The elongation of transcription of HIV RNA at the TAR (transactivation-response element) is highly regulated by positive and negative factors. The cellular negative transcription elongation factor NELF (negative elongation factor) was suggested to be involved in transcriptional regulation of HIV-1 (HIV type 1) by binding to the stem of the viral TAR RNA which is synthesized by cellular RNA polymerase II at the viral long terminal repeat. NELF is a heterotetrameric protein consisting of NELF A, B, C or the splice variant D, and E. In the present study, we determined the solution structure of the RRM (RNA-recognition motif) of the RNA-binding subunit NELF E and studied its interaction with the viral TAR RNA. Our results show that the separately expressed recombinant NELF E RRM has α-helical and β-strand elements adopting a βαββαβ fold and is able to bind to TAR RNA. Fluorescence equilibrium titrations with fluorescently labelled double- and single-stranded oligoribonucleotides representing the TAR RNA stem imply that NELF E RRM binds to the single-stranded TAR RNAs with Kd values in the low-micromolar range.

INTRODUCTION

After HIV-1 (HIV type 1) enters the host cell, the viral genomic RNA is reverse-transcribed into double-stranded DNA which is then integrated into the host genome. Subsequent synthesis of new viral RNA is tightly controlled by a complex interaction of viral and cellular proteins [1]. Transcription starts from the LTR (long terminal repeat) of the integrated proviral DNA by the cellular RNA Pol II (polymerase II). Once the initial sequence of 60 nucleotides of the viral RNA, the so-called TAR (transactivator-response element), is synthesized, a stable RNA stem–loop structure is formed which is present at the 5′ end of all viral transcripts [1]. At this step, several cellular transcription factors, including the cellular negative transcription elongation factor NELF (negative elongation factor) and DSIF [DRB (5,6-dichloro-1β-D-ribofuranosylbenzimidazole)-sensitivity-inducing factor], bind to the TAR element [25]. Elongation of transcription is possible only if the viral transactivator protein Tat (transactivator of transcription) recruits the cellular positive transcription elongation factor pTEFb (positive transcription elongation factor b) and binds to the bulge region of TAR [3,5,6]. Upon phosphorylation of RNA Pol II, NELF and DSIF by the kinase component CDK9 (cyclin-dependent kinase 9) of pTEFb, anti-termination occurs, leading to productive elongation of transcription [5]. NELF consists of four different subunits, namely NELF A, NELF B, alternatively spliced NELF C or D, and NELF E, also called RD because of its internal repeats of the amino acids arginine (R) and aspartic acid (D) [4,7]. NELF A exhibits sequence similarities to HDAg (hepatitis delta antigen) which binds to and activates RNA pol II [8]. NELF E contains an RRM (RNA-recognition motif). Furthermore, NELF E interacts with the NELF B subunit, probably via a leucine zipper motif. It has been shown that NELF E binds to various RNA elements [79]. Most importantly, isolated NELF E binds to the HIV TAR element in vitro [9]. Its TAR RNA-binding activity suggests that NELF E plays a role in the control of HIV transcription.

EXPERIMENTAL

Plasmid construct, expression and protein purification

A synthetically produced gene adapted to the Escherichia coli codon usage harbouring the RRM of NELF E and adjacent regions was cloned via the NdeI and BamHI restriction sites into the E. coli expression vector pET15b (Novagen). The soluble recombinant protein contained an N-terminal His6 tag (Figure 1).

Nucleotide and amino acid sequence of NELF E RRM

Figure 1
Nucleotide and amino acid sequence of NELF E RRM

The construct contains a His6 tag at the N-terminus, followed by a thrombin cleavage site indicated by the grey box. The amino acids defining the RRM are underlined, the amino acid sequence derived from NELF E is written in bold letters. The conserved ribonucleoprotein motifs RNP2 and RNP1 containing residues Tyr43 and Phe77 are marked by black boxes. Amino acid numbering is indicated on the right.

Figure 1
Nucleotide and amino acid sequence of NELF E RRM

The construct contains a His6 tag at the N-terminus, followed by a thrombin cleavage site indicated by the grey box. The amino acids defining the RRM are underlined, the amino acid sequence derived from NELF E is written in bold letters. The conserved ribonucleoprotein motifs RNP2 and RNP1 containing residues Tyr43 and Phe77 are marked by black boxes. Amino acid numbering is indicated on the right.

E. coli strain BL21(DE3) (Novagen) containing the recombinant plasmid was grown at 37 °C in LB (Luria–Bertani) medium containing ampicillin until a D600 of 1 was reached and then induced with 1 mM IPTG (isopropyl β-D-thiogalactoside). Cells were harvested 3–4 h after induction. For 15N- and 13C-labelling, cells were pelleted at 15 °C before induction and resuspended in ¼ vol. of M9 medium [10,11] supplemented with 15NH4Cl and 0.2% [13C]glucose as the sole nitrogen and carbon sources respectively [12]. After 1 h of shaking at 37 °C, induction was performed with 1 mM IPTG and cells were harvested after 3–4 h.

In order to purify NELF E RRM, bacterial cell pellets were lysed by sonication [6×45 s, 200 W; Labsonic U (B. Braun Biotech)] in 20 mM sodium phosphate buffer, pH 7.4, 0.5 M NaCl, 5 mM imidazole and 1 mM DTT (dithiothreitol). After centrifugation at 19100 g for 45 min, the supernatant was loaded on to a Ni2+-affinity column (His-trap chelating, GE Healthcare) and eluted by applying an imidazole step gradient. Peak fractions containing NELF E RRM were dialysed against 10 mM sodium phosphate buffer, pH 6.9, 250 mM NaCl and 1 mM DTT, and purified further on a heparin column (GE Healthcare) by an NaCl step gradient in the same buffer with up to 1 M NaCl. The eluted fractions containing NELF E RRM were dialysed against 10 mM sodium phosphate buffer, pH 6.9, 100 mM NaCl and 1 mM DTT, concentrated with Vivaspin concentrators (Vivascience, molecular-mass cut-off 5000 Da), divided into aliquots and then stored after shock-freezing at −80 °C.

