The [15N,1H]-HSQC data [Fig. 3(B–D)] indicate that the PAI subdomain is the primary sequence of SB transposase that binds to DR-core and DRi DNA. Therefore, we studied the DNA-binding properties of an isolated PAI subdomain (residues 1–57) of the SB DBD.
NMR solution structure of PAI subdomain
Based on the amino acid sequence similarity to Pax proteins and Tc3 transposase, it was proposed that the PAI subdomain consists of three helices, two of which form the HTH motif.[2, 14] Although secondary structure predictions suggested that more than half of the PAI subdomain is alpha-helical [Fig. 2(C)], our NMR data on the full-length DBD show that it is mostly unstructured in solution. The far-UV CD spectrum of the PAI subdomain collected at the same experimental conditions (20 mM sodium acetate buffer, pH 5.2), also lacks the typical signatures of alpha-helical structure (two negative bands at 208 nm and 222 nm and a positive band at 193 nm) and shows a negative band near 200 nm and low ellipticity above 210 nm, indicating that the protein is predominantly unstructured [Fig. 4(A), dashed line] in agreement with our NMR data on the full-length SB DBD. We found that the folding of the PAI subdomain can be achieved by raising the NaCl concentrations (up to 600 mM) and increasing pH to 7.0. At these conditions, the CD spectrum of PAI subdomain reflects a mixture of disordered and alpha-helical conformations [Fig. 4(A), solid line]. Fitting the CD data with a linear combination of disordered and alpha-helical basis spectra indicated the presence of about 50% alpha-helical structure. Chemical shift changes following the folding of PAI subdomain are exemplified for some residues in Figure 4(B). The [15N,1H]-HSQC spectrum displays well-dispersed cross-peaks, which signifies the presence of a folded structure [Fig. 5(A)]. The backbone and side chain assignments for the PAI subdomain were derived from 2D and 3D 15N-HSQC TOCSY and 15N-HSQC NOESY experiments using a 15N-labeled PAI subdomain. A consensus CSI of Hα protons [Fig. 5(B)] indicates three alpha-helical regions (residues 12–22, 29–33, and 39–55) within the PAI subdomain that are consistent with secondary structure predictions [Fig. 2(C)].
Figure 4. A: Far-UV CD spectra of the PAI subdomain of the DNA-binding domain of SB transposase. The spectra were collected at room temperature in 20 mM sodium acetate buffer at pH 5.2 (dashed line) and in the presence of 600 mM NaCl in 25 mM sodium phosphate buffer at pH 7.0 (solid line). CD spectra show that the PAI subdomain is folded at the latter conditions. B: Expansion of the [15N,1H]-HSQC spectra of unfolded (light blue) and folded (black) PAI subdomain exemplifies observed chemical shift changes. The spectra were collected at 5°C at the same buffer conditions as CD spectra shown in panel (A).
Download figure to PowerPoint
Figure 5. A: Assigned [15N,1H]-HSQC spectrum of the PAI subdomain from the DNA-binding domain of SB transposase. The spectrum was recorded at 5°C in 25 mM sodium phosphate buffer at pH 7.0. B: The chemical shift index (CSI) obtained using Hα chemical shifts of PAI subdomain. Stretches of random coil (0) and helices (−1) are shown below the graph.
Download figure to PowerPoint
It was proposed that the PAI subdomain participates in protein–protein interactions when a protein–DNA complex is formed. However, during size-exclusion chromatography under the experimental conditions used in our experiments, the PAI subdomain elutes significantly later than the 13.8 kDa RNase A, as can be expected for a 7.5 kDa monomer. The structure of the PAI subdomain monomer was calculated as described in “Materials and Methods” and the structural statistics are provided in Table 1. Cα traces for alpha-helix-aligned structures of the representative ensemble are shown superimposed in Figure 6(left panel) and a ribbon representation of a representative structure is shown in Figure 6(right panel). Coordinates for the 15 lowest energy structures were deposited in the Protein Data Bank (PDB) under accession code 2m8e; BMRB ID for this entry is 19249. The PAI subdomain folds into a compact, three-helix domain. Helices 2 and 3 are connected by the loop of four residues and form a HTH motif.
