HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wojtuszewski, K. Articles by Mukerji, I. Search for Related Content PubMed PubMed Citation Articles by Wojtuszewski, K. Articles by Mukerji, I. Social Bookmarking                   What's this?

The HU–DNA binding interaction probed with UV resonance Raman spectroscopy: Structural elements of specificity

Department of Molecular Biology and Biochemistry, Molecular Biophysics Program, Wesleyan University, Middletown, Connecticut 06459-0175, USA

Reprint requests to: Ishita Mukerji, Department of Molecular Biology and Biochemistry, Molecular Biophysics Program, Wesleyan University, Middletown, CT 06459-0175, USA; e-mail: imukerji{at}wesleyan.edu; fax: (860) 685-2141.

(RECEIVED March 9, 2004; FINAL REVISION May 7, 2004; ACCEPTED May 18, 2004)

	Abstract

TOP Abstract Introduction Results Discussion Materials and methods References

 
The Escherichia coli protein HU functions as an architectural DNA-binding protein by facilitating DNA looping or bending to form multiprotein complexes. Although HU does not recognize a specific DNA sequence, site-specific binding to a number of discontinuous, looped, or bent DNA substrates has been observed. In this study UV resonance Raman (UVRR) spectroscopy is used to identify structural elements associated with low- and high-affinity binding by examining three different HU–DNA complexes. UVRR spectra obtained with an excitation wavelength of 210 nm, which preferentially enhances protein backbone amide vibrations, indicate that HU secondary structure content increases and the protein structure becomes more rigid upon binding to DNA. The increase in -helical content is attributed to the C-terminal helix, which interacts with the DNA and may play a role in binding affinity and specificity. UVRR spectra obtained with a 215 nm excitation wavelength demonstrate that Pro mode intensity at 1455 cm^–1 decreases upon complex formation. This intensity decrease is attributed to the intercalation of Pro residues between DNA base pairs to induce a bend in the DNA, as has been observed previously in the IHF–DNA and HU–DNA cocrystal structures. DNA vibrational modes are also indicative of significant base unstacking and opening of the minor groove upon protein binding, consistent with bending and distortion of the DNA. In all three complexes, A-DNA conformational features are indicated by deoxyribose-phosphate backbone modes. These and other results suggest that protein-induced bending plays an important role in HU site-specific binding and supports a model of a mutually induced fit.

Keywords: protein; DNA interactions; A-tracts; Raman spectroscopy; DNA bending; DNA binding

Abbreviations: bp, base pair • IHF, integration host factor • FRET, fluorescence resonance energy transfer • UVRR, UV resonance Raman spectroscopy.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.04730204.

	Introduction

TOP Abstract Introduction Results Discussion Materials and methods References

 
HU, a multifunctional DNA-binding protein, participates in several cellular functions including growth, recombination, replication and transposition (Drlica and Rouviere-Yaniv 1987). Originally identified as a histone-like protein (Rouviere-Yaniv and Gros 1975), more recent studies implicate HU as a DNA chaperone (Travers 1993) or architectural protein. HU binding mediates DNA looping or bending, which facilitates the formation of multiprotein complexes (Lavoie and Chaconas 1993; Lavoie et al. 1996; Aki and Adhya 1997; Lia et al. 2003). Although HU protein participates in specific functions within the cell and binds to specific DNA sites, as in the case of Gal repression and Mu transposition, it does not exhibit sequence specific recognition of DNA.

HU, a heterodimer in Escherichia coli, binds with high affinity and 1:1 stoichiometry (HU dimer:DNA duplex) to discontinuous or bent DNA conformations, such as gapped or nicked DNA; whereas the binding of HU to linear duplex DNA is characterized by micromolar affinities and protein:DNA ratios greater than 1:1 (Castaing et al. 1995; Lavoie et al. 1996; Kamashev et al. 1999; Kobryn et al. 1999; Pinson et al. 1999; Kamashev and Rouviere-Yaniv 2000; Wojtuszewski et al. 2001). The structures of the 1:1 HU–DNA complexes have been successfully modeled using the IHF–DNA cocrystal structure, in which the protein induces a bend of 160° (Lavoie et al. 1996; Rice et al. 1996; Kamashev et al. 1999). The modeling of these complexes indicates that in certain instances HU can bind DNA duplexes site-specifically and induce a bend in a manner similar to IHF. The recent Anabena HU–DNA X-ray crystal structure (Swinger et al. 2003) reveals that the protein-induced bend is variable and not necessarily as large as in the IHF–DNA structure. IHF and HU interact with DNA through some common motifs such as the intercalation of Pro residues to induce bending or kinking of the DNA and interaction of the C terminus with the DNA (Rice et al. 1996; Swinger et al. 2003). In both structures protein-induced bending leads to the wrapping of the DNA around the protein -helical body.