In vitro transcription and purification of TAR RNA

Synthesis of TAR RNA was performed by in vitro transcription using T7 RNA polymerase, a T7 RNA primer oligodeoxyribonucleotide and a template oligodeoxyribonucleotide coding for the corresponding 59 nucleotides of the TAR sequence (5′-GCCUCUCUCUGGUUAGACCAGAUCUGAGCCUGGGAGCUCUCUGGCUAACUAGGGAAGGC-3′) plus the region complementary to the T7 primer. The TAR RNA was then purified by denaturing polyacrylamide/urea gel electrophoresis as described in [13].

NMR spectrometry

All NMR experiments were performed at 298 K on Bruker DRX600, Avance 700 and Avance 800 spectrometers equipped with standard inverted or cryogenically cooled 1H/13C/15N triple-resonance probes with pulsed-field gradient capabilities. In order to obtain sequential backbone and side-chain resonance assignments, standard double- and triple-resonance NMR experiments were recorded [14,15] with uniformly 15N- or 13C/15N-labelled recombinant NELF E RRM at a concentration of 0.6 mM. Distance restraints for structure calculation were derived from three-dimensional 13C- and 15N-edited NOESY-HSQC (heteronuclear single-quantum coherence) experiments with mixing times of 120 ms [16,17]. Dihedral angle restraints were derived from 3J(HN, Hα) scalar coupling constants determined from the intensity ratios of cross and diagonal peaks of the HNHA spectrum [18]. A series of New MEXICO (measurement of exchange rates in isotopically labelled compounds) [19] experiments with different mixing times were recorded for characterizing amide proton exchange. 1D (one-bound) (1HN,15N) RDCs (residual dipolar couplings) were determined by the IPAP (in-phase/antiphase) method [20] using a weakly aligned sample of uniformly 15N-labelled NELF E RRM in a mixture of penta(ethylene glycol)-monododecyl ether (C12E5), hexanol and water [molar ratio of C12E5/hexanol=0.95, 3% (w/v) C12E5/water] [21]. {1H}15N NOE (nuclear Overhauser effect) values were determined using the pulse sequence of Dayie and Wagner [22] with a relaxation delay of 6 s including the 3 s saturation period with 120° high-power pulses for the saturated subspectrum. Chemical-shift changes of NELF E RRM upon binding to TAR RNA were observed in 1H/15N-HSQC spectra after gradually adding TAR RNA to a sample containing 15N-labelled NELF E RRM.

Normalized chemical-shift changes were expressed as the weighted geometric average of 1HN- and 15N-chemical-shift changes for each residue:

 
formula

Normalized chemical shift changes larger than 0.04 p.p.m. were considered significant [23]. NMR data were processed using in-house written software and analysed with the NMR-View 5.2.2 program [24].

Structure calculation

Distance restraints for structure calculation were derived from 15N-NOESY-HSQC and 13C-NOESY-HSQC spectra. NOESY cross-peaks were classified according to their relative intensities and converted into distance restraints with upper limits of 3.0 Å (1 Å=0.1 nm) (strong), 4.0 Å (medium), 5.0 Å (weak) and 6.0 Å (very weak). For ambiguous distance restraints, the r−6 summation over all assigned possibilities defined the upper limit [25].

The raw scalar coupling constants were multiplied with a correction factor of 1.1 to take into account the different relaxation rates of in-phase and antiphase components [18]. Residues with scalar coupling constants below 6 Hz were restrained to dihedral angles between −80° and −40°, residues showing coupling constants above 8 Hz were restricted to dihedral angles of −160° to −80° [26]. Glycine residues were omitted, since they were not stereospecifically assigned and the coupling constants are likely to be affected by cross-relaxation [18].

Hydrogen bonds were included in the final structure calculation if the acceptor of a slowly exchanging amide proton characterized by a missing signal in a 150 ms New MEXICO experiment [19] could be identified from the results of preceding structure calculations. Thus a hydrogen bond was assumed if the distance between the carboxy oxygen and the amide proton was below 2.6 Å, and the angle of the amide proton, the amide nitrogen and the carboxy oxygen was less than 60° in all accepted structures. For each hydrogen bond, the distance between the amide proton and the acceptor was restrained to less than 2.3 Å, and the distance between the amide nitrogen and the acceptor was restrained to less than 3.3 Å [26]. All proline residues were considered to adopt the trans-conformation as strong HA(i)-HD(i+1) and HN(i) and HD(i+1) NOEs could be observed [27].

The structure calculations were performed with the program XPLOR 3.8.5.1 using a three-step simulated annealing protocol [28,29] with floating assignment of prochiral groups [30]. Initial conformational space sampling was carried out for 120 ps with a time step of 3 fs at a temperature of 2000 K, followed by a cooling period of 120 ps down to 1000 K, and 60 ps cooling to 100 K, both with a time step of 2 fs. A modified conformational database potential for backbone and side-chain dihedral angles was applied [31,32]. The proline angles were modified according to Neudecker et al. [33]. After simulated annealing, the structures were subjected to 1000 steps of Powell minimization [34], and the final 500 steps were minimized without conformational database potential.

