The eukaryotic transcription elongation factor DSIF [DRB (5,6-dichloro-1-β-D-ribofuranosylbenzimidazole) sensitivity-inducing factor] is composed of two subunits, hSpt4 and hSpt5, which are homologous to the yeast factors Spt4 and Spt5. DSIF is involved in regulating the processivity of RNA polymerase II and plays an essential role in transcriptional activation of eukaryotes. At several eukaryotic promoters, DSIF, together with NELF (negative elongation factor), leads to promoter-proximal pausing of RNA polymerase II. In the present paper we describe the crystal structure of hSpt4 in complex with the dimerization region of hSpt5 (amino acids 176–273) at a resolution of 1.55 Å (1 Å=0.1 nm). The heterodimer shows high structural similarity to its homologue from Saccharomyces cerevisiae. Furthermore, hSpt5-NGN is structurally similar to the NTD (N-terminal domain) of the bacterial transcription factor NusG. A homologue for hSpt4 has not yet been found in bacteria. However, the archaeal transcription factor RpoE” appears to be distantly related. Although a comparison of the NusG-NTD of Escherichia coli with hSpt5 revealed a similarity of the three-dimensional structures, interaction of E. coli NusG-NTD with hSpt4 could not be observed by NMR titration experiments. A conserved glutamate residue, which was shown to be crucial for dimerization in yeast, is also involved in the human heterodimer, but is substituted for a glutamine residue in Escherichia coli NusG. However, exchanging the glutamine for glutamate proved not to be sufficient to induce hSpt4 binding.

INTRODUCTION

The human DSIF [DRB (5,6-dichloro-1-β-D-ribofuranosylbenzimidazole) sensitivity-inducing factor] is a key player in the control of eukaryotic transcriptional elongation and has complex functions. It is involved in early elongation control by inducing pausing of RNAPol II (RNA polymerase II) in conjunction with the NELF (negative elongation factor) [1,2]. This was illustrated by promoter proximal pausing during transcription of the HIV-1 proviral genome DSIF and NELF interact directly with RNAPol II, and NELF also binds to the HIV-1 TAR RNA. Release of RNAPol II requires the viral transactivator protein Tat, which recruits the cellular P-TEFb (positive transcription elongation factor b) to the TAR RNA. In turn, P-TEFb phosphorylates the C-terminal domain of the largest RNAPol II subunit, the E subunit of NELF and the hSpt5 subunit of DSIF. Only then does NELF detach from the transcription machinery and DSIF is converted into an activator, thus stimulating elongation [38].

Similar promoter proximal pausing events involving NELF and DSIF have been detected at cellular genes under control of heat-shock promoters in Drosophila melanogaster [911]. Furthermore, interactions of DSIF or its yeast homologue with RNA-modifying enzymes, such as capping enzyme and cap methyltransferase, and other cellular factors have been reported, underlining the diversity of DSIF in transcriptional regulation [1215].

DSIF is a heterodimeric protein consisting of the 14 kDa subunit hSpt4 and the 120 kDa subunit hSpt5, which are homologous to the transcription factors Spt4 and Spt5 from the yeast Saccharomyces cerevisiae (Figure 1) [16]. hSpt5 is well conserved in eukaryotes and shares homology with NusG proteins of bacteria and archaea.

Schematic representation of the domain structures of hSpt4 and hSpt5

Figure 1
Schematic representation of the domain structures of hSpt4 and hSpt5

(A) Domain structure of hSpt4 and secondary structure distribution of hSpt4 in the hSpt4–hSpt5-NGN heterodimer. The amino acid sequences of hSpt4 (PDB: 3H7H), Spt4 from S. cerevisiae (PDB: 2EXU) and archaeal RpoE” from P. furiosus (PDB: 1RYQ) were aligned. Conserved amino acids are highlighted by a black frame, conserved substitutions are represented by white letters on grey boxes and semi-conserved substitutions are highlighted by black letters on grey boxes. Secondary structure elements of complexed hSpt4 (PDB: 3H7H) are indicated on top of the alignment. Numbers underneath the alignment refer to amino acid positions of hSpt4. (B) Domain architecture of hSpt5 [21,22]. The amino acid sequences of hSpt5-NGN (residues: 176–273) (PDB: 3H7H), Spt5-NGN from S. cerevisiae (residues: 286–375) (PDB: 2EXU) and E. coli NusG-NTD (PDB: 2K06) are aligned. Labelling of conservation of amino acids is the same as in (A). The secondary structure composition of complexed hSpt5-NGN is displayed on top of the alignment. C, cysteine residues involved in zinc co-ordination; CTR 1 and 2, C-terminal repeat 1 and 2 respectively; acidic, N-terminal acidic region. Numbers underneath the alignment refer to amino acid positions of hSpt5.

Figure 1
Schematic representation of the domain structures of hSpt4 and hSpt5

(A) Domain structure of hSpt4 and secondary structure distribution of hSpt4 in the hSpt4–hSpt5-NGN heterodimer. The amino acid sequences of hSpt4 (PDB: 3H7H), Spt4 from S. cerevisiae (PDB: 2EXU) and archaeal RpoE” from P. furiosus (PDB: 1RYQ) were aligned. Conserved amino acids are highlighted by a black frame, conserved substitutions are represented by white letters on grey boxes and semi-conserved substitutions are highlighted by black letters on grey boxes. Secondary structure elements of complexed hSpt4 (PDB: 3H7H) are indicated on top of the alignment. Numbers underneath the alignment refer to amino acid positions of hSpt4. (B) Domain architecture of hSpt5 [21,22]. The amino acid sequences of hSpt5-NGN (residues: 176–273) (PDB: 3H7H), Spt5-NGN from S. cerevisiae (residues: 286–375) (PDB: 2EXU) and E. coli NusG-NTD (PDB: 2K06) are aligned. Labelling of conservation of amino acids is the same as in (A). The secondary structure composition of complexed hSpt5-NGN is displayed on top of the alignment. C, cysteine residues involved in zinc co-ordination; CTR 1 and 2, C-terminal repeat 1 and 2 respectively; acidic, N-terminal acidic region. Numbers underneath the alignment refer to amino acid positions of hSpt5.