The nanomolar binding affinities observed with flexible or discontinuous DNA substrates and the micromolar binding affinities observed with linear DNA suggests that HU dimer exhibits different modes of DNA binding, which are governed by structure and not sequence, as has been recently described for HMG1 (Murphy and Churchill 2000). Further evidence for an alternate mode of HU binding to DNA has been observed by scanning force microscopy, in which HU binding leads to DNA elongation rather than compaction (Dame and Goosen 2002). The molecular basis that underlies these differences in HU binding interactions is not known.

The goal of this study is to identify structural elements important in binding and recognition by examining three different HU–DNA complexes using UV resonance Raman spectroscopy (UVRR): two higher affinity complexes that exhibit protein-induced bending are compared with a low-affinity complex that exhibits no bending. The three complexes have been previously characterized by fluorescence spectroscopy and analytical ultracentrifugation (Wojtuszewski et al. 2001; Wojtuszewski and Mukerji 2003), and were chosen for further study using UVRR spectroscopy because of the observed differences in binding affinity and the amount of protein-induced bending. In the higher affinity complexes (K_a ~5 µM^–1) the duplexes are 20 bp long and contain an A₄T₄ repeating sequence either in the 5' to 3' direction or in the 3' to 5' direction, hereafter referred to as A4–20 and T4–20, respectively. The orientation of the A-tract leads to a marked difference in structural and energetic properties, which were attributed to DNA bending (Hagerman 1986; Burkhoff and Tullius 1988; Haran and Crothers 1989). Fluorescence resonance energy transfer measurements (FRET) determined a bend angle of 45° and 12° for the A4–20 and T4–20 duplexes and protein binding increased the total bend to 70° and 62°, respectively (Wojtuszewski and Mukerji 2003). Also examined was HU binding to a 13-bp duplex, the sequence of which is based on the IHF H1 consensus binding sequence. HU binds to this sequence with low affinity (0.5 µM^–1), a 2:1 stoichiometry and little to no induced bending, as determined by fluorescence intensity measurements (Wojtuszewski et al. 2001). UVRR spectroscopy is now used to probe the structural elements of these complexes to identify their similarities and differences particularly with respect to binding affinity. This study elucidates molecular details of the HU–DNA interaction, which could not be discerned from the previous fluorescence experiments and provides a framework from which the structural recognition of DNA by HU can be understood.

Previously, UV resonance Raman spectroscopy has been used successfully to analyze proteins (Harada and Takeuchi 1986; Austin et al. 1993) and DNA (Fodor and Spiro 1986; Toyama et al. 1991, 1993; Mukerji et al. 1995, 1996, 1998; Sokolov et al. 2000; Mukerji and Williams 2002), but the technique has only been sparingly applied to the analysis of protein–DNA interactions. In contrast, Raman spectroscopy has been successfully applied to the study of DNA structure and protein–DNA interactions, and many previously identified conformational markers are used to analyze the current data (Thomas and Tsuboi 1993; Peticolas 1995; Thomas 1999). However, conducting the experiment in a resonance enhanced manner affords several advantages including the differential investigation of the protein and DNA moieties in the complex and the ability to use concentrations that are only modestly increased relative to those used to characterize the binding interaction by other methods (Wojtuszewski et al. 2001; Wojtuszewski and Mukerji 2003). In this study, excitation wavelengths of 210 and 215 nm are used to probe HU secondary structure through the protein amide backbone modes, and the local environment of Phe and Pro residues, respectively (Fig. 1). An excitation wavelength of 215 nm also enhances deoxyribose phosphate backbone modes, DNA carbonyl stretching vibrations and ring-stretching vibrations, the intensities of which are a reliable measure of base stacking. Structural details of the complexes can be inferred from changes in frequency or intensity of these modes upon complexation.