In a first step, 200 structures were calculated (Table 1) using 1926 distance, 24 hydrogen bond and 32 dihedral angle restraints. The 40 structures with the lowest total energy were then refined using 55 1D (1HN,15N) RDCs with a harmonic potential [35]. Dipolar couplings of flexible residues showing a {1H} 15N NOE below 0.65 at 14.1 T were excluded from the calculations. The tensor components of the alignment were optimized with a grid search by varying the axial component Da and the rhombicity R in steps of 0.5 and 0.1 respectively. The initial values of Da and R were estimated from the distribution of the 1D (1HN,15N) [36], and a Molecular Dynamics run was performed for each pair of Da and R, yielding an axial component of 9.0 Hz, and a rhombicity of 0.4 for the energetically most-favourable combination of Da and R.

Table 1
Structural statistics

Epot, overall potential energy; Eimpr, potential energy of improper angle; Ecdih, potential energy for dihedral restraints; Esani, potential energy for RDC restraints.

Parameter Value 
Experimental restraints  
 Distance restraints  
  Total 1926 
  Intraresidual 468 
  Sequential 406 
  Medium-range 281 
  Long-range 623 
 Dihedral angles 32 
 Dipolar couplings 55 
 Hydrogen bonds (two restraints each) 24 
Molecular Dynamics statistics  
 Energies (kcal/mol)  
  Epot 14.1±1.4 
  Ebond 0.58±0.06 
  Eangle 6.7±0.6 
  Eimpr 2.3±0.2 
  Erepel 2.4±0.3 
  ENOE 1.3±0.5 
  Ecdih 0.03±0.03 
  Esani 0.7±0.2 
 RMSDs from ideal distances (Å)  
  Bond lengths 0.00068±0.00004 
  Distance restraints 0.0036±0.0007 
 RMSDs from ideal angles (°)  
  Bond angles 0.14±0.06 
  Dihedral angle restraints 0.15±0.15 
 RMSDs from dipolar couplings (Hz) 0.11±0.02 
Atomic co-ordinate precision (RMSD) (Å)  
 Backbone heavy atoms 0.28 (Gly39–Ala108
 Heavy atoms 0.64 (Gly39­–Ala108
Ramachandran plot statistics  
 Residues in  
  Most-favoured regions 91.0% 
  Allowed regions 9.0% 
Parameter Value 
Experimental restraints  
 Distance restraints  
  Total 1926 
  Intraresidual 468 
  Sequential 406 
  Medium-range 281 
  Long-range 623 
 Dihedral angles 32 
 Dipolar couplings 55 
 Hydrogen bonds (two restraints each) 24 
Molecular Dynamics statistics  
 Energies (kcal/mol)  
  Epot 14.1±1.4 
  Ebond 0.58±0.06 
  Eangle 6.7±0.6 
  Eimpr 2.3±0.2 
  Erepel 2.4±0.3 
  ENOE 1.3±0.5 
  Ecdih 0.03±0.03 
  Esani 0.7±0.2 
 RMSDs from ideal distances (Å)  
  Bond lengths 0.00068±0.00004 
  Distance restraints 0.0036±0.0007 
 RMSDs from ideal angles (°)  
  Bond angles 0.14±0.06 
  Dihedral angle restraints 0.15±0.15 
 RMSDs from dipolar couplings (Hz) 0.11±0.02 
Atomic co-ordinate precision (RMSD) (Å)  
 Backbone heavy atoms 0.28 (Gly39–Ala108
 Heavy atoms 0.64 (Gly39­–Ala108
Ramachandran plot statistics  
 Residues in  
  Most-favoured regions 91.0% 
  Allowed regions 9.0% 

The 20 structures showing the lowest values of the target function excluding the database potential were analysed further with X-PLOR 3.8.5.1, MOLMOL [38] and PROCHECK 3.5.4 [39,40]. The structural co-ordinates were deposited in the Protein Data Bank (PDB) under accession code 2BZ2.

Fluorescence equilibrium titrations

Measurements were performed using a Fluorolog spectrophotometer (HORIBA Jobin Yvon) in reaction buffer consisting of 10 mM sodium phosphate, pH 7.0, and 100 mM NaCl in a volume of 2 ml. The excitation wavelength was 495 nm, and the emission intensity was measured at 522 or 525 nm with slit widths set at 2 nm for both excitation and emission. To analyse the doublestranded TAR RNA stem, the oligoribonucleotide representing the 5′ end of HIV-1 TAR, TAR-(1–18) (5′-GGUCUCUCUGGUUAGACC-3′) was hybridized in reaction buffer to the complementary 5′-FAM (6-carboxyfluorescein)-labelled oligoribonucleotide (biomers.net) representing the 3′ end of the HIV-1 TAR RNA, TAR-(42–57) (5′-FAM-GGCUAACUAGGGAACC-3′). The solution was heated to 90 °C for 3 min, followed by slow cooling to room temperature. For analysis of the single strands, the corresponding 5′-FAM-labelled oligoribonucleotides were used. Double- or single-stranded substrate (50 nM) was titrated with increasing amounts of NELF E RRM. Values for the dissociation constant, Kd, were determined assuming a two-state model, using a quadratic equation for the fitting procedure:

 
formula

where Fobs is observed fluorescence, FL0 is starting fluorescence, [L0] is concentration of the RNA, ΔF0 is change in fluorescence, and [E0] is concentration of protein.

RESULTS AND DISCUSSION

Purification of NELF E RRM

To analyse the structure and function of NELF E RRM, a synthetic gene containing the corresponding DNA coding for an N-terminal His6 tag followed by a thrombin site and the NELF E RRM was constructed with the codon usage adapted to E. coli (Figure 1). The regions adjacent to the NELF E RRM coding region had to be introduced since plasmid constructs lacking these additional DNA stretches led to insoluble protein. The DNA sequence was deposited into GenBank® under the accession number DQ885937. The expressed and purified protein was stable and could be stored in 10 mM sodium phosphate buffer, pH 6.9, containing 100 mM NaCl and 1 mM DTT, by shock-freezing aliquots at −80 °C. After thawing and addition of 10% (v/v) 2H2O, the samples were used directly for NMR studies.

