- Email: [email protected]
Structural studies of HIV-1 tat protein
J. MoL Biol. (1995) 247, 529-535
JMB COMMUNICATION
Structural Studies of HIV-1 Tat Protein R Bayer 1, M. Kraft 2, A. Ejchart 1, M. Westendorp 3, R. Frank 2 and R R6sch 1Lehrstuhl fiir Biopolymere und Bayreuther Institut ffir Makromolekfilforschung Universit~t Bayreuth D-95440 Bayreuth, Germany 2Zentrum fiir Molekularbiologie Heidelberg D-69120 Heidelberg Germany 3Deutsches Krebsforschungszentrum D-69120 Heidelberg Germany
*Corresponding author
Tat (trans-activator) proteins are early RNA binding proteins regulating lentiviral transcription. These proteins are necessary components in the life cycle of all known lentiviruses, such as the human immunodeficiency viruses (HIV) or the equine infectious anemia virus (EIAV). Tat proteins are thus ideal targets for drugs intervening with lentiviral growth. The consensus RNA binding motif (TAR, trans-activation responsive element) of HIV-1 is well characterized. Structural features of the 86 amino acid HIV-1, Zaire 2 isolate (HVIZ2) Tat protein in solution were determined by two dimensional (2D) nuclear magnetic resonance (NMR) methods and molecular dynamics (MD) calculations. In general, sequence regions corresponded to structural domains of the protein. It exhibited a hydrophobic core of 16 amino acids and a glutamine-rich domain of 17 amino acids. Part of the NH2 terminus, Val4 to Pro14, was sandwiched between these domains. Two highly flexible domains corresponded to a cysteine-rich and a basic sequence region. The 16 amino acid sequence of the core region is strictly conserved among the known Tat proteins, and the three-dimensional fold of these amino acids of HVIZ2 Tat protein was highly similar to the structure of the corresponding EIAV Tat domain. HVIZ2 Tat protein contained a well defined COOH-terminal Arg-Gly-Asp (RGD) loop similar to the recently determined decorsin RGD loop.
Keywords: tat protein; solution structure; HIV; NMR; molecular dynamics
The sequence of the 86 amino acid Tat protein from HVIZ2 (human immunodeficiency virus type 1, Zaire 2 isolate) may be subdivided into different regions according to their homology with Tat proteins from other lentiviruses (Derse et al., 1991; Dorn et al., 1990), the cysteine-rich region, Thr20 to Cys31, comprising five cysteine residues, the core region, Tyr/Phe32 to Tyr47, the basic region, Arg49 to Arg/Lys57, and the glutamine-rich region, Gln60 to Gln76: MDPVDPNIEP WNHPGSQPKT ACNRCHCKKC CYHCQVCFIT KGLGISYGRK KRRQRRRPSQ GGQTHQDPIP KQPSSQPRGD PTGPKE. Basic and core regions are highly conserved among all known Tat proteins, whereas a cysteine-rich region is not present, for example, in the EIAV Tat protein (Dorn et al., 1990; Noiman et al., 1991). The Abbreviations used: HIV-1, human immunodeficiency virus type 1; Tat, trans-activator; EIAV,equine infections anemia virus; TAR, trans-activation response element; ARM, arginine-rich motif; MD, molecular dynamics; 2D, two-dimensional; SA, simulated annealing; NOESY, nuclear Overhauser enhancement spectroscopy. 0022-2836/95/140529-07 $08.00/0
glutamine-rich region is severely truncated in EIAV Tat protein as compared to Tat proteins from primate immunodeficiency viruses. The core sequence region corresponds to a well defined hydrophobic structural domain in EIAV Tat protein, and the basic sequence region has a tendency to form a helix type structure in this protein (Sticht et al., 1993; Willbold etal., 1993). The basic region of HIV-Tat protein is suggested to form an c~-hehx from NMR studies of a chimeric EIAV Tat (core region)/HIV-1 Tat (basic region) peptide (Mujeeb et al., 1994). The basic region is involved in RNA (TAR, trans-activation response element) binding (Karn & Graeble, 1992), and Tat proteins thus belong to the family of arginine-rich motif (ARM) RNA binding proteins (Burd & Dreyfuss, 1994). The core region of HIV-1 Tat protein is additionally required for sequence-specific TAR binding (Churcher et al., 1993). The role of the cysteine-rich region is disputed: early reports suggesting that it is a zinc-dependent dimerization domain (Frankel et al., 1989) are not universally acknowledged (Churcher et al., 1993), and the nature and role of metal ion binding to the HIV-1 Tat protein cysteine-rich region is unclear (Jeyapaul et al., 1990). © 1995 Academic Press Limited
530
Recent reports take into doubt the requirement of metal ion binding to the cysteine-rich region for trans-activation activity altogether (Koken et al., 1994). The glutamine-rich region is probably involved in TAR binding (Churcher et al., 1993). H V I Z 2 Tat protein is available in milligram quantities as a transcriptionally active protein from chemical synthesis (Kraft et al., 1994). 2D-NMR spectroscopy (Ernst, 1992) and MD calculations with a modified simulated annealing (SA) protocol (Br/.inger, 1993) were applied to determine the structure of H V I Z 2 Tat protein in solution. The NOESY crosspeak pattern derived from spectra as in Figure I and shown in Figure 2 does not indicate the presence of stable elements of regular secondary structure. Molecular dynamics calculations based on the experimental parameters
Communication
showed that H V I Z 2 Tat protein consisted of three-dimensional domains of highly different flexibility corresponding to the sequence regions defined earlier (Dorn et al., 1990). This was evidenced b y the fact that all observed long range NOESY crosspeaks originated from either the glutamine-rich or the core sequence region. These regions were clearly defined as three-dimensional domains in the final structure as shown by a cartoon of the global fold (Figure 3) and the pairwise root mean square deviations (Table 1). The rigidity of the three-dimensional structure m a d e u p of the core region amino acids was reflected by the agreement of the final ten backbone structures, Phe38 to Tyr47 and Tyr32 to Tyr47, resulting from the MD calculations (Table 1). The complete core domains, Tyr32 to Tyr47, did not s u p e r p o s e well