The bacterial transcription factor NusG fulfills various as well as species-specific functions. In Escherichia coli, it enhances transcriptional elongation as well as ρ-dependent termination [17,18]. In contrast, it acts as an anti-termination factor when phage λ N protein and other Nus proteins (A, B and E) are present [19,20].

hSpt5 is comprised of an N-terminal acidic region, whose function has not yet been established, followed by the NGN (NusG N-terminal homology domain) (residues 176–270) [21,22], which exhibits homology with the NTD (N-terminal domain) of NusG proteins from bacteria and archaea. The NGNs of hSpt5 and of the Spt5 proteins of S. cerevisiae and the fission yeast Schizosaccharomyces pombe have been identified as harbouring the binding region for the small subunit hSpt4 [3,16,24]. Four to six KOW (after Kyrpides, Ouzounis and Woese [26]) motifs, which directly interact with RNAPol II during transcription, have been assigned C-terminally to the NGN in hSpt5 [21,22,25]. KOW motifs have also been identified in the NusG family proteins and the proteins RL24, 26 and 27 of the large ribosomal subunit [26]. The C-terminus of hSpt5 contains the hepta- and octa-peptide repeats CTR1 and CTR2 [3,4]. Phosphorylation of threonine residues in the CTR1 region of hSpt5 as well as methylation of arginine residues near its KOW motifs affects DSIF activity [8,27].

hSpt4, the smaller subunit of DSIF, harbours an N-terminal four-cysteine zinc finger, and exhibits α/β-topology [28,29]. The presence of a bound Zn2+ was verified previously [29]. In contrast with hSpt5, no bacterial homologue exists for hSpt4, but it is uniformly conserved in archaea and eukaryotes. The archaeal transcription factor RpoE”, originally annotated as RNA polymerase subunit E”, is distantly related to hSpt4 and is regarded as its evolutionary ancestor [22,30].

In the present work we determined the crystal structure of the recombinant hSpt4 complexed with hSpt5-NGN [human Spt5(176–273)], representing the binding region for hSpt4, at a resolution of 1.55 Å (1 Å=0.1 nm). The structure confirms that hSpt4 is indeed an α/β-type protein with a zinc finger and that it forms a tight complex with the NGN of hSpt5. Comparisons with the crystal structure of the homologous Spt4–Spt5(286–375) complex from S. cerevisiae [31] verified the high structural similarity already indicated by sequence alignments. Biophysical analyses revealed that, in spite of a similar folding topology of NusG-NTD and hSpt5-NGN, NusG-NTD lacks the ability to interact with hSpt4. Moreover, we also show a structural relationship between hSpt4 and the archaeal transcription factor RpoE”.

EXPERIMENTAL

Plasmid construction

Construction of plasmid pET-GB1-hSpt4, harbouring the DNA sequences for the solubility tag GB1 (streptococcal immunoglobulin-binding domain of protein G), as well as a TEV (tobacco etch virus) protease cleavage site followed by the full-length hSpt4 gene was described previously [29]. A synthetically produced gene adapted to the E. coli codon usage harbouring the NGN region (amino acids 176–273) of the DSIF hSpt5 gene was cloned into the expression vector pET15b (Novagen) via its NdeI and BamHI restriction sites. The final plasmid pET15b-hSpt5-NGN encodes an N-terminal 6×His-tagged fusion protein with a thrombin cleavage site.

Gene expression and protein purification

The plasmids pET15b-hSpt5-NGN and pET-GB1-hSpt4 were co-transformed into the E. coli strain BL21 (DE3) (Novagen). Cells were grown as described previously [29] using kanamycin (30 μg/ml) and ampicillin (100 μg/ml) as antibiotics. Purification of the complex by Ni-NTA (Ni2+-nitrilotriacetate)-affinity chromatography was performed as described previously for hSpt4 [29]. Fractions containing the DSIF complex were subjected to overnight TEV cleavage at 4 °C, followed by a second Ni2+-affinity chromatography step to remove excess cleaved hSpt4 not complexed with hSpt5. The N-terminal 6×His-tag of hSpt5 was removed by thrombin cleavage at room temperature (20 °C) overnight. To remove the cleaved-off affinity tags GB1 and 6×His, the sample was applied to a Ni-NTA column. The DSIF complex was collected in the flow-through and dialysed in steps in 25 mM Tris/HCl, pH 7.4, 10 mM 2-mercaptoethanol and decreasing concentrations of NaCl down to 50 mM.

The sample was subjected to cation-exchange chromatography (HiTrap SP XL column; GE Healthcare), dialysed against 25 mM Tris/HCl, pH 7.0, 50 mM NaCl, 10 mM 2-mercaptoethanol and concentrated in Vivaspin concentrators MWCO (molecular mass cut-off) 5000 (Sartorius, Göttingen, Germany).

The genes expressing untagged NusG-NTD and NusG-NTD(Q72E) were cloned into pET11a (Novagen) and the proteins were purified from E. coli extracts via heparin-affinity and size-exclusion chromatography using a Superdex 75 HR 10/30 column (GE Healthcare). Samples were dialysed and concentrated as described above for the hSpt4–hSpt5-NGN complex.

Size-exclusion chromatography

Size-exclusion chromatography to determine the molecular mass of the hSpt4–hSpt5-NGN complex was performed using a Superdex 75 HR 10/30 column (GE Healthcare). As molecular mass standards, albumin (67.0 kDa, GE Healthcare) chymotrypsinogen A (25.0 kDa, GE Healthcare) and cytochrome c (12.4 kDa, Sigma–Aldrich) were used. Standard proteins were dissolved in 25 mM Tris/HCl, pH 7.0, 50 mM NaCl and 10 mM 2-mercaptoethanol, which was also used for equilibration and the column run. The sample was dialysed against the same buffer.

NMR spectroscopy

Two dimensional 15N-1H-HSQC (heteronuclear single-quantum coherence) NMR spectra were recorded on a Bruker AV 800 MHz NMR spectrometer equipped with a cryogenic probe and pulsed field-gradient capabilities at 298 K according to the FHSQC (fast HSQC) scheme [32]. For data processing and visualization, in-house software and NMRView [33] were used respectively.