Structural NMR analysis

To investigate the integrity of the structure of purified NELF E RRM we performed CD (results not shown) and one-dimensional 1H-NMR spectroscopy. Using standard double- and triple-resonance NMR techniques with isotopically labelled protein, 1H, 13C and 15N backbone and side-chain resonances were assigned. Complete backbone and nearly complete side-chain assignment was obtained for the region Ala35–Arg109. For residues N-formyl-Met1–Arg34, as well as Lys110–Ser121, several amide resonances could not be assigned because of missing signals due to conformational exchange or proton exchange with the solvent.

The 1H/15N-HSQC spectrum of NELF E RRM shows the characteristic dispersion of a protein with an intact tertiary structure. Using standard double- and triple-resonance NMR techniques, it was possible to assign nearly all 1H, 13C and 15N resonances. While for the sequence region Lys38–Arg109 nearly all (>95%) backbone resonances and more than 85% of the side-chain resonances could be assigned, the terminal regions, especially the N-terminus, remained partially unassigned owing to strong overlap or missing resonances. The {1H} 15N-heteronuclear steady-state NOE experiment at 14.1 T (Figure 2) shows values around 0.6–0.8 for the region Lys38–Arg109, while outside of this region the heteronuclear NOE decreases towards the termini. {1H} 15N steady-state NOE values cluster around 0.7, indicating the absence of pronounced motions on the picosecond-to-nanosecond time scale for nearly all residues in between Gly39 and Arg109. Additionally, the assigned residues from the terminal regions show chemical shifts typically found for highly flexible polypeptides. This is characteristic for a compactly folded domain within the region Lys38–Arg109 and unstructured termini. Missing assignments of amide resonances can therefore be explained by conformational or solvent exchange. Owing to the flexible character of the termini, residues N-formyl-Met1–Arg34 and Leu114–Ser121 were excluded from further structural determination.

Size of heteronuclear {1H} 15N steady-state NOE at 14.1 T along the amino acid sequence

Figure 2
Size of heteronuclear {1H} 15N steady-state NOE at 14.1 T along the amino acid sequence

The values in the range 0.6–0.85 for residues Leu40–Cys110 indicate a rigid protein backbone in this region.

Figure 2
Size of heteronuclear {1H} 15N steady-state NOE at 14.1 T along the amino acid sequence

The values in the range 0.6–0.85 for residues Leu40–Cys110 indicate a rigid protein backbone in this region.

During the iterative structure determination, a set of 2037 experimental restraints, consisting of 1926 NOE-derived distance restraints, 24 hydrogen bonds (two distance restraints for each hydrogen bond), 32 dihedral restraints and 55 RDCs could be derived from NMR data (Table 1). The final structure calculation resulted in an ensemble of 20 structures showing no distance restraint violation larger than 0.16 Å, no violation of a dihedral restraint larger than 2.6° and no violation of an RDC larger than 0.48 Hz. Only small deviations from the idealized covalent bond geometry were obtained (Table 1). The resulting ensemble of 20 structures shows a high-co-ordinate precision of 0.28 Å for the heavy backbone atoms and 0.64 Å for all heavy atoms for residues Gly39–Ala108, corresponding to the structurally defined domain, as well as good stereochemical properties reflected by the fact that 91% of residues are located in the most-favoured regions of the Ramachandran plot (Figure 3A).

Solution structure of NELF E RRM

Figure 3
Solution structure of NELF E RRM

(A) Overlay of the 20 structures (residues Ala35–Met113) showing the lowest values of the target function excluding the database potential. (B) Ribbon presentation of the NELF E RRM structure. The Figure was generated using MOLMOL [38]. (C) Packing of Phe58, Val78 and Leu93. The amino acid residues are displayed as space-filled atoms. (D) The highly conserved aromatic residues Tyr43 and Phe77 derived from RNP2 and RNP1 of NELF E RRM are represented as balls and sticks. The N- and C-termini are indicated.

Figure 3
Solution structure of NELF E RRM

(A) Overlay of the 20 structures (residues Ala35–Met113) showing the lowest values of the target function excluding the database potential. (B) Ribbon presentation of the NELF E RRM structure. The Figure was generated using MOLMOL [38]. (C) Packing of Phe58, Val78 and Leu93. The amino acid residues are displayed as space-filled atoms. (D) The highly conserved aromatic residues Tyr43 and Phe77 derived from RNP2 and RNP1 of NELF E RRM are represented as balls and sticks. The N- and C-termini are indicated.

The solution structure of NELF E RRM exhibits a compact βαββαβ fold with a four-stranded antiparallel β-sheet (β1=Asn40–Tyr45, β2=Ile64–Asp70, β3=Cys75–Tyr80, β4=Gln102–Ile107) that packs against two helices (h1=Pro51–Phe61, h2=Met83–Leu93) which are oriented approximately perpendicular to each other (interhelix angle=114.3±1.8°; Figure 3B). Numerous hydrophobic contacts involving Leu42, Val44, Leu54, Ala57, Phe58, Val78, Ala86, Val90, Leu93, Val98, Val105 and Ile107 stabilize this packing. For example, Phe58 from h1 contacts Val78 from the central strand β3 and Leu93 from h2 (Figure 3C). These residues are conserved in all other known RRMs and adopt a similar conformation. Residues Thr96 and Gln97 form an additional short β-strand aligned antiparallel to β4, thus extending the β-sheet. This is similar to other RRMs where a β-hairpin is found in the sequence region between h2 and β4 [41]. The highly conserved aromatic residues Tyr43 and Phe77 known to be involved in base-stacking interactions with RNA in other RRMs are highlighted in Figure 3(D) [41].