Figure 1. a, 200 ms NOESY spectra obtained in H,_O; NOEs: 1: Ile45(SH), Ser46(HN); 2: Thr64(yH), Gln76(HN); 3: Thr82(TH), GIy83(HN); 4: Ile69(yH), Glu86(HN); 5: Leu43(SH), IIe45(HN). b, Superposition of 200ms NOESY spectra obtained in H20 and 2H20; NOEs: 6: Trp11(H6), Va136([3H); 7: His65(H4), VaI36(~H); 8: His65(H4), Glu63(I~H); 9: His65(H4), Val4(~H) or Glu60(~H) (ambiguous); 10: Trp11(H6), GIu63(J3H); 11: His65(H4), Glu72(~H); 12: Trp11(H6), Ile69(I3H) (ambiguous). c, 200 ms symmetrized NOESY spectrum obtained in 2H20; NOEs: 13: Ile45(SH), Ser46(c~H);14: Ue45(yH), Ser46(uH); 15: VaI4(TH), Ser46(c~H); 16: Thr64(yH), His65(c~H); 17: Ile39(TH), Gln66(c~H); 18: Ile45(SH), Gly44(~H); 19: Ile45(yCH3), Gly44(~H); 20: Ile45(TCH2), Gly44(c~H);21: Thr64(TH), Ile8(~H) or Glu86(c~H) (ambiguous); 22: Thr64(yH), His65(~H); 23: Ile45(TCH2), Gly44(~H). HIV-1 Tat protein, SwissProt data bank sequence entry HVIZ2, wild-type and Thr40Lys mutant, was chemically synthesized, purified, and refolded (Kraft et al., 1994); Slice et al. 1992): 10 ml of a 1 mg/ml solution of Tat protein was unfolded in 1 litre buffer solution containing 20 mM potassium phosphate (pH 6.5), 100 raM, ZnCI2, 50 mM mannitol, 10 mM ascorbic acid, 0.5 mM phenylmethylsulphonyl fluoride (PMSF), 0.5 mM dithiothreitol (DTT) and 6 M urea. Stepwise refolding was performed in the buffer used for the unfolding procedure, but without ZnC12, with subsequent concentrations of 4 M, 2 M, and 0 M urea. The protein was finally dialysed twice against 10 mM ammonium acetate. Helimn degassed buffers were used throughout. The protein was lyophilized twice after either procedure to remove ammonium acetate and stored as a lyophilized powder. 110 ml H20, 30 mM NaC1, 0.02% iv/v) NaN3 was degassed for 2 minutes, pressure 1 mbar. The buffer was then bubbled with N2 to saturation and kept under N2 atmosphere. 20 ~M Na2S~O4, was added; the final pH was approximately 5.5. The lyophilized protein was dissolved rapidly in 900 ~1 buffer solution, still under N2 atmosphere. The sample was then concentrated in a Speedvac vacuum concentrator to approximately 450 ~1. An NMR sample tube was flushed with N2 for approximately 10 min. The sample was then centrifuged and filled into the NMR tube after pH was determined to be approximately 6.3. N2 was passed over the sample in the NMR tube and the sample tube sealed. Protein concentrations of up to 2.89 mM without precipitation could be obtained with this procedure. Precipitation of HVIZ2 Tat protein was observed in dialyses in the presence of DTT in excess of 5 mM concentration, and increase in molecular mass of the order of several DTT molecules was detected by matrix-assisted laser desorption ionization--time of flight (MALDI-TOF) mass spectrometry under these conditions (Kraft et al., 1994). HIV-1 Tat protein is inactivated by DTT (Koken et al. 1994), in accord with these results. Thus, Na2S204 was used as reducing agent. Resonance assignments were made using two-dimensional double quantum filtered correlated spectroscopy (DQF-COSY), nuclear Overhauser enhancement spectroscopy (NOESY), and total coherence spectroscopy with suppression of NOESY type cross peaks (CLEAN-TOCSY) with the published standard procedures (Wtithrich, 1987). The following sets of spectra were recorded on a Bruker AMX600 NMR; HVIZ2 wt; COSY-DQF, CLEAN-TOCSY, mixing time 80 ms, NOESY, mixing time 200 and 250 ms, respectivel~ temperature, 298 K, 1.8 mM protein concentration, 50 mM NaC1, 0.04% (v/v) NAN3, in H20/2H20 (9:1), 1 mM dithioerythriol (DTT), (pH 6.3 to 6.5). HVIZ2 Thr40Lys mutant: DQF-COSY, CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 200 ms, temperature, 278 K, 2.0 mM protein concentration, 50 mM NaC1, 50 ~M Na2S204 and 0.04% (v/v) NAN3, in H20/2H20 (9.1), pH 6.5; CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 200 ms, temperature, 293 K, 2.2 mM protein concentration, 50 nM NaCI, 50 laM Na2S204, 0.04% (v/v) NaN3 in 2H,.O; NOESY, mixing time 200 ms, DQF-COSY, conditions as before, but pH 6.5; CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 100, 200 and 250 ms, respectivel)4 temperature, 298 K, 2.8 mM protein concentration, 50 mM NaCI, 50 mM Na2S,O4, 0.04% NaN3 in H20/2H20 (9:1), pH 6.0; NOESY, mixing time 200 ms, as before, but in 2H20. Spectra in the Figures are all from the HVIZ2 Thr40Lys mutant. Typical spectral parameters w.ere: frequency width 6024 Hz, time-domain data-size 8 k x 0.5 k or 4 k x 0.5 k data points, frequency domain data size 8 k x I k and 4 k x 1 k data points; sinebell squared filter with ~/4 phase shift, NMR data were evaluated using the NDee program package on X-window workstations (Herrmann et al., unpublished). 352 intraresidual ([i-j[ = 0), 185 sequential ([i-j[ = 1), 61 medium range ([i-j[ = 2,3,4,5), and 25 long range (li-jI) > 5) NOESY crosspeaks could be extracted from these spectra. Distance information was extracted from NOESY spectra with mixing times of 100, 200 and 250 ms. Sequential and medium range ([i-j[ ~< 5) NOEs were extracted from the 100 and 200 ms NOESY spectra to minimize spin diffusion effects. Long range NOEs ([i-j[ > 5) were extracted from all 3 NOESY spectra as well as from NOESY spectra recorded in 2H20.
531
Communication
0
_1.0
2
3 _ 1.2
8!3
8!1
ppm
_~.9~..14= 1~-~'8 8,15~~7 68~]~t~-o o~ol~_ o. 9
~
2~
_2.1
_ 2.2
0
7.~5
o D~
. [0
6
-
~
0
~)l
(~
0 o
.
4!5
4!3
.