Crystallization and data collection

Recombinant hSpt4–hSpt5-NGN was crystallized by the hanging-drop vapour diffusion method at 290 K by mixing 2 μl of protein solution (1.2 mg/ml in 25 mM Tris/HCl, pH 7.0, and 10 mM 2-mercaptoethanol) plus 2 μl of precipitant {50 mM Tris/HCl, pH 7.0, 0.2 M (NH4)2SO4, 13–15% PEG [poly(ethylene glycol)] 3350 and 10 mM 2-mercaptoethanol}. Triangular-shaped crystals grew within 1 week with dimensions of approx. 300×150×50 μm. Crystals were harvested in 17 μl of precipitant plus 3 μl of 2R,3R-butandiol (Sigma–Aldrich) before flash-cooling in liquid nitrogen. Zinc-MAD (multi-wavelength anomalous dispersion) diffraction data were collected at 100 K at the synchrotron beam line 14.2 of BESSY (Berlin electron storage ring company for synchrotron radiation; Berlin, Germany) with a single crystal belonging to the orthorhombic system (space group P212121 with cell constants of a=41.096 Å, b=52.091 Å and c=97.099 Å). The crystal diffracted to a resolution of 1.55 Å at the zinc-MAD remote wavelength (0.91841 Å). There is one heterodimer per asymmetric unit with a solvent content of 40%. Data were processed and scaled with XDS (X-ray dectector software) [34]. Data statistics are reported in Table 1.

Table 1
Data collection and refinement statistics

Values in parentheses are given for the highest resolution shell (1.59–1.55 Å).

Parameter Data collection  
Wavelength (Å) 0.91841 (remote) 1.28308 (Zn-peak) 
dmin (Å) 19.11–1.55 19.11–1.60 
Space group P212121  
Cell dimensions (Å)   
a 41.096  
b 52.091  
c 97.099  
Total/unique reflections 115753/30454 105462/51330 
Completeness (%) 98.1 (86.9) 97.1 (94.7) 
Rs3.4 (23.6) 2.4 (10.8) 
I/σ 22.7 (4.7) 22.1 (7.3) 
Model refinement   
dmin (Å) 19.11–1.55  
 Reflections 28930  
R-factor/Rfree-factor† 0.186/0.222 (0.222/0.256)  
 RMSD-bond length (Å) 0.01  
 RMSD-bond angles (°) 1.277  
 Average B-factor (Å214.36  
 ESU‡ (Å) 0.054  
 Protein atoms 1746  
 Solvent atoms 196  
 Metal ions (Zn+2 
Parameter Data collection  
Wavelength (Å) 0.91841 (remote) 1.28308 (Zn-peak) 
dmin (Å) 19.11–1.55 19.11–1.60 
Space group P212121  
Cell dimensions (Å)   
a 41.096  
b 52.091  
c 97.099  
Total/unique reflections 115753/30454 105462/51330 
Completeness (%) 98.1 (86.9) 97.1 (94.7) 
Rs3.4 (23.6) 2.4 (10.8) 
I/σ 22.7 (4.7) 22.1 (7.3) 
Model refinement   
dmin (Å) 19.11–1.55  
 Reflections 28930  
R-factor/Rfree-factor† 0.186/0.222 (0.222/0.256)  
 RMSD-bond length (Å) 0.01  
 RMSD-bond angles (°) 1.277  
 Average B-factor (Å214.36  
 ESU‡ (Å) 0.054  
 Protein atoms 1746  
 Solvent atoms 196  
 Metal ions (Zn+2 
*

Rsh Σi|Ii(h)−<I(h)>|/ΣhΣiIi(h); where i are the independent observations of reflection h.

The Rfree-factor was calculated from a 5% subset of reflections (1523), which were removed at random before the refinement was carried out.

Estimated overall co-ordinate error (ESU) based on maximum likelihood.

Structure determination and refinement

Initial phases were obtained with the SHELX program [35] by SAD (single-wavelength anomalous dispersion) phasing using the wavelength 1.28308 Å corresponding to the peak of the zinc-MAD experiment. Automatic model building with ARP/wARP software [36] resulted in a largely complete polypeptide model. Manual model building and positioning of the Zn2+ atom associated with hSpt4 were performed with the program COOT [37]. Positional and temperature factor refinements were carried out with Refmac5 [38]. In the final refinement cycles, the B-factors of Zn2+ and the S atoms of the four co-ordinating cysteine residues were refined anisotropically and alternative conformations of side chains were included. Several side chains located at the surface of the molecules (16 from hSpt4 and 11 from hSpt5-NGN) were only partially visible in the electron density maps and were therefore modelled using one of the rotamers.

ARP/wARP and Refmac5 programs were used as implemented in the CPP4 suite [39]. The refined model contained one hSpt4 subunit (Gly1 to Thr117) and one hSpt5-NGN subunit (Asp176 to Glu269), one Zn2+ ion and 196 solvent molecules. The N-terminal methionine residue of hSpt4 (numbered zero in the PDB file 3H7H) and the N-terminal residue methionine of hSpt5-NGN (numbered 175 in the PDB file 3H7H) are cloning artifacts. The last four C-terminal residues of hSpt5-NGN (Val-Ala-Asn-Leu), and the next four residues (Gly-Ser-Gly-Cys; derived from the cloning strategy) were not visible in the electron density maps and are therefore not included in the refined model. The final refinement statistics are shown in Table 1.

Figures were generated by PyMOL v0.99 (DeLano Scientific; http://www.pymol.org) and electrostatic surface potential calculations for hSpt5-NGN, Spt5-NGN and E. coli NusG-NTD were performed using the APBS (adaptive Poisson–Boltzmann solver) tool [41] for the PyMOL program written by Michael Lerner.

All secondary structures shown were determined by PROCHECK [42]. ClustalW2 [43] was used for sequence alignments. The program LIGPLOT [44] was used to identify the amino acid residues involved in dimerization. The co-ordinates and structure factors were deposited in the Protein Data Bank with accession code 3H7H.

RESULTS AND DISCUSSION

To facilitate formation of the correctly folded heterodimeric hSpt4–hSpt5-NGN complex, we chose to co-express the two genes, which were encoded on separate plasmids in E. coli. The hSpt5 construct used (amino acids 176–273) encompasses the NGN [19,20]. Previous attempts to express the two genes, hSpt4 and hSpt5-NGN, in separate strains and then combine the proteins, failed due to insolubility of hSpt5-NGN. Only hSpt4 could be purified successfully [29].

The integrity of the purified hSpt4–hSpt5-NGN complex was verified by size-exclusion chromatography (Figure 2A). Only one peak was detected, which eluted at 12.4 ml, corresponding to an apparent molecular mass of 23.0 kDa. This in good agreement with the calculated molecular mass of the complex of 25.3 kDa. Peak fractions analysed by SDS/PAGE revealed two bands, corresponding to hSpt4 and hSpt5-NGN (Figure 2B). The identity of the proteins was verified by MS. These analyses confirm that the heterodimer observed in the crystal structure most probably reflects the physiological complex and is not an artifact from crystal packing.