View larger version (59K): [in this window] [in a new window]   Figure 1. Anabena HU bound to DNA. HU monomers are shown in gray and black to indicate the heterodimeric nature of E. coli HU. The four Pro residues present in E. coli HU, located at positions 63 and 77 for each monomer, are shown in maroon. The six Phe residues present in E. coli HU at positions, 47, 50, and 79 for each monomer, are shown in green. These residues are highly conserved among all HU proteins and IHF. A portion of a second DNA molecule, present in the asymmetric unit, is shown to illustrate the interaction between the C-terminal helix and the DNA. Diagram was generated using Weblab Viewer Pro 3.7 and X-ray crystallographic coordinates of Swinger et al. (2003), PDB ID: 1P51 [PDB] .

	Results

TOP Abstract Introduction Results Discussion Materials and methods References

 
Complexes formed between HU protein and 3 DNA duplexes (A4–20, T4–20, and H1–13; Table 1) were examined using excitation wavelengths of 210 and 215 nm. Discrimination between different groups of the protein as well as differentiation between DNA and protein contributions is achieved through the use of these two excitation wavelengths. Although many of the same bands are observed in the 210 nm and 215 nm-excited spectra, because of the resonance enhancement effect the 210 nm spectra are used primarily to detect changes in HU secondary structure through the backbone amide modes and the 215 nm-excited spectra reveal changes in DNA conformation and the relative intensities of Phe and Pro bands.

View larger version (27K): [in this window] [in a new window]   Figure 2. Illustration of the method used to generate the spectral comparisons. All spectra shown are excited with an excitation wavelength of 210 nm. (A) HU only spectrum obtained at a protein concentration of 0.14 mM. The resonance-enhanced amide I, II, III, and S modes are indicated. (B) A4–20 only spectrum obtained at a concentration of 0.15 mM. (C) HU:A4–20 complex spectrum in which the concentration of HU is 0.14 mM and the concentration of A4–20 is 0.127 mM. (D) Added spectrum of HU and A4–20 using the spectra shown in (A) and (B). Spectra were normalized to a buffer band at 602 cm–1 and then multiplied by the appropriate factors to match the concentrations used to obtain the complex spectra shown in (C). All of the samples were in a 10 mM Tris buffer at pH 7.6 containing 0.1 mM EDTA and 100 mM NaCl.

View larger version (39K): [in this window] [in a new window]   Figure 3. 210 nm-excited spectra of the HU–DNA complexes. (A) Spectral range from 1125 to 1430 cm–1. (B) Spectral range from 1500 to 1760 cm–1. Mode assignments are discussed in the text and are given in Table 2. (1) HU-only spectrum; (2) HU:A4–20 spectrum obtained under the same conditions as in Figure 2; (3) HU2:T4–20 spectrum obtained with a protein concentration of 0.14 mM and a DNA concentration of 0.132 mM; (4) HU2:H1–13 spectrum obtained with a protein concentration of 0.13 mM and a DNA concentration of 0.112 mM. Buffer conditions were the same as in Figure 2 except that 50 mM NaCl was used for the HU2:H1–13 complex.

View larger version (39K): [in this window] [in a new window]   Figure 4. Relative Raman intensities of amide II and amide S modes. Amide II relative intensities are shown in the solid gray columns and amide S in the hatched columns. The intensity ratios of the complex spectra relative to the spectra of the added components are depicted. All intensities were measured relative to a buffer band at 602 cm–1. Original spectra are shown in Figure 3.

HU₂:H1–13 complex
Interestingly, in the HU2:H1–13 complex spectra the amide I frequency shift is only -3 cm^–1 (Fig. 3). Similarly, a -5 cm^–1 shift is observed for the amide III mode (1255 1250 cm^–1) in the H1–13 complex, which contrasts with the 8–10 cm^–1 shift observed for the other complexes (Table 2). In the HU₂:H1–13 complex the intensity of the amide II and amide S modes are increased relative to the added spectra (Fig. 4).

Phe mode intensity probes local environment
HU protein
In the free protein, vibrational modes corresponding to Phe residues are observed at 1601, 1581, 1179, and 1003 cm^–1 (Table 3). Although observable with a 210 nm excitation wavelength, these modes are strongly enhanced using 215 nm excitation (data not shown; Sokolov and Mukerji 2000). Previously, it has been demonstrated that the intensities of these modes scale with the relative hydrophobicity of the local environment. The 1601 cm^–1 mode intensity, which is a symmetrical C—C stretching mode of the phenyl ring (F8a), has the steepest dependence on local environment (Hildebrandt et al. 1988; Sokolov and Mukerji 2000). In this study the 1601 and 1179 cm^–1 vibrational modes are used as an indicator of the local environment of Phe residues in HU. The relatively intense F8b mode at 1580 cm^–1 is not used for this comparison because of the considerable intensity contribution from a DNA ring-stretching mode.