4!1
3!9
ppm
Figure 1
because a flexible hinge, Cys37-Phe38, connected two structurally rigid parts of the core. This hinge is not present in the EIAV Tat protein structure, as a disulphide bond between Cys37 and Cys40 stabilizes the local structure of this protein. A pivotal amino acid in the turn-like core domain structure was Gly44, a residue highly conserved in all known Tat proteins. EIAV Tat protein and HVIZ2 Tat protein are similar in their core domains, Phe41 to Tyr49 and Phe38 to Tyr47, respectivel}~ which can be nearly ideally superposed onto each other (Figure 4), in spite of the fact that HVIZ2 Tat protein contains one additional amino acid, Thr40, in this sequence
region. This strengthens the view that integrity of the core domain is crucial for activity of Tat proteins, and all conclusions arrived at for the EIAV Tat protein concerning the core domain are equally valid for the HVIZ2 Tat protein. For example, specific R N A binding is only possible with a peptide containing the turn region around Gly44 (Churcher et al., 1993). Lys41, suspected to interact with transcription factor I I D (Kashanchi et al., 1994), is solvent-exposed and stabilized as part of the hydrophobic core. From its general location, Lys41 could well contribute to RNA binding as suggested recently (Churcher et al., 1993). The glutamine-rich sequence region, which
532
Communication
formed the second rigid domain in HVIZ2 Tat protein, has no full length counterpart in EIAV Tat protein. As summarized in a cartoon (Figure 3), core domain and glutamine-rich domain both show NOESY crosspeaks to a stretch of amino acids close to the NH2 terminus. Val4 to Trp11 were thus sandwiched between core and glutamine-rich domains. This relates to observations that HIV-1 Tat protein is inactive in amino acid 2 to 6 (Asp2 to Pro6 in HV1Z2) deletion mutants (Kuppuswamy et al., 1989). The COOH terminus is also fixed to the glutamine-rich domain. This leaves the stretch of amino acids Pro14 to Lys19, the cysteine-rich sequence region, and the basic sequence region highly flexible. Overall, in the case of HV1Z2 Tat protein, it seems to be justified to speak of a stable molecular centre made up of core sequence region, glutamine-rich sequence region, and NH2 terminus. For further studies of Tat and TAR interactions by fluorescence spectroscop~ Trpll is nearly ideally located in the centre of the molecule, sandwiched between the core and basic domains (Figure 3). This should make Trp11 a sensitive probe for structural changes imposed on the molecule for example by TAR interaction. The Arg78-Gly79-Asp80 (RGD loop), supposed to be a general key site for adhesive recognition and receptor interaction (Hynes & Lander, 1992), is also clearly solvent-exposed at the tip of a hairpin structure, which is experimentally well defined by several NOESY crosspeaks (Ser74-Gly83; Set75-
Glu86; Gln66-Thr82; His65-Asp80; Thr64-Glu86; Gly62-Thr82) as reflected in the low r.m.s.d, value of this loop (Table 1). Although RGD loop sequences seem to be very flexible in most proteins studied so far (Adler et al., 1991; Klaus et al., 1993; Saudek et al., 1991), the recently determined structure of decorsin (Krezel et al., 1994) shows a rigid RGD loop similar to that found for HV1Z2 Tat protein. The rigidity of the HV1Z2 Tat protein RGD loop structure may well be caused by two proline residues, Pro77 and Pro81, flanking the loop as proline residues have only a very small qb torsion angle space. Pro81 is highly conserved in all known HIV-1 Tat sequences, and Pro77 is only conservatively replaced by Ser77 in some sequences. A similar way of stabilization is suggested for decorsin, where one Pro residue directly flanks the RGD loop, whereas the other proline residue is spaced from the loop by two residues (Krezel et al., 1994). The question as to the conformation of the domain made up of the basic amino acid region in HVIZ2 Tat protein has to go partially unanswered. On one hand, in our experiments this basic region did not show any helix-type NOESY cross peaks, which can be observed, although weakl~ for the EIAV Tat protein basic region (Willbold et al., 1994). On the other hand, we would not have observed these cross peaks under the present experimental conditions for HV1Z2 Tat protein (2.8 mM concentration) if they were as weak as they are in the EIAV Tat protein spectra (8 mM concentration). Our results thus suggested that the basic domain is not a rigid helix in full length HVIZ2 Tat protein, contrasting the
~ PVDPNIEPWNHPGSQPKTACNRCHCKKCCYHCQVCFITKGLGI SYGRKKRRQRRRPSQGGQTHQD PI PKQPS SQPRGD PTGPE~
101
dNN(i,i+l) d~(i,i+l)
_
201
301
401
~
--
=
,
501
i
~ a X
601
~ X
m
701
--
.
~
801
~
m
dpN(i,i+l) d~(i,i+l) dNN(i,i+2) d~N(i,i+2) d~(i,i+2)
m
m m
dl~N(i,i+3) d~(i,i+3) dpx(i,i+4) Figure 2. Intraresidue and medium range (amino acids i and j with li-jl < 4) NOESY connectivities involving backbone protons. The height of the bars symbolizes qualitatively the relative crosspeak strength (weak, medium, strong).
533
Communication
Core
H
Gln
Figure 3. Cartoon of the global fold of HVIZ2 Tat protein. The 2 spheres symbolize rigid substructures, that is the core and the glutamine-rich domains. Several of the observed long range NOE connectivities are indicated. Colour code: blue, NH2 terminus and COOH terminus; yellow, cysteine-rich domain; light blue, core domain; pink, basic domain; green, glutamine-rich domain; red-orange, RGD loop. MD calculations based on the experimental NOESY data were performed with the XPLOR 3.1 program package basically with the standard hybrid distance geometry/simulated annealing (dgsa) protocol. NOE cross peaks were grouped according to their intensity into 3 categories: strong (0.18 to 0.28 nm), medium (0.18 to 0.40 rim) and weak (0.18 to 0.55 nm). The structures were calculated from the NMR data according to the standard X-PLOR distance geometry and refinement protocols with minor modifications. After bound smoothing, embedding and regularization a family of 30 substructures was produced. Full embedding and subsequent SA refinement with 3.75 ps heating to 1000 K and a square NOE potential was used. MD calculation (12.5 ps) at high temperature with increased weight on geometry and additional cooling in 50 K steps down on 100 K were followed by 300 steps of energy minimization. After another round of refinement with inclusion of electrostatic interactions (¢ = 1) and 600 steps of energy minimization, 10 structures were selected on the criteria of smallest NOE violations as well as energy and r.m.s.d, values. notion of a stable u-helix observed for the chimeric EIAV Tat (core region)/HIV-1 Tat (basic region) peptide (Mujeeb et al., 1994). Induction of u-helical
Table 1 Structural Statistics HIV-1 Tat protein R.m.s.d. from ideal geometry Angles (deg) Bonds (nm) Impropers (deg) NOE (nm)
1.542 0.0013 1.092 0.015
Average energies (kJ/mol) ENOE EvDw E,~.l
1610 -3956 -2117 R.m.s.d. protein backbone (nm)
Phe38 to Tyr47 Tyr32 to Tyr47 Gln63 to Gin72 Pro76 to Pro80 Whole protein
0.08 0.17 0.15 0.07 0.42
conformation in the basic d o m a i n b y addition of trifluoroethanol was possible for EIAV Tat protein (Sticht et al., 1993) but not for H V I Z 2 Tat protein (data not shown). So far there is no conclusive evidence that the cysteine-rich region contributes to the transcriptional activity of H V I Z 2 Tat protein. A different role for the cysteine-rich region may be suggested b y the fact that H V I Z 2 Tat protein, but not EIAV Tat protein, contains the adhesive RGD loop, and that RGD proteins, such as proteins of the disintegrin famil3~ contain six conserved cysteine residues in a total of about 40 amino acids (Klaus et al., 1993; Krezel et al., 1994). Thus, the H V I Z 2 Tat protein cysteine-rich region may be required for cell adhesion rather than TAR recognition, which would clarify m a n y of the recent ambiguities concerning this sequence region. The network of intramolecular disulphide bonds postulated for RGD proteins was not observed for the H V I Z 2 Tat. This could be an experimental artifact caused b y the reducing conditions u n d e r which w e
Communication
534
References
o
.
*
,
e °
• •
.