Analysis of purified hSpt4–hSpt5-NGN

Figure 2
Analysis of purified hSpt4–hSpt5-NGN

(A) Size-exclusion chromatography using a Superdex 75 HR 10/30 column (GE Healthcare). The elution volumes and molecular masses (in kDa) of the standard proteins used for column calibration are indicated. The numbers 1–3 correspond to the lanes shown in (B). (B) SDS/PAGE (19% gel). The gel shows eluted fractions from (A) of hSpt4–hSpt5-NGN. The bands corresponding to hSpt4 and hSpt5-NGN are indicated. Numbers on the left indicate the molecular masses of standard proteins (lane M).

Figure 2
Analysis of purified hSpt4–hSpt5-NGN

(A) Size-exclusion chromatography using a Superdex 75 HR 10/30 column (GE Healthcare). The elution volumes and molecular masses (in kDa) of the standard proteins used for column calibration are indicated. The numbers 1–3 correspond to the lanes shown in (B). (B) SDS/PAGE (19% gel). The gel shows eluted fractions from (A) of hSpt4–hSpt5-NGN. The bands corresponding to hSpt4 and hSpt5-NGN are indicated. Numbers on the left indicate the molecular masses of standard proteins (lane M).

Structure of the hSpt4–hSpt5-NGN complex

The crystal structure of hSpt4–hSpt5-NGN was determined to a resolution of 1.55 Å by zinc-SAD phasing. Crystal parameters, data collection and refinement statistics are summarized in Table 1. The crystal structure of the hSpt4–hSpt5-NGN complex shows a compact heterodimer with an overall surface area of 11547 Å2 and an overall buried surface of 3159 Å2 [45].

The hSpt5-NGN of the DSIF heterodimer is an α/β type protein with β-strand and α-helical contents of 17.3% and 34.7% respectively. It exhibits a β1α1β2β3α2α3β4α4α5β5 folding topology (Figure 3) with a central four-stranded antiparallel β-sheet that is sandwiched between three helices (α1, α2 and α3) and two short single helices (α4 and α5). Helices α2 and α3 span across the β-sheet, thereby shielding parts of the sheet from the solvent. Furthermore, the largest helix (α1) of hSpt5-NGN is aligned along β2 and both are part of the dimerization interface.

Structure of the hSpt4–hSpt5-NGN heterodimer

Figure 3
Structure of the hSpt4–hSpt5-NGN heterodimer

(A) Ribbon diagram of the hSpt4–hSpt5-NGN heterodimer. hSpt5-NGN is presented in green, hSpt4 in blue; three of the four cysteine residues (Cys16, Cys19, Cys33 and Cys36) co-ordinating the zinc ion (grey sphere) are shown in yellow; for clarity not all secondary structure elements are indicated. (B) Topology of the secondary structure elements in the hSpt4–hSpt5-NGN complex. Green: hSpt5-NGN; blue: hSpt4. β-strands are represented by arrows, α-helices by cylinders. (C) Electron density map. 2FobsFcal electron density map (blue) contoured at 1.0 σ of the zinc finger region of hSpt4. (D) Dimerization interface of the complex. hSpt5-NGN is depicted in green, the surface of hSpt4 is represented semi-transparently. The secondary structure elements of hSpt4 are shown in blue, unstructured regions in light grey. Residues of hSpt4 interacting with hSpt5-NGN are shown in red (polar) or dark grey (hydrophobic). Residues of hSpt5-NGN involved in hydrophobic interactions with hSpt4 are coloured in black, polar residues are coloured in red. The interacting residues of hSpt5-NGN are labelled. The conserved Glu228 of hSpt5 and its interaction partner Ser69 (in italics) of hSpt4 are highlighted in orange.

Figure 3
Structure of the hSpt4–hSpt5-NGN heterodimer

(A) Ribbon diagram of the hSpt4–hSpt5-NGN heterodimer. hSpt5-NGN is presented in green, hSpt4 in blue; three of the four cysteine residues (Cys16, Cys19, Cys33 and Cys36) co-ordinating the zinc ion (grey sphere) are shown in yellow; for clarity not all secondary structure elements are indicated. (B) Topology of the secondary structure elements in the hSpt4–hSpt5-NGN complex. Green: hSpt5-NGN; blue: hSpt4. β-strands are represented by arrows, α-helices by cylinders. (C) Electron density map. 2FobsFcal electron density map (blue) contoured at 1.0 σ of the zinc finger region of hSpt4. (D) Dimerization interface of the complex. hSpt5-NGN is depicted in green, the surface of hSpt4 is represented semi-transparently. The secondary structure elements of hSpt4 are shown in blue, unstructured regions in light grey. Residues of hSpt4 interacting with hSpt5-NGN are shown in red (polar) or dark grey (hydrophobic). Residues of hSpt5-NGN involved in hydrophobic interactions with hSpt4 are coloured in black, polar residues are coloured in red. The interacting residues of hSpt5-NGN are labelled. The conserved Glu228 of hSpt5 and its interaction partner Ser69 (in italics) of hSpt4 are highlighted in orange.

The secondary structure distribution of hSpt4 is similar to hSpt5, with approx. 32.2% α-helical and 19.4% β-strand content (Figure 1). A secondary structure prediction, based on CD data for separately purified monomeric hSpt4 [29], yielded a similar structural distribution as observed in the heterodimer. It consists of two β-sheets of three and two antiparallel β-strands (β1′–β3′, β4′–β5′) respectively. Both β-sheets are packed orthogonally against each other and face towards the dimer interface where β4′ is directly involved in the interaction with hSpt5-NGN. The solvent-exposed side of hSpt4 is exclusively α-helical with three short α helices (α2′–α4′) and α7′ (Figures 3A and 3B). Additionally, in the hSpt4 centre, a Zn2+ is tetrahedrally co-ordinated by four cysteine residues (Cys16, Cys19, Cys33 and Cys36) forming a zinc finger (Figure 3C). Cysteine residues involved in Zn2+ co-ordination are located in the loop regions between β1′ and β2′ (Cys16 and Cys19) and α2′ and α3′ (Cys33 and Cys36), and presumably fulfill a major structural function by arranging the large helix-loop region, from amino acids 23–54, in close proximity to the β-sheet with β1′ and β2′.