Table 1. Structural Statistics and Restraint Information for the NMR Structure of the PAI Subdomain
|Restraints and statistics|| |
|NOE distance restraints (total)||418|
|Average number of NOE restraints per residue||7.4|
|Medium and long-range NOEs (j − i >3)||130|
|Residual dipolar couplings (Hz)||28|
|TALOS derived dihedral angle restraints||51|
|NOE distance violations >0.5 Å||0|
|Dihedral angle violations >5°||0|
|RMS deviation from mean structure (Å)|| |
|All heavy atoms||1.4|
|Ramachandran statisticsa (%)|| |
|Most favored region||96.7|
|Additionally allowed region||2.8|
|Generously allowed region||0.5|
Figure 6. NMR solution structure of the PAI subdomain. Cα traces of superimposed 15 lowest energy structures (left) and the cartoon representation of the representative structure of the PAI subdomain (right) are shown.
Download figure to PowerPoint
Interactions of PAI subdomain with DNA
We initially used an electrophoretic mobility shift assay (EMSA) to verify that the expressed protein binds to DNA. Figure 7(A) shows the EMSA analysis of PAI subdomain binding to the 18-bp core sequence (left) and 32-bp inner DRi (right). In both cases, the PAI subdomain was capable of binding the DNA producing a distinct gel shift. Two bands, corresponding to the DR-core or DRi free (bottom) or bound to the PAI subdomain (top), were observed. No distinct bands corresponding to higher order complexes were detected at the protein:DNA ratios studied.
Figure 7. Binding of the PAI subdomain of the DNA-binding domain of SB transposase to DR-core (left panel) and DRi (right panel) DNA sequences. A: EMSA shows that the PAI subdomain binds to the DR-core and DRi DNA. Protein–DNA complexes were formed in 20 mM HEPES buffer at pH 7.5 in the presence of 0.1 mg/mL BSA, 1 µg [di:dC], 1 mM DTT, 150 mM NaCl, and 1 mM MgCl2. Probe quantity was fixed at 0.21 pM and the quantities of PAI subdomain varied from 0 to 2.24 nM as indicated by triangles above gels. B: [15N,1H]-HSQC spectra of pure PAI subdomain (blue) and of the PAI subdomain in the presence of a twofold molar excess of DR-core and DRi DNA sequences. The spectra were collected at 5°C in aqueous (5% D2O/95% H2O) 25 mM sodium phosphate buffer at pH 7.0 in the presence of 300 mM NaCl. C: Cartoon structures of PAI subdomain show residues affected by the binding of DR-core (left) and DRi (right) DNA sequences in red.
Download figure to PowerPoint
The binding of the PAI subdomain to the DR-core and DRi DNA sequences was further investigated by recording the [15N,1H]-HSQC spectra at a 1:2 PAI:DNA molar ratio because these sequences preferentially bind the PAI subdomain in the full-length DBD of SB transposase. The initial set of experiments was carried out in the presence of 600 mM NaCl to maintain the PAI subdomain in folded form. However, only negligibly small changes in [15N,1H]-HSQC spectrum of PAI subdomain were observed that reflected essentially no binding (data not shown), which was likely due to the presence of the high salt concentration. Accordingly, binding experiments were conducted at the intermediate, 300 mM NaCl. Under this condition, spectral changes were indicative of binding of the PAI subdomain to the DR-core [Fig. 7(B), left panel] and DRi [Fig. 7(B), right panel]; significant line broadening and chemical shift changes were observed for a set of residues.