Q
Figure 4. Superposition of part of the core domain backbone of EIAV Tat protein (green), Phe41 to Tyr49, and HVIZ2 Tat protein (orange), Phe38 to Tyr47. Lys41 in the HVIZ2 Tat protein is coloured blue.
had to keep the protein to prevent its aggregation under NMR concentrations. It is interesting to note that a region of high homology to the Cys-rich region in HIV Tat proteins is found in the proto-oncogenic Wnt5a gene products (Clark et al., 1993), the vertebrate relatives of the developmental Drosophila melanogaster wingless gene product: HVIZ2 Wnt5A (mouse)
22 37 - CNRCHCK -- KCCYH-CQVC -TERCHCKFHWCCYVKCKKC351 369
Wnt proteins are s u p p o s e d to be signalling molecules acting over a short distance. It is now possible, with determination of basic features of the EIAV Tat protein and the H V I Z 2 Tat protein structures, to address more specific questions related to the nature of the Tat/TAR interactions b y spectroscopic and biochemical techniques. For a full understanding, however, knowledge of the structure of the Tat/TAR complexes is a necessary prerequisite.
Acknowledgements F. Herrmann helped with the NDee-program and art work. M. Churcher, J. Karn, T. Parslow, G. Varani, and T. James made their results accessible before publication. Financial support was granted by the Fonds der Chemischen Industrie to P. B. and E R., by the Deutsche Forschungsgemeinschaft and the European Union BIOMED 1-AIDS RESEARCH program to P. R. and R. F. The NDee program is available on request. Co-ordinates were deposited in the Brookhaven Protein Data Bank and are available on request. The accession number is not available at the time of publication.
Adler, M., Lazarus, R. A., Dennis, M. S. & Wagner, G. (1991). Solution structure of kistrin, a potent platelet aggregation inhibitor and GP IIb-IIIa antagonist. Science, 253, 445-8. Briinger, A. (1993). X-PLOR 3.1 Manual. Yale University Press, New Haven. Burd, C. & Dreyfuss, G. (1994). Conserved structures and diversity of functions of RNA-binding proteins. Science, 265, 615-621. Churcher, M. J., Lamont, C., Hamy, F., Dingwall, C., Green, S. M., Lowe, A. D., Butler, J. G., Gait, M. J. & Karn, J. (1993). High affinity binding of TAR RNA by the human immunodeficiency virus type-1 tat protein requires base-pairs in the RNA stem and amino acid residues flanking the basic region. J. Mol. Biol. 230, 90-110. Clark, C. C., Cohen, I., Eichstetter, I., Cannizzaro, L. A., McPherson, J. D., Wasmuth, J. J. & Iozzo, R. W. (1993). Molecular cloning of the human proto-oncogene Wnt-5A and mapping of the gene (WNT5A) to chromosome 3p14-p21. Genomics, 18, 249-60. Derse, D., Carvalho, M., Carroll, R. & Peterlin, B. M. (1991). A minimal lentivirus Tat. J. Virol. 65, 7012-5. Dorn, P., DaSilva, L., Martarano, L. & Derse, D. (1990). Equine infectious anemia virus tat: Insights into the structure, function, and evolution of lentivirus trans-activator proteins. J. Virol. 64, 1616-1624. Ernst, R. R. (1992). Nuclear magnetic resonance fourier transforms spectroscop~ Angew. Chemie., Int. Ed. Engl. 31, 805-823. Frankel, A. D., Biancalana, S. & Hudson, D. (1989). Activity of synthetic peptides from the Tat protein of human immunodeficiency virus type 1. Proc. Nat. Acad. Sci., U.S.A. 86, 7397-401. Hynes, R. O. & Lander, A. D. (1992). Contact and adhesive specificities in the associations, migrations, and targeting of cells and axons. Cell, 68, 303-22. Jeyapaul, J., Reddy, M. R. & Khan, S. A. (1990). Activity of synthetic tat peptides in human immunodeficiency virus type 1 long terminal repeat-promoted transcription in a cell-free system. Proc. Nat. Acad. Sci., U.S.A. 87, 7030-4. Karn, J. & Graeble, M. A. (1992). New insights into the mechanism of HIV-1 trans-activation. Trends Genet. 8, 365-8. Kashanchi, F., Piras, G., Radonovich, M. F., Duvall, J. F., Fattaey, A., Chiang, C. M., Roeder, R. G. & Brady, J. N. (1994). Direct interaction of human TFIID with the HIV-1 transactivator tat. Nature (London), 367, 295-9. Klaus, W., Broger, C., Gerber P. & Senn, H. (1993). Determination of the disulphide bonding pattern in proteins by local and global analysis of nuclear magnetic resonance data. Application to flavoridin. J. Mol. Biol. 232, 897-906. Koken, S. E., Greijer, A. E., Verhoef, K., van-Wamel, J., Bukrinskaya, A. G. & Berkhout, B. (1994). Intracellular analysis of in vitro modified HIV Tat protein. J. Biol. Chem. 269, 8366-75. Kraft, M., Westendorp, M., Krammer, P., Bayer, P., R6sch, P. & Frank, R. W. (1994)..In Peptides 1994 (Maja, H. L. S., ed.), Escom, Leiden. Krezel, A. M., Wagner, G., Seymour-Ulmer, J. & Lazarus, R. A. (1994). Structure of the RGD protein decorsin: conserved motif and distinct function in leech proteins that affect blood clotting. Science, 264, 1994--1947. Kuppuswamy, M., Subramanian, T., Srinivasan, A. & Chinnadurai, G. (1989). Multiple functional domains of
Communication
Tat, the trans-activator of HIV-1, defined by mutational analysis. Nucl. Acids Res 17, 3551-61. Mujeeb, A., Bishop, K., Matija, E B., Turck, C., Parslow, T. G. & James, T. L. (1994). NMR structure of a biologically active peptide containing the RNA-binding domain of HW-1 Tat. Proc. Nat. Acad. Sci., U.S.A. 91, 8248-8252. Noiman, S., Yaniv, A., Tsach, T., Miki, T., Tronick, S. R. & Gazit, A. (1991). The Tat protein of equine infectious anemia virus is encoded by at least three types of transcripts. Virology, 184, 521-30. Saudek, V., Atkinson, R. A. & Petton, J. T. (1991). Three-dimensional structure of echistatin, the smallest active RGD protein. Biochemistry, 30, 7369-72. Slice, L. W., Codner, E., Antelman, D., Holl~ M., Wegrzynski, B., Wang, J., Toome, V., Hsu, M. C. & Nalin, C. M. (1992). Characterization of recombinant
535
HW-1 Tat and its interaction with TAR RNA. Biochemistry, 31, 12062-8. Sticht, H., Willbold, D., Bayer, P., Ejchart, A., Herrmann, F., Rosin-Arbesfeld, R., Gazit, A., Yaniv, A., Frank, R. & R6sch, P. (1993). Equine infectious anemia virus Tat is a predominantly helical protein. Eur. J. Biochem. 218, 973--6. Willbold, D., Kriiger, U., Frank, R., Rosin-Arbesfeld, R., Gazit, A., Yaniv, A. & R6sch, E (1993). Sequencespecific resonance assignments of the 1H-NMR spectra of a synthetic, biologically active EIAV Tat protein. Biochemistry, 32, 8439-8445. WiUbold, D., Rosin-Arbesfeld, R., Sticht, H., Frank, R. & R6sch, P. (1994). The structure of the equine infectious anemia virus Tat protein. Science, 264, 1584-1587. Wfithrich, K. (1987). NMR of Proteins and Nucleic Acids. Wile3~ New York.
