cation of cis-acting regulatory sequences in the human T cell specific DNA binding by a short peptide dimer. Science lymphotrophic virus type III (HTLV-III/LAV) long terminal 249: 769-771. repeat. Cell 41: 813-823. Viglianti, G.A. and J.I. Mullins. 1988. Functional comparison of Rosen, C.A., J.G. Sodroski, W.C. Goh, A.I. Dayton, J. Lippke, transactivation by simian immunodeficiency virus from and W.A. Haseltine. 1986. Post-transcriptional regulation rhesus macaques and human immunodeficiency virus type accounts for the trans-activation of the human T- 1./. Virol 62:4523-4532. lymphotrophic virus type III. Nature 319: 555-559. Weeks, K.M., C. Ampe, S.C. Schultz, T.A. Steitz, and D.M. Rould, M.A., J.J. Perone, D. Soil, and T.A. Steitz. 1989. Structure Crothers. 1990. Fragments of the HIV-1 Tat protein specifi­ of E. coli glutaminyl-tRNA synthetase complexed with cally bind TAR RNA. Science 249: 1281-1285. tRNA*^'" and ATP at 2.8A resolution. Science 246: 1135- Weiss, M.A. 1990. Thermal unfolding of a leucine zipper do­ 1142. main and its specific DNA complex: Implications for scis- Roy, S., U. Delling, C.-H. Chen, C.A. Rosen, and N. Sonenberg. sor's grip recognition. Biochemistry 29: 8020-8024. 1990a. A bulge structure in HIV-1 TAR RNA is required for Wright, CM., B.K. Felber, H. Paskalis, and G.N. Pavlakis. 1986. Tat binding and Tat-mediated trans-activation. Genes e) Expression and characterization of the trans-activator of Dev. 4: 1365-1373. HTLV-III/LAV virus. Science 234: 988-992. Roy, S., M.G. Katze, N.T. Parkin, I. Edery, A.G. Hovanessian, and N. Sonenberg. 1990b. Control of the interferon-induced 68-kilodalton protein kinase by the HIV-1 tat gene product. Science 1A7: 1216-1219. Roy, S., N.T. Parkin, C. Rosen, J. Itovitch, and N. Sonenberg. 1990c. Structural requirements for Hans activation of hu­ man immunodeficiency virus type 1 long terminal repeat- directed gene expression by tat: Importance of base pairing, loop sequence, and bulges in the tat-responsive sequence. /. Virol 64: 1402-1406. Ruben, S., A. Perkins, R. Purcell, K. Joung, R. Sia, R. Burghoff, W.A. Haseltine, and C.A. Rosen. 1989. Structural and func­ tional characterization of human immunodeficiency virus tat protein. /. Virol 63: 1-8. Sadaie, M.R., J. Rappaport, T. Bcntcr, S.F. Josephs, R. Willis, and F. Wong-Staal. 1988. Missensc mutations in an infectious human immunodeficiency viral genome: Functional map­ ping of tat and identification of the rev splice acceptor. Proc. Natl Acad. Sci. 85: 9224-9228. Sclby, M.J. and B.M. Peterlin. 1990. Trans-activation by HlV-1 Tat via a heterologous RNA binding protein. Cell 62: 769- 776. Selby, M.J., E.S. Bain, P.A. Luciw, and B.M. Peterlin. 1989. Structure, sequence, and position of the stem-loop in tar determine transcriptional elongation by tat through the HlV-1 long terminal repeat. Genes &> Dev. 3: 547-558. Sigler, P.B. 1988. Acid blobs and negative noodles. Nature 333:210-212. Siomi, H., H. Shida, M. Maki, and M. Hatanaka. 1990. Effects of a highly basic region of human immunodeficiency virus tat protein on nucleolar localization. /. Virol. 64: 1803-1807. Sodroski, J., C. Rosen, F. Wong-Staal, S. Salahuddin, M. Popovic, S. Arya, R.C. Gallo, and W.A. Haseltine. 1985a. Transacting transcriptional regulation of human T-cell leukemia virus type III long terminal repeat. Science 227: 171-173. Sodroski, J., R. Patarca, C. Rosen, F. Wong-Staal, and W. Hasel­ tine. 1985b. Location of the trans-activating region on the genome of human T-cell lymphotrophic virus type III. Sci­ ence 229: 74-77. Southgate, C, M.L. Zapp, and M.R. Green. 1990. Activation of transcription by HIV-1 Tat protein tethered to nascent RNA through another protein. Nature 345: 640-642. Stem, S., T. Powers, L.-M. Changchien, and H.F. Noller. 1989. RNA-protein interactions in 30S ribosomal subunits: Fold­ ing and functions of 16s rRNA. Science 244: 783-790. Subramanian, T., M. Kuppuswamy, L. Venkatesh, A. Srini- vasan, and G. Chinnadurai. 1990. Functional substitution of the basic domain of the HIV-1 trans-activator. Tat, with the basic domain of the functionally heterologous Rev. Virology 176:178-183. Talanian, R.V., C.J. McKnight, and P.S. Kim. 1990. Sequence-

210 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Analysis of arginine-rich peptides from the HIV Tat protein reveals unusual features of RNA-protein recognition.

B J Calnan, S Biancalana, D Hudson, et al.

Genes Dev. 1991, 5: Access the most recent version at doi:10.1101/gad.5.2.201

References This article cites 74 articles, 37 of which can be accessed free at: http://genesdev.cshlp.org/content/5/2/201.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.

Copyright © Cold Spring Harbor Laboratory Press