Three distinct processes, namely, the conformational rearrangement of PAI subdomain in the presence of DNA, exchange between DNA-bound and unbound states of the PAI subdomain, and the dimerization (or higher order oligomerization) of PAI-DNA complexes, could cause spectral line broadening if they follow intermediate exchange kinetics on an NMR timescale. Varying the temperature, pH, and solvent systems failed to improve spectrum quality significantly. Consequently, we could not determine the full NMR structure of the PAI-DNA complex. To gain structural insight in PAI-DNA binding, we labeled the residues that demonstrated severely broadened NMR signals (absent from [15N,1H]-HSQC spectrum) on the structure of PAI subdomain [highlighted in red in Fig. 7(C), left panel]. The majority of these residues (28, 29, 31, 33–36, 38–43, 47) are located on the second and third alpha helices and on the loop connecting these two helices of the HTH motif. Accordingly, we predict that these residues form the DNA-binding site of PAI subdomain. Figure 8 shows that this prediction is consistent with the expected location of the DNA-binding site based on the comparison with available structures of the DNA-bound form of the two closest family members, Tc3 and Mos1 transposases,[30-32] and with the paired box transcription factors, with which the SB PAI subdomain has amino acid sequence similarity. The structure of Pax5 transcription factor has the highest amino acid sequence similarity to the SB PAI subdomain, as reflected in the similarity of their structures.
Figure 8. Structures of the PAI subdomain of SB transposase and the DNA-bound N-terminal specific DNA-binding subdomains of Tc3 transposase (PDB codes 1tc3 and 1u78, Refs.  and ), Mos1 transposase (PDB code 3hos, Ref. ), and Pax5 transcription factor (PDB code 1k78, Ref. ). Protein structures were superimposed and then merely shifted along x (horizontal) axis. Residues on the second and third alpha-helices of SB PAI subdomain affected by DNA-binding are shown in black. Truncated DNA sequences from the PDB structures are shown in black to highlight the DNA-binding site of Mos1, Tc3, and Pax5 proteins.
Download figure to PowerPoint
Other residues affected by DNA-binding are located on the first helix. Interestingly, NMR signals of these residues broaden similarly upon either the addition of DNA or at decreasing the NaCl concentration, which suggests similar conformational folding/unfolding of the PAI subdomain in both cases. Indeed, the binding experiments were done at intermediate NaCl concentrations at which only part of the PAI subdomain molecules may adopt the fully structured conformation. Several N-terminal residues (3, 7, 9, and 10) demonstrate severely broadened signals. These residues are located on the unstructured and highly flexible N-terminus of the PAI subdomain and may be affected by binding to DNA. Additionally, the dimerization/oligomerization of the PAI-DNA complexes could be expected because the chemical steps of SB transposition occur within a stable protein–DNA complex. The minimal PAI-DNA complex would be the dimer of protein with the two DNA molecules bound, corresponding to a 37 kDa complex for a 18-bp DR-core sequence or 50 kDa complex for a 32-bp DRi or DRo sequences. The formation of stable protein-DNA complexes of this size would lead to the broadening of all peaks in the HSQC spectra, which is not the case. However, the contribution of the transient protein–protein interactions to the broadening of signals originated from the first alpha helix in addition to the folding/unfolding of PAI subdomain remains a possibility based on comparison with Mos1 and Tc3 transposases.[30-32]
We note that the spectra of the PAI subdomain bound to either the 32-bp DRi or the 18-bp DR-core sequences are highly similar. Thus, the DR-core is the primary binding site of PAI-subdomain. Additional residues demonstrated broadened signals due to the binding of DRi are circled on [15N,1H]-HSQC spectrum of PAI subdomain [Fig. 7(B), right panel] and indicated on the PAI structure in red [Fig. 7(C), right panel]. Some of these residues are clustered on the HTH motif; however, residues 11, 16, and 18 are located on helix-1. Although the binding to DRi causes greater changes on helix-1 than the binding to DR-core, only 1:1 complexes of PAI-DNA are formed as indicated by EMSA. Thus, these greater changes upon DRi binding are again related to the conformational exchange due to folding.