Edited by P. E. Wright
(Received 11 October 1994; accepted 3 January 1995)
JMB COMMUNICATION
Structural Studies of HIV-1 Tat Protein R Bayer 1, M. Kraft 2, A. Ejchart 1, M. Westendorp 3, R. Frank 2 and R R6sch 1Lehrstuhl fiir Biopolymere und Bayreuther Institut ffir Makromolekfilforschung Universit~t Bayreuth D-95440 Bayreuth, Germany 2Zentrum fiir Molekularbiologie Heidelberg D-69120 Heidelberg Germany 3Deutsches Krebsforschungszentrum D-69120 Heidelberg Germany
*Corresponding author
Tat (trans-activator) proteins are early RNA binding proteins regulating lentiviral transcription. These proteins are necessary components in the life cycle of all known lentiviruses, such as the human immunodeficiency viruses (HIV) or the equine infectious anemia virus (EIAV). Tat proteins are thus ideal targets for drugs intervening with lentiviral growth. The consensus RNA binding motif (TAR, trans-activation responsive element) of HIV-1 is well characterized. Structural features of the 86 amino acid HIV-1, Zaire 2 isolate (HVIZ2) Tat protein in solution were determined by two dimensional (2D) nuclear magnetic resonance (NMR) methods and molecular dynamics (MD) calculations. In general, sequence regions corresponded to structural domains of the protein. It exhibited a hydrophobic core of 16 amino acids and a glutamine-rich domain of 17 amino acids. Part of the NH2 terminus, Val4 to Pro14, was sandwiched between these domains. Two highly flexible domains corresponded to a cysteine-rich and a basic sequence region. The 16 amino acid sequence of the core region is strictly conserved among the known Tat proteins, and the three-dimensional fold of these amino acids of HVIZ2 Tat protein was highly similar to the structure of the corresponding EIAV Tat domain. HVIZ2 Tat protein contained a well defined COOH-terminal Arg-Gly-Asp (RGD) loop similar to the recently determined decorsin RGD loop.
Keywords: tat protein; solution structure; HIV; NMR; molecular dynamics
The sequence of the 86 amino acid Tat protein from HVIZ2 (human immunodeficiency virus type 1, Zaire 2 isolate) may be subdivided into different regions according to their homology with Tat proteins from other lentiviruses (Derse et al., 1991; Dorn et al., 1990), the cysteine-rich region, Thr20 to Cys31, comprising five cysteine residues, the core region, Tyr/Phe32 to Tyr47, the basic region, Arg49 to Arg/Lys57, and the glutamine-rich region, Gln60 to Gln76: MDPVDPNIEP WNHPGSQPKT ACNRCHCKKC CYHCQVCFIT KGLGISYGRK KRRQRRRPSQ GGQTHQDPIP KQPSSQPRGD PTGPKE. Basic and core regions are highly conserved among all known Tat proteins, whereas a cysteine-rich region is not present, for example, in the EIAV Tat protein (Dorn et al., 1990; Noiman et al., 1991). The Abbreviations used: HIV-1, human immunodeficiency virus type 1; Tat, trans-activator; EIAV,equine infections anemia virus; TAR, trans-activation response element; ARM, arginine-rich motif; MD, molecular dynamics; 2D, two-dimensional; SA, simulated annealing; NOESY, nuclear Overhauser enhancement spectroscopy. 0022-2836/95/140529-07 $08.00/0
glutamine-rich region is severely truncated in EIAV Tat protein as compared to Tat proteins from primate immunodeficiency viruses. The core sequence region corresponds to a well defined hydrophobic structural domain in EIAV Tat protein, and the basic sequence region has a tendency to form a helix type structure in this protein (Sticht et al., 1993; Willbold etal., 1993). The basic region of HIV-Tat protein is suggested to form an c~-hehx from NMR studies of a chimeric EIAV Tat (core region)/HIV-1 Tat (basic region) peptide (Mujeeb et al., 1994). The basic region is involved in RNA (TAR, trans-activation response element) binding (Karn & Graeble, 1992), and Tat proteins thus belong to the family of arginine-rich motif (ARM) RNA binding proteins (Burd & Dreyfuss, 1994). The core region of HIV-1 Tat protein is additionally required for sequence-specific TAR binding (Churcher et al., 1993). The role of the cysteine-rich region is disputed: early reports suggesting that it is a zinc-dependent dimerization domain (Frankel et al., 1989) are not universally acknowledged (Churcher et al., 1993), and the nature and role of metal ion binding to the HIV-1 Tat protein cysteine-rich region is unclear (Jeyapaul et al., 1990). © 1995 Academic Press Limited
530
Recent reports take into doubt the requirement of metal ion binding to the cysteine-rich region for trans-activation activity altogether (Koken et al., 1994). The glutamine-rich region is probably involved in TAR binding (Churcher et al., 1993). H V I Z 2 Tat protein is available in milligram quantities as a transcriptionally active protein from chemical synthesis (Kraft et al., 1994). 2D-NMR spectroscopy (Ernst, 1992) and MD calculations with a modified simulated annealing (SA) protocol (Br/.inger, 1993) were applied to determine the structure of H V I Z 2 Tat protein in solution. The NOESY crosspeak pattern derived from spectra as in Figure I and shown in Figure 2 does not indicate the presence of stable elements of regular secondary structure. Molecular dynamics calculations based on the experimental parameters
Communication
showed that H V I Z 2 Tat protein consisted of three-dimensional domains of highly different flexibility corresponding to the sequence regions defined earlier (Dorn et al., 1990). This was evidenced b y the fact that all observed long range NOESY crosspeaks originated from either the glutamine-rich or the core sequence region. These regions were clearly defined as three-dimensional domains in the final structure as shown by a cartoon of the global fold (Figure 3) and the pairwise root mean square deviations (Table 1). The rigidity of the three-dimensional structure m a d e u p of the core region amino acids was reflected by the agreement of the final ten backbone structures, Phe38 to Tyr47 and Tyr32 to Tyr47, resulting from the MD calculations (Table 1). The complete core domains, Tyr32 to Tyr47, did not s u p e r p o s e well