Identification of the binding interface

To obtain information on the binding of NELF E RRM to the HIV-1 TAR RNA, we performed additional NMR experiments. The binding interface could be clearly defined by the NMR titration studies (Figure 4). Amide (1HN,15N) chemical shifts are very sensitive to local structural changes. Therefore observation of chemical-shift changes on titration of a binding partner to a 15N-labelled protein provides a powerful method for mapping of the binding interface. Addition of TAR RNA results in remarkable chemical-shift changes for resonances located in the central β-sheet (Asn41, Leu42, Tyr45, Cys75 and Phe77) and in the N-terminal region of h2 (Glu84 and Asp87; Figure 4A). For several residues in particular in strand β4 (Lys104–Ile107), signals were absent in the 1H/15N-HSQC spectra, probably due to exchange processes on the intermediate time scale, suggesting that these residues are involved in binding. This is characteristic for affinities in the low-micromolar range. The chemical-shift changes for residues located in strands β1 and β3 indicate the typical binding of RNA to the RRM by stacking of bases on to the two conserved aromatic residues Tyr43 and Phe77 (Figure 4B). Figure 4(C) shows a surface representation of the protein, highlighting the binding interface and the residues that exhibit chemical shift changes upon binding of TAR RNA.

Determination of the binding interface

Figure 4
Determination of the binding interface

(A) Overlay of the 1H/15N-HSQC spectra of free NELF E RRM (black) and the NELF E–TAR RNA complex (red). The amino acid resonances showing significant chemical-shift changes upon addition of an equimolar amount of TAR RNA are indicated by arrows. (B) Normalized chemical-shift changes of NELF E RRM upon TAR RNA binding. The normalized chemical-shift changes (weighted geometric average of 1HN and 15N chemical-shift changes) are shown as a function of the primary sequence. Changes larger than 0.04 (dotted line) were considered significant. Resonances disappearing after TAR binding due to extreme line broadening by chemical exchange are marked with an ×. (C) Surface representation of NELF E RRM highlighting the binding interface. Residues with resonances showing significant chemical-shift changes upon TAR binding (0.04<Δδ≤0.1), including the highly conserved Tyr43 and Phe77 of the RNP2 and RNP1 sequence motifs, are shown in yellow, those with resonances showing chemical-shift changes with Δδ>0.1 are indicated in light red and residues whose resonances could not be detected due to extreme line broadening are presented in light green.

Figure 4
Determination of the binding interface

(A) Overlay of the 1H/15N-HSQC spectra of free NELF E RRM (black) and the NELF E–TAR RNA complex (red). The amino acid resonances showing significant chemical-shift changes upon addition of an equimolar amount of TAR RNA are indicated by arrows. (B) Normalized chemical-shift changes of NELF E RRM upon TAR RNA binding. The normalized chemical-shift changes (weighted geometric average of 1HN and 15N chemical-shift changes) are shown as a function of the primary sequence. Changes larger than 0.04 (dotted line) were considered significant. Resonances disappearing after TAR binding due to extreme line broadening by chemical exchange are marked with an ×. (C) Surface representation of NELF E RRM highlighting the binding interface. Residues with resonances showing significant chemical-shift changes upon TAR binding (0.04<Δδ≤0.1), including the highly conserved Tyr43 and Phe77 of the RNP2 and RNP1 sequence motifs, are shown in yellow, those with resonances showing chemical-shift changes with Δδ>0.1 are indicated in light red and residues whose resonances could not be detected due to extreme line broadening are presented in light green.

The large chemical-shift changes seen in the Met113 and Ala116 resonances on TAR titration (Figure 4B) imply a structural change of this region which is highly flexible in free NELF E RRM. Structural changes of the C-terminus are observed, e.g. in hnRNP1 (heteronuclear ribonucleoprotein 1) RRM1, where the corresponding region is also unstructured in the free state and adopts a 310 helix upon RNA binding [42]. The additional β-strand of the NELF E RRM (Thr96–Gln97) does not exhibit significant chemical-shift changes, indicating that this region is not involved in RNA recognition. This is in contrast with the RRM from TcUBP1 (Trypanosoma cruzi U-rich-RNA-binding protein 1), where large chemical-shift changes of the resonances of the additional β-hairpin indicate interaction with RNA [43]. Unusual chemical-shift changes or disappearing signals are found for several residues in the loop preceding h2 (Glu81–Glu84 and Asp87). This region is not involved in RNA binding in complexes with known structures [42,4446], and the distance between the RNA-contact sites and h2 is rather large for the induction of secondary chemical-shift changes. We thus analysed whether a second NELF E RRM molecule is binding simultaneously to TAR.

Substrate binding

In order to obtain information on the stoichiometry of the complex, more detailed structural and dynamical characterizations, i.e. 15N-relaxation measurements and temperature-dependent titrations, were performed. However, analysis of the data was hampered by severe line broadening due to the large molecular mass of the complex, and only ambiguous data could be obtained (results not shown).

Therefore fluorescence titrations were performed. It has been shown previously that NELF E binds to the lower region of the TAR RNA stem [47]. However, none of the RRMs analysed so far exhibits high affinity for double-stranded RNA [41]. To determine whether NELF E RRM binds to single- or double-stranded TAR RNA, fluorescently labelled single or double-stranded oligoribonucleotides representing the lower stem of TAR were used for fluorescence equilibrium titrations. The experimental data obtained with the single-stranded RNA oligomers could be closely described by a two-state model, and Kd values of approx. 8.2 μM for TAR-(1–18) and 2.6 μM for TAR-(42–57) respectively could be derived (Figures 5A and 5B). Binding to the double-stranded RNA oligomer, however, could not be described by such a simple model (Figure 5C). However, further experiments will be necessary to obtain the precise binding region on the TAR RNA.