View larger version (38K): [in this window] [in a new window]   Figure 5. 215 nm-excited spectra of the HU–DNA complexes. (A) Spectra from 1775 to 1425 cm–1. Mode assignments are discussed in the text and are given in Table 3. Top: HU:A4–20; Middle: HU2:T4–20; Bottom: HU2:H1–13. Complexes prepared as in Figure 3. Buffer conditions were the same as in Figure 2 except that 50 mM NaCl was used for the HU2:H1–13 complex. (B) Spectral range: 1450–1100 cm–1. Top: HU:A4–20; Middle: HU2:T4–20; Bottom: HU2:H1–13.

View larger version (34K): [in this window] [in a new window]   Figure 6. Relative Raman intensities of Phe modes at 1179 and 1601 cm–1. Relative intensities of 1179 cm–1 mode are shown in the hatched columns and 1601 cm–1 in the solid gray columns. The intensity ratios of the complex spectra relative to the spectra of the added components are depicted. All intensities shown represent an average of at least two different data sets and were measured relative to a buffer band at 602 cm–1. Original spectra are shown in Figure 5.

Pro mode intensity a measure of intercalation into the DNA
HU protein
Vibrational modes corresponding to Pro residues are observed at 1455 and 1480 cm^–1. The frequency of the imide II Pro mode is sensitive to H-bond donation at the X-Pro carbonyl. Because this vibrational mode primarily arises from C—N bond stretching similar to amide II, stronger H-bonding leads to higher imide II frequencies because of increased electron density on the C—N bond (Takeuchi and Harada 1990; Jordan et al. 1996). The E. coli HU protein has only two Pro residues per subunit located at positions 63 and 77. The higher frequency shoulder, which corresponds to Pro residues with stronger H-bonding, most probably arises from the Pro residues located at 77 and 77 in the "saddle" region of the protein (Fig. 1). Because the -ribbon arms are mobile and relatively unordered the backbone carbonyl moieties of the 63 and 63 Pro residues are probably H-bonded to solvent in the absence of DNA and are assigned to the peak occurring at 1455 cm^–1.

HU:A4–20 and HU₂:T4–20 complexes
HU binding to the A4–20 and T4–20 duplexes leads to an ~30% reduction in Pro mode intensity at 1455 cm^–1 relative to the added spectra (Fig. 7). This decrease in Pro intensity is attributed to intercalation of the 63 and 63 Pro residues in between DNA bases. The reduction in Pro mode intensity is approximately the same in the two complexes, suggesting that the degree of intercalation for the complexes is the same. The intensity of the Pro mode at 1480 cm^–1 is not monitored in the complex spectra because of interference from the relatively intense DNA band occurring at ~1485 cm^–1 (Table 3).

View larger version (13K): [in this window] [in a new window]   Figure 7. Relative Raman intensity of Pro mode at 1455 cm–1. The intensity ratios of the complex spectra relative to the spectra of the added components are depicted. All intensities shown represent an average of at least two different data sets and were measured relative to a buffer band at 602 cm–1. Original spectra are shown in Figure 5.

DNA structural changes upon complex formation
Structural changes of the DNA molecules are detected in the 215 nm spectra shown in Figure 5. Many of these features are also observed with an excitation wavelength of 210 nm; however, with 215 nm, the DNA features are easier to detect because of the reduced intensity contributions from the protein amide modes.

HU:A4–20 and HU₂:T4–20: Changes in DNA backbone conformation
For the complexes formed with the A4–20 and T4–20 duplexes, a large peak at 1578 cm^–1 is observed, which has contributions from Phe and adenine residues (Table 3). The dA bands occurring at 1578 and 1485 cm^–1 arise primarily from C5 = C4 and C4—N3 ring stretching motions (Fodor et al. 1985) and become more defined relative to the added spectrum (Figs. 2, 5A). This relative increase in intensity upon complexation is suggestive of bases unstacking. Backbone ribosyl modes associated with dA residues are broader in the complex because of increased intensity at 1407 and 1308 cm^–1, frequencies that are characteristic of the A-form (Fig. 5B; Toyama et al. 1993; Peticolas 1995). A peak is also observed at 1337 cm^–1 (1343 cm^–1 in the T4–20 complex), which is indicative of the C₂'-sendo conformation and B-form DNA (Thomas and Tsuboi 1993; Peticolas 1995). The broadness of the peaks indicates that the duplex molecules adopt a mixture of A and B conformations in the complexes.