Figure 1. a, 200 ms NOESY spectra obtained in H,_O; NOEs: 1: Ile45(SH), Ser46(HN); 2: Thr64(yH), Gln76(HN); 3: Thr82(TH), GIy83(HN); 4: Ile69(yH), Glu86(HN); 5: Leu43(SH), IIe45(HN). b, Superposition of 200ms NOESY spectra obtained in H20 and 2H20; NOEs: 6: Trp11(H6), Va136([3H); 7: His65(H4), VaI36(~H); 8: His65(H4), Glu63(I~H); 9: His65(H4), Val4(~H) or Glu60(~H) (ambiguous); 10: Trp11(H6), GIu63(J3H); 11: His65(H4), Glu72(~H); 12: Trp11(H6), Ile69(I3H) (ambiguous). c, 200 ms symmetrized NOESY spectrum obtained in 2H20; NOEs: 13: Ile45(SH), Ser46(c~H);14: Ue45(yH), Ser46(uH); 15: VaI4(TH), Ser46(c~H); 16: Thr64(yH), His65(c~H); 17: Ile39(TH), Gln66(c~H); 18: Ile45(SH), Gly44(~H); 19: Ile45(yCH3), Gly44(~H); 20: Ile45(TCH2), Gly44(c~H);21: Thr64(TH), Ile8(~H) or Glu86(c~H) (ambiguous); 22: Thr64(yH), His65(~H); 23: Ile45(TCH2), Gly44(~H). HIV-1 Tat protein, SwissProt data bank sequence entry HVIZ2, wild-type and Thr40Lys mutant, was chemically synthesized, purified, and refolded (Kraft et al., 1994); Slice et al. 1992): 10 ml of a 1 mg/ml solution of Tat protein was unfolded in 1 litre buffer solution containing 20 mM potassium phosphate (pH 6.5), 100 raM, ZnCI2, 50 mM mannitol, 10 mM ascorbic acid, 0.5 mM phenylmethylsulphonyl fluoride (PMSF), 0.5 mM dithiothreitol (DTT) and 6 M urea. Stepwise refolding was performed in the buffer used for the unfolding procedure, but without ZnC12, with subsequent concentrations of 4 M, 2 M, and 0 M urea. The protein was finally dialysed twice against 10 mM ammonium acetate. Helimn degassed buffers were used throughout. The protein was lyophilized twice after either procedure to remove ammonium acetate and stored as a lyophilized powder. 110 ml H20, 30 mM NaC1, 0.02% iv/v) NaN3 was degassed for 2 minutes, pressure 1 mbar. The buffer was then bubbled with N2 to saturation and kept under N2 atmosphere. 20 ~M Na2S~O4, was added; the final pH was approximately 5.5. The lyophilized protein was dissolved rapidly in 900 ~1 buffer solution, still under N2 atmosphere. The sample was then concentrated in a Speedvac vacuum concentrator to approximately 450 ~1. An NMR sample tube was flushed with N2 for approximately 10 min. The sample was then centrifuged and filled into the NMR tube after pH was determined to be approximately 6.3. N2 was passed over the sample in the NMR tube and the sample tube sealed. Protein concentrations of up to 2.89 mM without precipitation could be obtained with this procedure. Precipitation of HVIZ2 Tat protein was observed in dialyses in the presence of DTT in excess of 5 mM concentration, and increase in molecular mass of the order of several DTT molecules was detected by matrix-assisted laser desorption ionization--time of flight (MALDI-TOF) mass spectrometry under these conditions (Kraft et al., 1994). HIV-1 Tat protein is inactivated by DTT (Koken et al. 1994), in accord with these results. Thus, Na2S204 was used as reducing agent. Resonance assignments were made using two-dimensional double quantum filtered correlated spectroscopy (DQF-COSY), nuclear Overhauser enhancement spectroscopy (NOESY), and total coherence spectroscopy with suppression of NOESY type cross peaks (CLEAN-TOCSY) with the published standard procedures (Wtithrich, 1987). The following sets of spectra were recorded on a Bruker AMX600 NMR; HVIZ2 wt; COSY-DQF, CLEAN-TOCSY, mixing time 80 ms, NOESY, mixing time 200 and 250 ms, respectivel~ temperature, 298 K, 1.8 mM protein concentration, 50 mM NaC1, 0.04% (v/v) NAN3, in H20/2H20 (9:1), 1 mM dithioerythriol (DTT), (pH 6.3 to 6.5). HVIZ2 Thr40Lys mutant: DQF-COSY, CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 200 ms, temperature, 278 K, 2.0 mM protein concentration, 50 mM NaC1, 50 ~M Na2S204 and 0.04% (v/v) NAN3, in H20/2H20 (9.1), pH 6.5; CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 200 ms, temperature, 293 K, 2.2 mM protein concentration, 50 nM NaCI, 50 laM Na2S204, 0.04% (v/v) NaN3 in 2H,.O; NOESY, mixing time 200 ms, DQF-COSY, conditions as before, but pH 6.5; CLEAN-TOCSY, mixing time 90 ms, NOESY, mixing time 100, 200 and 250 ms, respectivel)4 temperature, 298 K, 2.8 mM protein concentration, 50 mM NaCI, 50 mM Na2S,O4, 0.04% NaN3 in H20/2H20 (9:1), pH 6.0; NOESY, mixing time 200 ms, as before, but in 2H20. Spectra in the Figures are all from the HVIZ2 Thr40Lys mutant. Typical spectral parameters w.ere: frequency width 6024 Hz, time-domain data-size 8 k x 0.5 k or 4 k x 0.5 k data points, frequency domain data size 8 k x I k and 4 k x 1 k data points; sinebell squared filter with ~/4 phase shift, NMR data were evaluated using the NDee program package on X-window workstations (Herrmann et al., unpublished). 352 intraresidual ([i-j[ = 0), 185 sequential ([i-j[ = 1), 61 medium range ([i-j[ = 2,3,4,5), and 25 long range (li-jI) > 5) NOESY crosspeaks could be extracted from these spectra. Distance information was extracted from NOESY spectra with mixing times of 100, 200 and 250 ms. Sequential and medium range ([i-j[ ~< 5) NOEs were extracted from the 100 and 200 ms NOESY spectra to minimize spin diffusion effects. Long range NOEs ([i-j[ > 5) were extracted from all 3 NOESY spectra as well as from NOESY spectra recorded in 2H20.
531
Communication
0
_1.0
2
3 _ 1.2
8!3
8!1
ppm
_~.9~..14= 1~-~'8 8,15~~7 68~]~t~-o o~ol~_ o. 9
~
2~
_2.1
_ 2.2
0
7.~5
o D~
. [0
6
-
~
0
~)l
(~
0 o
.
4!5
4!3
.