Fluorescence equilibrium titration with the TAR RNA stem

Figure 5
Fluorescence equilibrium titration with the TAR RNA stem

FAM-labelled (A) TAR-(1–18), (B) TAR-(42–57) or (C) double-stranded TAR RNA stem (50 nM) was titrated with NELF E RRM in 10 mM sodium phosphate buffer, pH 7.0, 100 mM NaCl. The curves in (A) and (B) show the best fit to a quadratic equation describing the binding equilibria with Kd values of 8.2±1.0 μM and 2.6±0.2 μM respectively. No curve could be fitted to the data in (C) using a simple two-state model.

Figure 5
Fluorescence equilibrium titration with the TAR RNA stem

FAM-labelled (A) TAR-(1–18), (B) TAR-(42–57) or (C) double-stranded TAR RNA stem (50 nM) was titrated with NELF E RRM in 10 mM sodium phosphate buffer, pH 7.0, 100 mM NaCl. The curves in (A) and (B) show the best fit to a quadratic equation describing the binding equilibria with Kd values of 8.2±1.0 μM and 2.6±0.2 μM respectively. No curve could be fitted to the data in (C) using a simple two-state model.

In parallel to our structure determination of the RRM, the NMR structure of the RRM of Parp14 which is identical with the NELF E RRM, was determined by the RIKEN Structural Genomics/Proteomics Initiative (PDB accession code 1X5P). The average structures of both ensembles display a backbone RMSD (root mean square deviation) of 1.3 Å showing a high similarity of the three-dimensional fold. The N- and C-terminal extensions of our construct, however, proved to be crucial for RNA binding.

The presented results thus demonstrate that NELF E RRM by itself is sufficient for binding to TAR RNA. The structures of RRMs generally exhibit a βαββαβ fold. The loops between the secondary-structure elements of RRMs are usually disordered and of various lengths. It is assumed that the different specificities and binding affinities of RRMs are, among other factors, modulated by variations in these loop regions [41]. Although full-length NELF E can bind to various RNA elements via its RRM, the precise binding regions have not yet been determined [9,47]. Thus analysis of the interaction of NELF E RRM with different single- and double-stranded RNAs will contribute to the understanding of the function of NELF E in the regulation of HIV transcription in particular, and eukaryotic transcription in general, and to identify new drug targets against HIV.

We thank Katrin Weiss and Nadine Herz for excellent technical assistance. This work was supported by grants from the Deutsche Forschungsgemeinschaft (Ro617/8-4) and the Sonderforschungsbereich 466.