Furthermore, a short 310-helix (α1′) is located at the N-terminus and a second one (α2′) is between β4′ and β5′. In the β-strands β4′ and β5′, a β-bulge interruption resides on equivalent spatial positions forcing both strands to bend away from the interface region. At the C-terminus of hSpt4, the short β-strand β6′ is involved in a two-stranded parallel β-sheet with the second strand, β5, derived from hSpt5-NGN.

Dimerization interface

In the heterodimer, β4′ of hSpt4 aligns antiparallelly with β2 of hSpt5-NGN, leading to a large six-stranded β-sheet that stabilizes the dimer (Figure 3B). Helix α1 of hSpt5-NGN packs against β2 of hSpt5-NGN and β4′ of hSpt4. Non-polar residues point towards the interface, thereby stabilizing the interaction by hydrophobic contacts with hSpt4 (Figure 3D). Thus the central interface comprises predominantly of hydrophobic residues. Only a few polar side chains (e.g. Gln204 and Glu219 of hSpt5-NGN) at the border of the interface also contribute to dimerization.

For the S. cerevisiae homologues Spt4 and Spt5, it was reported that the salt bridge between the conserved Glu338 in Spt5 and Ser58 of Spt4 is crucial for binding, and yeast mutant screens showed that mutation of Glu338 resulted in loss of dimerization [31]. In hSpt5, the corresponding conserved residue Glu228 is located at the C-terminus of β3 and faces towards the N-terminus of α5′ of hSpt4. It forms interactions at hydrogen bond distance with backbone amide protons of residues Val72 and Trp71 and with the hydroxy group of Ser69 of hSpt4 (Figure 3D). To analyse whether the conserved Glu228 in hSpt5 (Glu338 in Spt5 from S. cerevisiae) is also essential for dimerization with hSpt4, we exchanged Glu228 for a glutamine residue and co-expressed wild-type hSpt4 with mutated hSpt5-NGN-E228Q. However, the E228Q mutation obviously destroyed the structural integrity of the complex. In contrast with the wild-type protein, hSpt5-NGN-E228Q was found to be insoluble and could not be co-purified with hSpt4 (results not shown), indicating an important structural role for this residue.

hSpt4–hSpt5-NGN shares an identity with Spt4–Spt5-NGN from S. cerevisiae (PDB: 2EXU) of 37.5% (ClustalW2). The structures of the two complexes are remarkably conserved and superposition of the complexes clearly show identical folding topologies [RMSD (root mean square deviation): 1.36 Å over 860 backbone atoms] (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/425/bj4250373add.htm). The two Spt4 proteins reveal a slightly less conserved tertiary structure as compared with the Spt5 proteins, although the Spt4 proteins share a higher sequence identity (Spt4 proteins: 46%; Spt5 proteins: 28%; ClustalW2). Especially, helices α2′ and α4′ of hSpt4 do not superimpose perfectly with yeast Spt4, and hSpt4 harbours 10 additional N-terminal residues. Except for the short 310-helix α1′, these residues show no significant secondary structural elements (see Supplementary Figure S1). However, residues Lys3 and Val6 of the N-terminal region of hSpt4 are involved in hydrophobic interactions with hSpt5-NGN, which is not the case in the S. cerevisiae complex. These interactions probably stabilize the human DSIF complex.

Moreover, in S. cerevisiae Spt5-NGN, helix α3 is replaced by a long β-strand which superimposes at its C-terminus with β4 of the hSpt5-NGN. The 310 helices α4 and α5 in hSpt5-NGN are substituted for a single long helix enclosing both hSpt5-NGN helices in yeast Spt5-NGN (see Supplementary Figure S1). In addition, the small β5-strand at the C-terminus of hSpt5-NGN is missing in Spt5-NGN of S. cerevisiae.

Comparison of hSpt5-NGN with E. coli NusG-NTD

The physiological function of the NGN of DSIF has not yet been identified. In addition to the NGN, NusG proteins from bacteria and archaea also contain KOW motifs, supporting the hypothesis that hSpt5 is the eukaryotic counterpart of NusG.

Comparison of hSpt5-NGN with E. coli NusG-NTD (PDB: 2K06) [46] does indeed reveal structural homology [RMSD: 3.1 Å over 386 backbone atoms, sequence identity: 16% (ClustalW2)] (Figure 4A). The two proteins exhibit identical folding topologies, but also possess unique characteristics such as the 310 helices α4 and α5 in hSpt5-NGN, which are replaced by a single long C-terminal helix in E. coli NusG-NTD. The small β5 of hSpt5 is lacking in E. coli NusG-NTD and the major β-sheet of hSpt5-NGN also does not superimpose perfectly with its E. coli counterpart. Additionally, discrepancies in secondary structure elements can be seen for helix α3 of hSpt5-NGN, which is substituted by a β-strand in E. coli NusG-NTD. In contrast, the small β-strand (β4) in hSpt5-NGN is missing in E. coli NusG-NTD. Similar structural differences can be detected between human and yeast Spt5-NGN, indicating that the structures of NTDs and NGNs are more conserved between prokaryotes and S. cerevisiae than between S. cerevisiae and humans.

Comparison of hSpt4–hSpt5-NGN with E. coli NusG-NTD and archaeal RpoE”

Figure 4
Comparison of hSpt4–hSpt5-NGN with E. coli NusG-NTD and archaeal RpoE”

(A) Superposition of hSpt5-NGN with E. coli NusG-NTD. hSpt5-NGN is shown in green, E. coli NusG-NTD (PDB: 2K06) in brown; secondary structure elements mentioned in the text are indicated. (B) Electrostatic surface potential of hSpt5-NGN with E. coli NusG-NTD. Structures are displayed in identical orientations. The conserved Glu228 in hSpt5-NGN and the corresponding Gln72 in E. coli NusG-NTD are shown as yellow sticks in the ribbon diagrams. The electrostatic surface potential mapped on the surfaces of hSpt5-NGN and E. coli NusG-NTD at a contour level of ±20 kT is presented below; negative potentials are shown in red, positive potentials in blue and neutral potentials in white; Glu228 in hSpt5-NGN and Gln72 in E. coli NusG-NTD are indicated by a dotted circle. (C) Superposition of hSpt4 with RpoE”. Ribbon diagrams of hSpt4 (blue) and RpoE” from P. furiosus (PDB: 1RYQ) (orange); cysteine residues involved in zinc co-ordination are shown as sticks with the sulphur atoms in yellow. The zinc ion of hSpt4 is displayed as a grey sphere. The zinc ion of RpoE” is not shown, but is located at a similar position; cysteine residue numbers of RpoE” are in italics. α4 is not visible on the left ribbon digram.