4!1
3!9
ppm
Figure 1
because a flexible hinge, Cys37-Phe38, connected two structurally rigid parts of the core. This hinge is not present in the EIAV Tat protein structure, as a disulphide bond between Cys37 and Cys40 stabilizes the local structure of this protein. A pivotal amino acid in the turn-like core domain structure was Gly44, a residue highly conserved in all known Tat proteins. EIAV Tat protein and HVIZ2 Tat protein are similar in their core domains, Phe41 to Tyr49 and Phe38 to Tyr47, respectivel}~ which can be nearly ideally superposed onto each other (Figure 4), in spite of the fact that HVIZ2 Tat protein contains one additional amino acid, Thr40, in this sequence
region. This strengthens the view that integrity of the core domain is crucial for activity of Tat proteins, and all conclusions arrived at for the EIAV Tat protein concerning the core domain are equally valid for the HVIZ2 Tat protein. For example, specific R N A binding is only possible with a peptide containing the turn region around Gly44 (Churcher et al., 1993). Lys41, suspected to interact with transcription factor I I D (Kashanchi et al., 1994), is solvent-exposed and stabilized as part of the hydrophobic core. From its general location, Lys41 could well contribute to RNA binding as suggested recently (Churcher et al., 1993). The glutamine-rich sequence region, which
532
Communication
formed the second rigid domain in HVIZ2 Tat protein, has no full length counterpart in EIAV Tat protein. As summarized in a cartoon (Figure 3), core domain and glutamine-rich domain both show NOESY crosspeaks to a stretch of amino acids close to the NH2 terminus. Val4 to Trp11 were thus sandwiched between core and glutamine-rich domains. This relates to observations that HIV-1 Tat protein is inactive in amino acid 2 to 6 (Asp2 to Pro6 in HV1Z2) deletion mutants (Kuppuswamy et al., 1989). The COOH terminus is also fixed to the glutamine-rich domain. This leaves the stretch of amino acids Pro14 to Lys19, the cysteine-rich sequence region, and the basic sequence region highly flexible. Overall, in the case of HV1Z2 Tat protein, it seems to be justified to speak of a stable molecular centre made up of core sequence region, glutamine-rich sequence region, and NH2 terminus. For further studies of Tat and TAR interactions by fluorescence spectroscop~ Trpll is nearly ideally located in the centre of the molecule, sandwiched between the core and basic domains (Figure 3). This should make Trp11 a sensitive probe for structural changes imposed on the molecule for example by TAR interaction. The Arg78-Gly79-Asp80 (RGD loop), supposed to be a general key site for adhesive recognition and receptor interaction (Hynes & Lander, 1992), is also clearly solvent-exposed at the tip of a hairpin structure, which is experimentally well defined by several NOESY crosspeaks (Ser74-Gly83; Set75-
Glu86; Gln66-Thr82; His65-Asp80; Thr64-Glu86; Gly62-Thr82) as reflected in the low r.m.s.d, value of this loop (Table 1). Although RGD loop sequences seem to be very flexible in most proteins studied so far (Adler et al., 1991; Klaus et al., 1993; Saudek et al., 1991), the recently determined structure of decorsin (Krezel et al., 1994) shows a rigid RGD loop similar to that found for HV1Z2 Tat protein. The rigidity of the HV1Z2 Tat protein RGD loop structure may well be caused by two proline residues, Pro77 and Pro81, flanking the loop as proline residues have only a very small qb torsion angle space. Pro81 is highly conserved in all known HIV-1 Tat sequences, and Pro77 is only conservatively replaced by Ser77 in some sequences. A similar way of stabilization is suggested for decorsin, where one Pro residue directly flanks the RGD loop, whereas the other proline residue is spaced from the loop by two residues (Krezel et al., 1994). The question as to the conformation of the domain made up of the basic amino acid region in HVIZ2 Tat protein has to go partially unanswered. On one hand, in our experiments this basic region did not show any helix-type NOESY cross peaks, which can be observed, although weakl~ for the EIAV Tat protein basic region (Willbold et al., 1994). On the other hand, we would not have observed these cross peaks under the present experimental conditions for HV1Z2 Tat protein (2.8 mM concentration) if they were as weak as they are in the EIAV Tat protein spectra (8 mM concentration). Our results thus suggested that the basic domain is not a rigid helix in full length HVIZ2 Tat protein, contrasting the
~ PVDPNIEPWNHPGSQPKTACNRCHCKKCCYHCQVCFITKGLGI SYGRKKRRQRRRPSQGGQTHQD PI PKQPS SQPRGD PTGPE~
101
dNN(i,i+l) d~(i,i+l)
_
201
301
401
~
--
=
,
501
i
~ a X
601
~ X
m
701
--
.
~
801
~
m
dpN(i,i+l) d~(i,i+l) dNN(i,i+2) d~N(i,i+2) d~(i,i+2)
m
m m
dl~N(i,i+3) d~(i,i+3) dpx(i,i+4) Figure 2. Intraresidue and medium range (amino acids i and j with li-jl < 4) NOESY connectivities involving backbone protons. The height of the bars symbolizes qualitatively the relative crosspeak strength (weak, medium, strong).
533
Communication
Core
H
Gln
Figure 3. Cartoon of the global fold of HVIZ2 Tat protein. The 2 spheres symbolize rigid substructures, that is the core and the glutamine-rich domains. Several of the observed long range NOE connectivities are indicated. Colour code: blue, NH2 terminus and COOH terminus; yellow, cysteine-rich domain; light blue, core domain; pink, basic domain; green, glutamine-rich domain; red-orange, RGD loop. MD calculations based on the experimental NOESY data were performed with the XPLOR 3.1 program package basically with the standard hybrid distance geometry/simulated annealing (dgsa) protocol. NOE cross peaks were grouped according to their intensity into 3 categories: strong (0.18 to 0.28 nm), medium (0.18 to 0.40 rim) and weak (0.18 to 0.55 nm). The structures were calculated from the NMR data according to the standard X-PLOR distance geometry and refinement protocols with minor modifications. After bound smoothing, embedding and regularization a family of 30 substructures was produced. Full embedding and subsequent SA refinement with 3.75 ps heating to 1000 K and a square NOE potential was used. MD calculation (12.5 ps) at high temperature with increased weight on geometry and additional cooling in 50 K steps down on 100 K were followed by 300 steps of energy minimization. After another round of refinement with inclusion of electrostatic interactions (¢ = 1) and 600 steps of energy minimization, 10 structures were selected on the criteria of smallest NOE violations as well as energy and r.m.s.d, values. notion of a stable u-helix observed for the chimeric EIAV Tat (core region)/HIV-1 Tat (basic region) peptide (Mujeeb et al., 1994). Induction of u-helical
Table 1 Structural Statistics HIV-1 Tat protein R.m.s.d. from ideal geometry Angles (deg) Bonds (nm) Impropers (deg) NOE (nm)
1.542 0.0013 1.092 0.015
Average energies (kJ/mol) ENOE EvDw E,~.l
1610 -3956 -2117 R.m.s.d. protein backbone (nm)
Phe38 to Tyr47 Tyr32 to Tyr47 Gln63 to Gin72 Pro76 to Pro80 Whole protein
0.08 0.17 0.15 0.07 0.42
conformation in the basic d o m a i n b y addition of trifluoroethanol was possible for EIAV Tat protein (Sticht et al., 1993) but not for H V I Z 2 Tat protein (data not shown). So far there is no conclusive evidence that the cysteine-rich region contributes to the transcriptional activity of H V I Z 2 Tat protein. A different role for the cysteine-rich region may be suggested b y the fact that H V I Z 2 Tat protein, but not EIAV Tat protein, contains the adhesive RGD loop, and that RGD proteins, such as proteins of the disintegrin famil3~ contain six conserved cysteine residues in a total of about 40 amino acids (Klaus et al., 1993; Krezel et al., 1994). Thus, the H V I Z 2 Tat protein cysteine-rich region may be required for cell adhesion rather than TAR recognition, which would clarify m a n y of the recent ambiguities concerning this sequence region. The network of intramolecular disulphide bonds postulated for RGD proteins was not observed for the H V I Z 2 Tat. This could be an experimental artifact caused b y the reducing conditions u n d e r which w e
Communication
534
References
o
.
*
,
e °
• •
.