Abbreviations

     
  • C12E5

    penta(ethylene glycol)monododecyl ether

  •  
  • DSIF

    DRB (5,6-dichloro-1β-D-ribofuranosylbenzimidazole)-sensitivity-inducing factor

  •  
  • DTT

    dithiothreitol

  •  
  • FAM

    6-carboxyfluorescein

  •  
  • HIV-1

    HIV type 1

  •  
  • HSQC

    heteronuclear single-quantum coherence

  •  
  • IPTG

    isopropyl β-D-thiogalactoside

  •  
  • MEXICO

    measurement of exchange rates in isotopically labelled compounds

  •  
  • NELF

    negative elongation factor

  •  
  • NOE

    nuclear Overhauser effect

  •  
  • Pol

    II, polymerase II

  •  
  • pTEFb

    positive transcription elongation factor b

  •  
  • RDC

    residual dipolar coupling

  •  
  • RMSD

    root mean square deviation

  •  
  • RRM

    RNA-recognition motif

  •  
  • TAR

    transactivation-response element

References

References
1
Taube
R.
Fujinaga
K.
Wimmer
J.
Barboric
M.
Peterlin
B. M.
Tat transactivation: a model for the regulation of eukaryotic transcriptional elongation
Virology
1999
, vol. 
264
 (pg. 
245
-
253
)
2
Kao
S. Y.
Calman
A. F.
Luciw
P. A.
Peterlin
B. M.
Anti-termination of transcription within the long terminal repeat of HIV-1 by tat gene product
Nature
1987
, vol. 
330
 (pg. 
489
-
493
)
3
Wada
T.
Orphanides
G.
Hasegawa
J.
Kim
D. K.
Shima
D.
Yamaguchi
Y.
Fukuda
A.
Hisatake
K.
Oh
S.
Reinberg
D.
Handa
H.
FACT relieves DSIF/NELF-mediated inhibition of transcriptional elongation and reveals functional differences between P-TEFb and TFIIH
Mol. Cell
2000
, vol. 
5
 (pg. 
1067
-
1072
)
4
Yamaguchi
Y.
Takagi
T.
Wada
T.
Yano
K.
Furuya
A.
Sugimoto
S.
Hasegawa
J.
Handa
H.
NELF, a multisubunit complex containing RD, cooperates with DSIF to repress RNA polymerase II elongation
Cell
1999
, vol. 
97
 (pg. 
41
-
51
)
5
Ping
Y. H.
Rana
T. M.
DSIF and NELF interact with RNA polymerase II elongation complex and HIV-1 tat stimulates P-TEFb-mediated phosphorylation of RNA polymerase II and DSIF during transcription elongation
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
12951
-
12958
)
6
Zhu
Y.
Pe'ery
T.
Peng
J.
Ramanathan
Y.
Marshall
N.
Marshall
T.
Amendt
B.
Mathews
M. B.
Price
D. H.
Transcription elongation factor P-TEFb is required for HIV-1 tat transactivation in vitro
Genes Dev.
1997
, vol. 
11
 (pg. 
2622
-
2632
)
7
Narita
T.
Yamaguchi
Y.
Yano
K.
Sugimoto
S.
Chanarat
S.
Wada
T.
Kim
D. K.
Hasegawa
J.
Omori
M.
Inukai
N.
, et al. 
Human transcription elongation factor NELF: identification of novel subunits and reconstitution of the functionally active complex
Mol. Cell. Biol.
2003
, vol. 
23
 (pg. 
1863
-
1873
)
8
Yamaguchi
Y.
Filipovska
J.
Yano
K.
Furuya
A.
Inukai
N.
Narita
T.
Wada
T.
Sugimoto
S.
Konarska
M. M.
Handa
H.
Stimulation of RNA polymerase II elongation by hepatitis delta antigen
Science
2001
, vol. 
293
 (pg. 
124
-
127
)
9
Yamaguchi
Y.
Inukai
N.
Narita
T.
Wada
T.
Handa
H.
Evidence that negative elongation factor represses transcription elongation through binding to a DRB sensitivity-inducing factor/RNA polymerase II complex and RNA
Mol. Cell. Biol.
2002
, vol. 
22
 (pg. 
2918
-
2927
)
10
Meyer
O.
Schlegel
H. G.
Biology of aerobic carbon monoxide-oxidizing bacteria
Annu. Rev. Microbiol.
1983
, vol. 
37
 (pg. 
277
-
310
)
11
Sambrook
J.
Maniatis
T.
Fritsch
E. F.
Molecular Cloning: a Laboratory Manual
1989
Cold Spring Harbor
Cold Spring Harbor Laboratory Press
12
Marley
J.
Lu
M.
Bracken
C.
A method for efficient isotopic labeling of recombinant proteins
J. Biomol. NMR
2001
, vol. 
20
 (pg. 
71
-
75
)
13
Metzger
A. U.
Schindler
T.
Willbold
D.
Kraft
M.
Steegborn
C.
Volkmann
A.
Frank
R. W.
Rösch
P.
Structural rearrangements on HIV-1 tat (32–72) TAR complex formation
FEBS Lett.
1996
, vol. 
384
 (pg. 
255
-
259
)
14
Bax
A.
Grzesiek
S.
Methodological advances in protein NMR
Acc. Chem. Res.
1993
, vol. 
26
 (pg. 
131
-
138
)
15
Schleucher
J.
Schwendinger
M.
Sattler
M.
Schmidt
P.
Schedletzky
O.
Glaser
S. J.
Sorensen
O. W.
Griesinger
C.
A general enhancement scheme in heteronuclear multidimensional NMR employing pulsed field gradients
J. Biomol. NMR
1994
, vol. 
4
 (pg. 
301
-
306
)
16
Ikura
M.
Kay
L. E.
Tschudin
R.
Bax
A.
Three-dimensional NOESY-HMQC spectroscopy of a 13C-labeled compound
J. Magn. Reson.
1990
, vol. 
86
 (pg. 
204
-
209
)
17
Talluri
S.
Wagner
G.
An optimized 3D NOESY-HSQC
J. Magn. Reson. Ser. B.
1996
, vol. 
112
 (pg. 
200
-
205
)
18
Vuister
W. G.
Bax
A.
Quantitative J correlation: a new approach for measuring homonuclear three-bond J(HNHα) coupling constants in 15N-enriched proteins
J. Am. Chem. Soc.
1993
, vol. 
115
 (pg. 
7772
-
7777
)
19
Koide
S.
Jahnke
W.
Wright
P. E.
Measurement of intrinsic exchange rates of amide protons in a 15N-labeled peptide
J. Biomol. NMR
1995
, vol. 
6
 (pg. 
306
-
312
)
20
Ottiger
M.
Delaglio
F.
Bax
A.
Measurement of J and dipolar couplings from simplified two-dimensional NMR spectra
J. Magn. Reson.
1998
, vol. 
131
 (pg. 
373
-
378
)
21
Otting
G.
Ruckert
M.
Levitt
M. H.
Moshref
A.
NMR experiments for the sign determination of homonuclear scalar and residual dipolar couplings
J. Biomol. NMR
2000
, vol. 
16
 (pg. 
343
-
346
)
22
Dayie
K. T.
Wagner
G.
Relaxation-rate measurements of 15N-1H groups with pulsed-field gradients and preservation of coherence pathways
J. Magn. Reson. Ser. A
1994
, vol. 
111
 (pg. 
121
-
126
)
23
Hajduk
P. J.
Dinges
J.
Miknis
G. F.
Merlock
M.
Middleton
T.
Kempf
D. J.
Egan
D. A.
Walter
K. A.
Robins
T. S.
Shuker
S. B.
, et al. 