Figure 4
Comparison of hSpt4–hSpt5-NGN with E. coli NusG-NTD and archaeal RpoE”

(A) Superposition of hSpt5-NGN with E. coli NusG-NTD. hSpt5-NGN is shown in green, E. coli NusG-NTD (PDB: 2K06) in brown; secondary structure elements mentioned in the text are indicated. (B) Electrostatic surface potential of hSpt5-NGN with E. coli NusG-NTD. Structures are displayed in identical orientations. The conserved Glu228 in hSpt5-NGN and the corresponding Gln72 in E. coli NusG-NTD are shown as yellow sticks in the ribbon diagrams. The electrostatic surface potential mapped on the surfaces of hSpt5-NGN and E. coli NusG-NTD at a contour level of ±20 kT is presented below; negative potentials are shown in red, positive potentials in blue and neutral potentials in white; Glu228 in hSpt5-NGN and Gln72 in E. coli NusG-NTD are indicated by a dotted circle. (C) Superposition of hSpt4 with RpoE”. Ribbon diagrams of hSpt4 (blue) and RpoE” from P. furiosus (PDB: 1RYQ) (orange); cysteine residues involved in zinc co-ordination are shown as sticks with the sulphur atoms in yellow. The zinc ion of hSpt4 is displayed as a grey sphere. The zinc ion of RpoE” is not shown, but is located at a similar position; cysteine residue numbers of RpoE” are in italics. α4 is not visible on the left ribbon digram.

Moreover, the loop between β2 and β3 is truncated in hSpt5-NGN (Figure 4A). Structural comparisons showed that in archaeal NusG proteins Glu338 of Spt5 is also conserved [31], whereas several bacterial NusG proteins, similar to the one from E. coli, harbour a glutamine residue at this position. The physiological function of the corresponding glutamate residue in archaea has not yet been identified. However, sequence comparisons with the DSIF complex from S. cerevisiae suggested that the conserved glutamate might play a role in the interaction of archaeal NusG with RpoE” [31].

Since we can show that the bacterial NusG is structurally similar to hSpt5, we analysed whether the NTD domain of E. coli NusG interacts with hSpt4. However, the co-expression experiments and NMR titrations (see Supplementary Figure S2A at http://www.BiochemJ.org/bj/425/bj4250373add.htm) we performed with hSpt4 and E. coli NusG-NTD did not suggest any interactions.

Since E. coli NusG possesses a glutamine residue (Gln72) at the position corresponding to the conserved Glu228 of hSpt5, we mutated Gln72 of NusG-NTD to a glutamate residue to make it more similar to its eukaryotic and archaeal homologues. We then tested if this amino acid exchange is sufficient for interaction with hSpt4. NMR titration experiments with15N-labelled hSpt4 and unlabelled E. coli NusG-NTD-Q72E showed no significant chemical shift changes in the HSQC spectrum of hSpt4 (see Supplementary Figure S2B), implying that even with the mutated protein no interaction occurs. This result indicates that the conserved glutamate in eukaryotes and archaea cannot by itself account for the interaction, even in the context of high overall structural similarity.

Comparison of the electrostatic surfaces of hSpt5-NGN and E. coli wild-type NusG-NTD revealed (Figure 4B) that in E. coli NusG-NTD negative charges predominate in the region corresponding to the dimerization interface, whereas in hSpt5, positive and negative charges appear to be more evenly distributed. These differences could also influence the ability to interact.

It has been suggested that a hydrophobic surface on NusG mediates binding to E. coli RNA polymerase [4648]. The region on NusG is located opposite to the surface which corresponds to the Spt4-binding interface of the yeast Spt5-NGN. A hydrophobic surface region is also present in hSpt5 at a similar position, although it is smaller and interrupted by hydrophilic residues (see Supplementary Figure S3 at http://www.BiochemJ.org/bj/425/bj4250373add.htm)

hSpt4 and RpoE” share core elements

hSpt4 shows distant structural homology to the archaeal RpoE” from Pyrococcus furiosus (PDB: 1RYQ) (Figure 4C) [30,31]. Albeit much smaller, structure alignment with hSpt4 reveals topological similarities for the zinc finger and the corresponding hSpt5-interaction region [RMSD: 1.6 Å, over 341 backbone atoms; sequence identity 23% (ClustalW2)]. The three helical elements α2′, α3′ and α4′ of hSpt4 are exchanged for unstructured loop regions, which are stabilized by the zinc finger formed by the four cysteine residues Cys6, Cys9, Cys18 and Cys21 of RpoE”. Apart from Cys18, these cysteines are arranged in the same relative orientation as the corresponding cysteine residues of hSpt4. Both proteins share the β-sheet ensemble, with one two-stranded antiparallel sheet and a second small three-stranded sheet arranged perpendicularly to each other.

Consequently, RpoE” contains all the basic and important elements for stability and interaction with NGNs. Potential binding of the two proteins was suggested by docking experiments of RpoE” with the crystal structure of NusG from Methanocaldococcus jannaschii [49] and GST pull-down experiments detected interaction of RpoE” with NusG from M. jannaschii [31]. The role of RpoE” or Spt4 proteins is still not fully understood. However, despite the structural homology between hSpt4 and RpoE”, this may not necessarily translate into similar functional roles or even a similar localization or interface with the RNA polymerase. So far, all interactions of DSIF with other transcription factors or RNA polymerases appear to be mediated via Spt5 proteins. hSpt4 might thus facilitate interaction of Spt5 with other factors or could simply stabilize the NGN of hSpt5.

Abbreviations

     
  • DSIF

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

  •  
  • GB1

    streptococcal immunoglobulin-binding domain of protein G

  •  
  • hSpt5-NGN

    human Spt5(176–273)

  •  
  • HSQC

    heteronuclear single-quantum coherence

  •  
  • KOW motif

    Kyrpides, Ouzounis and Woese motif

  •  
  • MAD

    multi-wavelength anomalous dispersion

  •  
  • NELF

    negative elongation factor

  •  
  • NGN

    NusG N-terminal homology domain

  •  
  • Ni-NTA

    Ni2+-nitrilotriacetate

  •  
  • NTD

    N-terminal domain

  •  
  • P-TEFb

    positive transcription elongation factor b

  •  
  • RMSD

    root mean square deviation

  •  
  • RNAPol II

    RNA polymerase II

  •  
  • SAD

    single-wavelength anomalous dispersion

  •  
  • TEV

    tobacco etch virus

AUTHOR CONTRIBUTION

Birgitta Wöhrl conceived and co-ordinated the study. Sabine Wenzel performed all the experiments and participated in data analysis. Berta Martins participated in designing the experiments and performed the structure data analysis. Sabine Wenzel, Birgitta Wöhrl and Berta Martins wrote the paper. Paul Rösch provided conceptual input and critical advice. All authors read and approved the manuscript.