HU:A4–20 and HU₂:T4–20 complexes: The minor groove
Other features observable in the complex spectra (Fig. 5A) that can be assigned to the DNA occur in the high frequency region of the spectrum (1650–1750 cm^–1). These features observed at 1646, 1676, and 1710 cm^–1 arise from the exocyclic carbonyl modes of the dC, dG, and dT residues (Fodor and Spiro 1986; Grygon and Spiro 1990). The 1646 cm^–1 mode arises from the dT C4 = O stretching vibration and dC C2 = O stretching vibration (Fodor et al. 1985; Fodor and Spiro 1986; Mukerji and Williams 2002). The 1676 cm^–1 mode is assigned to the dG C6 = O stretching vibration (Fodor and Spiro 1986; Mukerji et al. 1996). The 1711 cm^–1 mode, although weak, has been shown to arise from the dT C2 = O stretching vibration, and this mode is preferentially enhanced with 215 nm excitation (Grygon and Spiro 1990; Toyama et al. 1991; Tsuboi et al. 1997; Zhang et al. 1998). The observation of these shoulders is consistent with an increase in intensity of these modes and a shift to higher frequency upon complexation. These intensity increases and frequency upshifts arise from a decrease in base stacking interactions and a reduction in H-bonding strength, which potentially results from a significant distortion of the DNA helix and a disruption of base pair H-bonding.

Of most interest is the observation of shoulders corresponding to the dC C2 = O and dT C2 = O stretching vibrations at 1647 and 1711 cm^–1 (Fig. 5A; Fodor and Spiro 1986; Grygon and Spiro 1990; Tsuboi et al. 1997). These groups are located in the minor groove and the intensity increase of the associated vibrational modes suggests that the minor groove opens or widens upon complexation, leading to increased H-bonding with the solvent.

HU₂:H1–13 complex: No evidence for DNA distortion
Interestingly, many of the structural features indicating a distortion of the DNA are not detected in the HU₂:H1–13 complex spectrum, which is dominated by Phe modes (Fig. 5). Some smaller features associated with the DNA can be detected; in particular, the spectral features at 1308, 1337, and 1415 cm^–1 suggestive of a mixture of A- and B- form DNA are observable, as well as the strong features at 1484 and 1579 cm^–1 indicative of base unstacking (Table 3; Thomas and Tsuboi 1993; Peticolas 1995; Thomas 1999). In the 1600–1750 cm^–1 region shoulders are observed at 1641 and 1682 cm^–1, corresponding to the dT C4 = O and dG C6 = O stretching vibrations (Fig. 5A; Fodor and Spiro 1986; Mukerji et al. 1996; Mukerji and Williams 2002). Strikingly, the dT C2 = O stretching vibration is not observed.

	Discussion

TOP Abstract Introduction Results Discussion Materials and methods References

 
Comparison of the complexes: Bent versus linear
Although UVRR spectroscopic results reveal that the three HU–DNA complexes examined share many common features, there are pronounced differences in the structures that can be related to binding affinity and site specificity. In general, the HU:A4–20 and HU₂:T4–20 complexes are very similar structurally, despite the difference in HU binding stoichiometry; yet, these complexes share only a few structural features with the HU₂:H1–13 complex.

Determination of secondary structure content of free HU by UVRR is consistent with the known structure of HU as determined from X-ray crystallography (Tanaka et al. 1984; White et al. 1989, 1999), NMR (Vis et al. 1995), and Raman (Serban et al. 2003a) methods, which indicate that the -helical core of the protein is well folded, but the -ribbon arms are relatively unordered and mobile in the absence of DNA. In the HU:A4–20 and HU₂:T4–20 complexes, however, HU exhibits a greater amount of helical content upon binding DNA and the DNA spectral bands indicative of distortion and minor groove opening are comparable in intensity and frequency, suggesting that HU interacts with these two duplexes in a similar manner to yield a bent complex. These results are in good agreement with FRET measurements that had previously determined HU binding to either the A4–20 or T4–20 duplex led to a very similar bend angle, 70° and 62°, respectively (Wojtuszewski and Mukerji 2003).

In the current study the observed opening of the minor groove is consistent with an interaction between the protein -arms and the minor groove as observed in the crystal structure of Anabena HU bound to DNA (Swinger et al. 2003; Fig. 1). This structural feature appears to be an important element of binding and recognition, as has been previously suggested by X-ray crystallography (Swinger et al. 2003) and Raman (Serban et al. 2003b) studies.