NMR-based discovery of lead inhibitors that block DNA binding of the human papillomavirus E2 protein
J. Med. Chem.
1997
, vol. 
40
 (pg. 
3144
-
3150
)
24
Johnson
B. A.
Blevins
R. A.
NMRview: a computer program for the visualization and analysis of NMR data
J. Biomol. NMR
1994
, vol. 
4
 (pg. 
603
-
614
)
25
Nilges
M.
Calculation of protein structures with ambiguous distance restraints. automated assignment of ambiguous NOE crosspeaks and disulphide connectivities
J. Mol. Biol.
1995
, vol. 
245
 (pg. 
645
-
660
)
26
Schweimer
K.
Hoffmann
S.
Bauer
F.
Friedrich
U.
Kardinal
C.
Feller
S. M.
Biesinger
B.
Sticht
H.
Structural investigation of the binding of a herpesviral protein to the SH3 domain of tyrosine kinase lck
Biochemistry
2002
, vol. 
41
 (pg. 
5120
-
5130
)
27
Wüthrich
K.
NMR of proteins and nucleic acids
1986
New York
Wiley
28
Nilges
M.
Gronenborn
A. M.
Brunger
A. T.
Clore
G. M.
Determination of three-dimensional structures of proteins by simulated annealing with interproton distance restraints: application to crambin, potato carboxypeptidase inhibitor and barley serine proteinase inhibitor 2
Protein Eng.
1988
, vol. 
2
 (pg. 
27
-
38
)
29
Nilges
M.
Clore
G. M.
Gronenborn
A. M.
Determination of three-dimensional structures of proteins from interproton distance data by dynamical simulated annealing from a random array of atoms. circumventing problems associated with folding
FEBS Lett.
1988
, vol. 
239
 (pg. 
129
-
136
)
30
Folmer
R. H.
Hilbers
C. W.
Konings
R. N.
Nilges
M.
Floating stereospecific assignment revisited: application to an 18 kDa protein and comparison with J-coupling data
J. Biomol. NMR
1997
, vol. 
9
 (pg. 
245
-
258
)
31
Kuszewski
J.
Clore
G. M.
Sources of and solutions to problems in the refinement of protein NMR structures against torsion angle potentials of mean force
J. Magn. Reson.
2000
, vol. 
146
 (pg. 
249
-
254
)
32
Neudecker
P.
Sticht
H.
Rösch
P.
Improving the efficiency of the Gaussian conformational database potential for the refinement of protein and nucleic acid structures
J. Biomol. NMR
2001
, vol. 
21
 (pg. 
373
-
375
)
33
Neudecker
P.
Nerkamp
J.
Eisenmann
A.
Nourse
A.
Lauber
T.
Schweimer
K.
Lehmann
K.
Schwarzinger
S.
Ferreira
F.
Rösch
P.
Solution structure, dynamics, and hydrodynamics of the calcium-bound cross-reactive birch pollen allergen bet v 4 reveal a canonical monomeric two EF-hand assembly with a regulatory function
J. Mol. Biol.
2004
, vol. 
336
 (pg. 
1141
-
1157
)
34
Powell
M. J. D.
Restart procedures for conjugate gradient method
Math. Program.
1977
, vol. 
12
 (pg. 
241
-
254
)
35
Tjandra
N.
Grzesiek
S.
Bax
A.
Magnetic field dependence of nitrogen-proton J splittings in 15N-enriched human ubiquitin resulting from relaxation interference and residual dipolar coupling
J. Am. Chem. Soc.
1996
, vol. 
118
 (pg. 
6264
-
6272
)
36
Clore
G. M.
Gronenborn
A. M.
Bax
A.
A robust method for determining the magnitude of the fully asymmetric alignment tensor of oriented macromolecules in the absence of structural information
J. Magn. Reson.
1998
, vol. 
133
 (pg. 
216
-
221
)
37
Reference deleted
38
Koradi
R.
Billeter
M.
Wüthrich
K.
MOLMOL: a program for display and analysis of macromolecular structures
J. Mol. Graphics
1996
, vol. 
14
 (pg. 
51
-
55
)
39
Morris
A. L.
MacArthur
M. W.
Hutchinson
E. G.
Thornton
J. M.
Stereochemical quality of protein structure coordinates
Proteins
1992
, vol. 
12
 (pg. 
345
-
364
)
40
Laskowski
R. A.
Rullmannn
J. A.
MacArthur
M. W.
Kaptein
R.
Thornton
J. M.
AQUA and PROCHECK-NMR: programs for checking the quality of protein structures solved by NMR
J. Biomol. NMR
1996
, vol. 
8
 (pg. 
477
-
486
)
41
Maris
C.
Dominguez
C.
Allain
F. H.
The RNA recognition motif, a plastic RNA-binding platform to regulate post-transcriptional gene expression
FEBS J.
2005
, vol. 
272
 (pg. 
2118
-
2131
)
42
Allain
F. H.
Gubser
C. C.
Howe
P. W.
Nagai
K.
Neuhaus
D.
Varani
G.
Specificity of ribonucleoprotein interaction determined by RNA folding during complex formulation
Nature
1996
, vol. 
380
 (pg. 
646
-
650
)
43
Volpon
L.
D'Orso
I.
Young
C. R.
Frasch
A. C.
Gehring
K.
NMR structural study of TcUBP1, a single RRM domain protein from Trypanosoma cruzi: contribution of a β hairpin to RNA binding
Biochemistry
2005
, vol. 
44
 (pg. 
3708
-
3717
)
44
Deo
R. C.
Bonanno
J. B.
Sonenberg
N.
Burley
S. K.
Recognition of polyadenylate RNA by the poly(A)-binding protein
Cell
1999
, vol. 
98
 (pg. 
835
-
845
)
45
Handa
N.
Nureki
O.
Kurimoto
K.
Kim
I.
Sakamoto
H.
Shimura
Y.
Muto
Y.
Yokoyama
S.
Structural basis for recognition of the tra mRNA precursor by the sex-lethal protein
Nature
1999
, vol. 
398
 (pg. 
579
-
585
)
46
Wang
X.
Tanaka Hall
T. M.
Structural basis for recognition of AU-rich element RNA by the HuD protein
Nat. Struct. Biol.
2001
, vol. 
8
 (pg. 
141
-
145
)
47
Fujinaga
K.
Irwin
D.
Huang
Y.
Taube
R.
Kurosu
T.
Peterlin
B. M.
Dynamics of human immunodeficiency virus transcription: P-TEFb phosphorylates RD and dissociates negative effectors from the transactivation response element
Mol. Cell. Biol.
2004
, vol. 
24
 (pg. 
787
-
795
)

Author notes

1

These authors contributed equally to this work.

Structural co-ordinates for NELF (negative elongation factor) E RNA-recognition motif have been deposited in the Protein Data Bank under accession code 2BZ2.

The nucleotide sequence data for NELF (negative elongation factor) E RNA-recognition motif have been deposited in the DDBJ, EMBL, GenBank® and GSDB Nucleotide Sequence Databases under the accession number DQ885937.