We thank Britta Zimmermann for excellent technical assistance. We acknowledge the staff of beamline BL14.2 (BESSY, Berlin, Germany) for support during data collection.

FUNDING

The project was funded by grants from the Deutsche Forschungsgemeinschaft (DFG) [grant number Ro617/16–1] and the Study Program-Macromolecular Science from the Elite Network of Bavaria. B.M.M. thanks the Fonds of the Chemical Industry (FCI) for financial support.

References

References
1
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
)
2
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
)
3
Ivanov
D.
Kwak
Y. T.
Guo
J.
Gaynor
R. B.
Domains in the SPT5 protein that modulate its transcriptional regulatory properties
Mol. Cell. Biol.
2000
, vol. 
20
 (pg. 
2970
-
2983
)
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
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
)
6
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
)
7
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
)
8
Yamada
T.
Yamaguchi
Y.
Inukai
N.
Okamoto
S.
Mura
T.
Handa
H.
P-TEFb-mediated phosphorylation of hSpt5 C-terminal repeats is critical for processive transcription elongation
Mol. Cell
2006
, vol. 
21
 (pg. 
227
-
237
)
9
Lis
J. T.
Mason
P.
Peng
J.
Price
D. H.
Werner
J.
P-TEFb kinase recruitment and function at heat shock loci
Genes Dev.
2000
, vol. 
14
 (pg. 
792
-
803
)
10
Andrulis
E. D.
Guzman
E.
Doring
P.
Werner
J.
Lis
J. T.
High-resolution localization of Drosophila Spt5 and Spt6 at heat shock genes in vivo: roles in promoter proximal pausing and transcription elongation
Genes Dev.
2000
, vol. 
14
 (pg. 
2635
-
2649
)
11
Wu
C. H.
Yamaguchi
Y.
Benjamin
L. R.
Horvat-Gordon
M.
Washinsky
J.
Enerly
E.
Larsson
J.
Lambertsson
A.
Handa
H.
Gilmour
D.
NELF and DSIF cause promoter proximal pausing on the hsp70 promoter in Drosophila
Genes Dev.
2003
, vol. 
17
 (pg. 
1402
-
1414.
)
12
Lindstrom
D. L.
Squazzo
S. L.
Muster
N.
Burckin
T. A.
Wachter
K. C.
Emigh
C. A.
McCleery
J. A.
Yates
J. R.
3rd
Hartzog
G. A.
Dual roles for Spt5 in pre-mRNA processing and transcription elongation revealed by identification of Spt5-associated proteins
Mol. Cell. Biol.
2003
, vol. 
23
 (pg. 
1368
-
1378
)
13
Mandal
S. S.
Chu
C.
Wada
T.
Handa
H.
Shatkin
A. J.
Reinberg
D.
Functional interactions of RNA-capping enzyme with factors that positively and negatively regulate promoter escape by RNA polymerase II
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
7572
-
7577
)
14
Sims
R. J.
III
Belotserkovskaya
R.
Reinberg
D.
Elongation by RNA polymerase II: the short and long of it
Genes Dev.
2004
, vol. 
18
 (pg. 
2437
-
2468
)
15
Krogan
N. J.
Kim
M.
Ahn
S. H.
Zhong
G.
Kobor
M. S.
Cagney
G.
Emili
A.
Shilatifard
A.
Buratowski
S.
Greenblatt
J. F.
RNA polymerase II elongation factors of Saccharomyces cerevisiae: a targeted proteomics approach
Mol. Cell. Biol.
2002
, vol. 
22
 (pg. 
6979
-
6992
)
16
Wada
T.
Takagi
T.
Yamaguchi
Y.
Ferdous
A.
Imai
T.
Hirose
S.
Sugimoto
S.
Yano
K.
Hartzog
G. A.
Winston
F.
, et al. 
DSIF, a novel transcription elongation factor that regulates RNA polymerase II processivity, is composed of human Spt4 and Spt5 homologs
Genes Dev.
1998
, vol. 
12
 (pg. 
343
-
356
)
17
Sullivan
S. L.
Gottesman
M. E.
Requirement for. E. coli NusG protein in factor-dependent transcription termination
. Cell
1992
, vol. 
68
 (pg. 
989
-
994
)
18
Cardinale
C. J.
Washburn
R. S.
Tadigotla
V. R.
Brown
L. M.
Gottesman
M. E.
Nudler
E.
Termination factor rho and its cofactors NusA and NusG silence foreign DNA in
E coli. Science
2008
, vol. 
320
 (pg. 
935
-
938.
)
19
DeVito
J.
Das
A.
Control of transcription processivity in phage lambda: Nus factors strengthen the termination-resistant state of RNA polymerase induced by N antiterminator
Proc. Natl. Acad. Sci. U.S.A.
1994
, vol. 
91
 (pg. 
8660
-
8664
)
20
Torres
M.
Balada
J. M.
Zellars
M.
Squires
C.
Squires
C. L.
In vivo effect of NusB and NusG on rRNA transcription antitermination
J. Bacteriol.
2004
, vol. 
186
 (pg. 
1304
-
1310
)
21
Yamaguchi
Y.
Wada
T.
Watanabe
D.
Takagi
T.
Hasegawa
J.
Handa
H.
Structure and function of the human transcription elongation factor DSIF
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
8085
-
8092
)
22
Ponting
C. P.
Novel domains and orthologues of eukaryotic transcription elongation factors
Nucleic Acids Res.
2002
, vol. 
30
 (pg. 
3643
-
3652
)
23
Reference deleted
24
Schwer
B.
Schneider
S.
Pei
Y.
Aronova
A.
Shuman
S.
Characterization of the Schizosaccharomyces pombe Spt5-Spt4 complex
RNA
2009
, vol. 
15
 (pg. 
1241
-
1250
)
25
Hartzog
G. A.
Wada
T.
Handa
H.
Winston
F.
Evidence that Spt4, Spt5, and Spt6 control transcription elongation by RNA polymerase II in Saccharomyces cerevisiae
Genes Dev.