Q
Figure 4. Superposition of part of the core domain backbone of EIAV Tat protein (green), Phe41 to Tyr49, and HVIZ2 Tat protein (orange), Phe38 to Tyr47. Lys41 in the HVIZ2 Tat protein is coloured blue.
had to keep the protein to prevent its aggregation under NMR concentrations. It is interesting to note that a region of high homology to the Cys-rich region in HIV Tat proteins is found in the proto-oncogenic Wnt5a gene products (Clark et al., 1993), the vertebrate relatives of the developmental Drosophila melanogaster wingless gene product: HVIZ2 Wnt5A (mouse)
22 37 - CNRCHCK -- KCCYH-CQVC -TERCHCKFHWCCYVKCKKC351 369
Wnt proteins are s u p p o s e d to be signalling molecules acting over a short distance. It is now possible, with determination of basic features of the EIAV Tat protein and the H V I Z 2 Tat protein structures, to address more specific questions related to the nature of the Tat/TAR interactions b y spectroscopic and biochemical techniques. For a full understanding, however, knowledge of the structure of the Tat/TAR complexes is a necessary prerequisite.
Acknowledgements F. Herrmann helped with the NDee-program and art work. M. Churcher, J. Karn, T. Parslow, G. Varani, and T. James made their results accessible before publication. Financial support was granted by the Fonds der Chemischen Industrie to P. B. and E R., by the Deutsche Forschungsgemeinschaft and the European Union BIOMED 1-AIDS RESEARCH program to P. R. and R. F. The NDee program is available on request. Co-ordinates were deposited in the Brookhaven Protein Data Bank and are available on request. The accession number is not available at the time of publication.
Adler, M., Lazarus, R. A., Dennis, M. S. & Wagner, G. (1991). Solution structure of kistrin, a potent platelet aggregation inhibitor and GP IIb-IIIa antagonist. Science, 253, 445-8. Briinger, A. (1993). X-PLOR 3.1 Manual. Yale University Press, New Haven. Burd, C. & Dreyfuss, G. (1994). Conserved structures and diversity of functions of RNA-binding proteins. Science, 265, 615-621. Churcher, M. J., Lamont, C., Hamy, F., Dingwall, C., Green, S. M., Lowe, A. D., Butler, J. G., Gait, M. J. & Karn, J. (1993). High affinity binding of TAR RNA by the human immunodeficiency virus type-1 tat protein requires base-pairs in the RNA stem and amino acid residues flanking the basic region. J. Mol. Biol. 230, 90-110. Clark, C. C., Cohen, I., Eichstetter, I., Cannizzaro, L. A., McPherson, J. D., Wasmuth, J. J. & Iozzo, R. W. (1993). Molecular cloning of the human proto-oncogene Wnt-5A and mapping of the gene (WNT5A) to chromosome 3p14-p21. Genomics, 18, 249-60. Derse, D., Carvalho, M., Carroll, R. & Peterlin, B. M. (1991). A minimal lentivirus Tat. J. Virol. 65, 7012-5. Dorn, P., DaSilva, L., Martarano, L. & Derse, D. (1990). Equine infectious anemia virus tat: Insights into the structure, function, and evolution of lentivirus trans-activator proteins. J. Virol. 64, 1616-1624. Ernst, R. R. (1992). Nuclear magnetic resonance fourier transforms spectroscop~ Angew. Chemie., Int. Ed. Engl. 31, 805-823. Frankel, A. D., Biancalana, S. & Hudson, D. (1989). Activity of synthetic peptides from the Tat protein of human immunodeficiency virus type 1. Proc. Nat. Acad. Sci., U.S.A. 86, 7397-401. Hynes, R. O. & Lander, A. D. (1992). Contact and adhesive specificities in the associations, migrations, and targeting of cells and axons. Cell, 68, 303-22. Jeyapaul, J., Reddy, M. R. & Khan, S. A. (1990). Activity of synthetic tat peptides in human immunodeficiency virus type 1 long terminal repeat-promoted transcription in a cell-free system. Proc. Nat. Acad. Sci., U.S.A. 87, 7030-4. Karn, J. & Graeble, M. A. (1992). New insights into the mechanism of HIV-1 trans-activation. Trends Genet. 8, 365-8. Kashanchi, F., Piras, G., Radonovich, M. F., Duvall, J. F., Fattaey, A., Chiang, C. M., Roeder, R. G. & Brady, J. N. (1994). Direct interaction of human TFIID with the HIV-1 transactivator tat. Nature (London), 367, 295-9. Klaus, W., Broger, C., Gerber P. & Senn, H. (1993). Determination of the disulphide bonding pattern in proteins by local and global analysis of nuclear magnetic resonance data. Application to flavoridin. J. Mol. Biol. 232, 897-906. Koken, S. E., Greijer, A. E., Verhoef, K., van-Wamel, J., Bukrinskaya, A. G. & Berkhout, B. (1994). Intracellular analysis of in vitro modified HIV Tat protein. J. Biol. Chem. 269, 8366-75. Kraft, M., Westendorp, M., Krammer, P., Bayer, P., R6sch, P. & Frank, R. W. (1994)..In Peptides 1994 (Maja, H. L. S., ed.), Escom, Leiden. Krezel, A. M., Wagner, G., Seymour-Ulmer, J. & Lazarus, R. A. (1994). Structure of the RGD protein decorsin: conserved motif and distinct function in leech proteins that affect blood clotting. Science, 264, 1994--1947. Kuppuswamy, M., Subramanian, T., Srinivasan, A. & Chinnadurai, G. (1989). Multiple functional domains of
Communication
Tat, the trans-activator of HIV-1, defined by mutational analysis. Nucl. Acids Res 17, 3551-61. Mujeeb, A., Bishop, K., Matija, E B., Turck, C., Parslow, T. G. & James, T. L. (1994). NMR structure of a biologically active peptide containing the RNA-binding domain of HW-1 Tat. Proc. Nat. Acad. Sci., U.S.A. 91, 8248-8252. Noiman, S., Yaniv, A., Tsach, T., Miki, T., Tronick, S. R. & Gazit, A. (1991). The Tat protein of equine infectious anemia virus is encoded by at least three types of transcripts. Virology, 184, 521-30. Saudek, V., Atkinson, R. A. & Petton, J. T. (1991). Three-dimensional structure of echistatin, the smallest active RGD protein. Biochemistry, 30, 7369-72. Slice, L. W., Codner, E., Antelman, D., Holl~ M., Wegrzynski, B., Wang, J., Toome, V., Hsu, M. C. & Nalin, C. M. (1992). Characterization of recombinant
535
HW-1 Tat and its interaction with TAR RNA. Biochemistry, 31, 12062-8. Sticht, H., Willbold, D., Bayer, P., Ejchart, A., Herrmann, F., Rosin-Arbesfeld, R., Gazit, A., Yaniv, A., Frank, R. & R6sch, P. (1993). Equine infectious anemia virus Tat is a predominantly helical protein. Eur. J. Biochem. 218, 973--6. Willbold, D., Kriiger, U., Frank, R., Rosin-Arbesfeld, R., Gazit, A., Yaniv, A. & R6sch, E (1993). Sequencespecific resonance assignments of the 1H-NMR spectra of a synthetic, biologically active EIAV Tat protein. Biochemistry, 32, 8439-8445. WiUbold, D., Rosin-Arbesfeld, R., Sticht, H., Frank, R. & R6sch, P. (1994). The structure of the equine infectious anemia virus Tat protein. Science, 264, 1584-1587. Wfithrich, K. (1987). NMR of Proteins and Nucleic Acids. Wile3~ New York.
Edited by P. E. Wright
(Received 11 October 1994; accepted 3 January 1995)