1998
, vol. 
12
 (pg. 
357
-
369
)
26
Kyrpides
N. C.
Woese
C. R.
Ouzounis
C. A.
KOW: a novel motif linking a bacterial transcription factor with ribosomal proteins
Trends Biochem. Sci.
1996
, vol. 
21
 (pg. 
425
-
426
)
27
Kwak
Y. T.
Guo
J.
Prajapati
S.
Park
K. J.
Surabhi
R. M.
Miller
B.
Gehrig
P.
Gaynor
R. B.
Methylation of SPT5 regulates its interaction with RNA polymerase II and transcriptional elongation properties
Mol. Cell
2003
, vol. 
11
 (pg. 
1055
-
1066
)
28
Chiang
P. W.
Wang
S. Q.
Smithivas
P.
Song
W. J.
Crombez
E.
Akhtar
A.
Im
R.
Greenfield
J.
Ramamoorthy
S.
Van Keuren
M.
, et al. 
Isolation and characterization of the human and mouse homologues (SUPT4H and Supt4h) of the yeast SPT4 gene
Genomics
1996
, vol. 
34
 (pg. 
368
-
375
)
29
Wenzel
S.
Schweimer
K.
Rösch
P.
Wöhrl
B. M.
The small hSpt4 subunit of the human transcription elongation factor DSIF is a Zn-finger protein with α/β type topology
Biochem. Biophys. Res. Commun.
2008
, vol. 
370
 (pg. 
414
-
418
)
30
Werner
F.
Structure and function of archaeal RNA polymerases
Mol. Microbiol.
2007
, vol. 
65
 (pg. 
1395
-
1404
)
31
Guo
M.
Xu
F.
Yamada
J.
Egelhofer
T.
Gao
Y.
Hartzog
G. A.
Teng
M.
Niu
L.
Core structure of the yeast Spt4-Spt5 complex: A conserved module for regulation of transcription elongation
Structure
2008
, vol. 
16
 (pg. 
1649
-
1658
)
32
Mori
S.
Abeygunawardana
C.
Johnson
M. O.
van Zijl
P. C.
Improved sensitivity of HSQC spectra of exchanging protons at short interscan delays using a new fast HSQC (FHSQC) detection scheme that avoids water saturation
J. Magn. Reson. B
1995
, vol. 
108
 (pg. 
94
-
98
)
33
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
)
34
Kabsch
W.
Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants
J. Appl. Chrystallogr.
1993
, vol. 
26
 (pg. 
795
-
800
)
35
Sheldrick
G. M.
A short history of SHELX
Acta Crystallogr. A
2008
, vol. 
64
 (pg. 
112
-
122
)
36
Morris
R. J.
Perrakis
A.
Lamzin
V. S.
ARP/wARP and automatic interpretation of protein electron density maps
Methods Enzymol.
2003
, vol. 
374
 (pg. 
229
-
244
)
37
Emsley
P.
Cowtan
K.
COOT: Model-building tools for molecular graphics
Acta Crystallogr. D Biol. Crystallogr.
2004
, vol. 
60
 (pg. 
2126
-
2132
)
38
Murshudov
G. N.
Vagin
A. A.
Dodson
E. J.
Refinement of macromolecular structures by the maximum-likelihood method
Acta Crystallogr. D Biol. Crystallogr.
1997
, vol. 
53
 (pg. 
240
-
255
)
39
Collaborative Computational Project, Number 4
The CCP4 suite: programs for protein crystallography
Acta Crystallogr. D Biol. Crystallogr.
1994
, vol. 
50
 (pg. 
760
-
763
)
40
Reference deleted
41
Baker
N. A.
Sept
D.
Joseph
S.
Holst
M. J.
McCammon
J. A.
Electrostatics of nanosystems: application to microtubules and the ribosome
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
10037
-
10041
)
42
Laskowski
R. A.
MacArthur
M. W.
Moss
D. S.
Thornton
J. M.
PROCHECK: A program to check the stereochemical quality of protein structures
J. Appl. Cryst.
1993
, vol. 
26
 (pg. 
283
-
291
)
43
Larkin
M. A.
Blackshields
G.
Brown
N. P.
Chenna
R.
McGettigan
P. A.
McWilliam
H.
Valentin
F.
Wallace
I. M.
Wilm
A.
Lopez
R.
, et al. 
Clustal W and clustal X version 2.0
Bioinformatics
2007
, vol. 
23
 (pg. 
2947
-
2948
)
44
Wallace
A. C.
Laskowski
R. A.
Thornton
J. M.
LIGPLOT: A program to generate schematic diagrams of protein-ligand interactions
Protein Eng.
1995
, vol. 
8
 (pg. 
127
-
134
)
45
Krissinel
E.
Henrick
K.
Inference of macromolecular assemblies from crystalline state
J. Mol. Biol.
2007
, vol. 
372
 (pg. 
774
-
797
)
46
Mooney
R. A.
Schweimer
K.
Rösch
P.
Gottesman
M. E.
Landick
R.
Two structurally independent domains of. E. coli NusG create regulatory plasticity via distinct interactions with RNA polymerase and regulators
J. Mol. Biol.
2009
, vol. 
391
 (pg. 
341
-
358
)
47
Belogurov
G. A.
Vassylyeva
M. N.
Svetlov
V.
Klyuyev
S.
Grishin
N. V.
Vassylyev
D. G.
Artsimovitch
I.
Structural basis for converting a general transcription factor into an operon-specific virulence regulator
Mol. Cell.
2007
, vol. 
26
 (pg. 
117
-
129
)
48
Nickels
B. E.
Genetic assays to define and characterize protein-protein interactions involved in gene regulation
Methods
2009
, vol. 
47
 (pg. 
53
-
62
)
49
Zhou
H.
Liu
Q.
Gao
Y.
Teng
M.
Niu
L.
Crystal structure of NusG N-terminal (NGN) domain from Methanocaldococcus jannaschii and its interaction with RpoE”
Proteins
2009
, vol. 
76
 (pg. 
787
-
793
)

Author notes

The structural co-ordinates reported for the human transcription elongation factor DSIF [DRB (5,6-dichloro-1-β-D-ribofuranosylbenzimidazole) sensitivity-inducing factor], hSpt4/hSpt5 (176–273) will appear in the Protein Data Bank under accession code 3H7H.

Supplementary data