By contrast, the UVRR spectral features of the HU₂:H1–13 complex are attributed to a different and low affinity protein binding interaction. In this complex, smaller frequency shifts are observed for both amide I and III and spectral features indicative of an increase in -helical content and Pro intercalation are reduced relative to the other complexes. These observations suggest that the -helical secondary structure is less rigid and forms weaker H-bonds in the HU₂:H1–13 complex. The lack of observable C=O bands at high frequency and less intense DNA ring-stretching modes argues against a significant amount of minor groove opening and DNA distortion in this complex, consistent with previous fluorescence measurements (Wojtuszewski et al. 2001).

All three complexes exhibit UVRR spectral signatures of A- and B-form DNA, suggesting that protein binding modifies DNA conformation. This finding is consistent with a previous Raman study of Bacillus stearothermophilus HU binding to two different DNA dodecamers, in which a mixture of A- and B-form DNA was observed as well as the unstacking of the purine bases (Serban et al. 2003b). Similarly, a twist of 31°/bp was also observed in the Anabena HU-DNA cocrystal structure, indicating that the DNA is underwound with respect to canonical B-form DNA and is closer to that of the A-form conformation (Swinger et al. 2003).

Pro intercalation and induced bending
Pro vibrational modes in the A4–20 and T4–20 complexes decrease in intensity upon complex formation. E. coli HU only has two Pro residues per subunit located at positions 63 and 77, in contrast to B. stearothermophilus HU and Anabena HU, which have six and eight Pro residues, respectively. The Pro residues in E. coli HU are highly conserved, and are also found in Anabena HU and IHF (Rice 1997). In the X-ray crystal structures Pro 63 was observed to intercalate and stack with DNA base pairs to induce bending (Rice et al. 1996; Swinger et al. 2003). Therefore, the reduction in Pro intensity is attributed to the intercalation of Pro residues (63 and 63) into the DNA, potentially between TT bases, as detected in the IHF cocrystal structure (Fig. 8).

View larger version (33K): [in this window] [in a new window]   Figure 8. Pro intercalation and stacking interactions in HU–DNA and IHF–DNA cocrystal structures. (Left) Stacking interaction between Pro 63 and T5 in the HU–DNA cocrystal structure (Swinger et al. 2003). (Right) Pro 65 intercalation and stacking interaction between two thymine residues (T37 and T38) in the IHF–DNA cocrystal structure (Rice et al. 1996). Diagram was generated using Weblab Viewer Pro 3.7 and X-ray crystallographic coordinates of Rice et al. (1996; Swinger et al. 2003), pdbid: 1P51 [PDB] and 1IHF [PDB] .

In contrast to the bent complexes, the Pro vibrational mode intensity in the "linear" HU₂:H1–13 complex is increased relative to that obtained for the added components, indicating that there is no intercalation of Pro residues. The lack of intercalation is consistent with the absence of induced bending detected for this complex by fluorescence spectroscopy (Wojtuszewski et al. 2001). These findings suggest that the relatively low affinity and non-site-specific HU₂:H1–13 complex is structurally distinct from either the HU:A4–20 or HU2:T4–20 complexes. Furthermore, these observations suggest that the relative UVRR intensities of Pro vibrational modes can be used as an indicator of Pro intercalation and induced bending.

Comparison of the complexes: Flexibility of the monomer–monomer interface
By analogy with the structure from B. stearothermophilus HU (White et al. 1999), six Phe residues (47, 50, and 79 from each dimer; see Fig. 1) form an aromatic cluster that links the top of the -helical region to the hydrophobic inner surface of the -sheet region. The intensities of Phe vibrational modes scale linearly with hydrophobicity (Hildebrandt et al. 1988; Sokolov and Mukerji 2000); therefore, the decrease in Phe mode intensity in the A4–20 and T4–20 complexes suggests that the hydrophobic interactions in this cluster become less pronounced as a consequence of the binding interaction. This would be consistent with an opening of the dimeric interface as the protein adapts its conformation to bind and bend the DNA. Recent Raman experiments of B. stearothermophilus HU that monitored deuteration of the amide II mode in the absence of DNA have indicated that the dimer interface is more flexible than previously thought (Serban et al. 2003a).

In contrast, the strongest and most distinctive features of the 215 nm-excited HU₂:H1–13 complex spectra arise from the Phe residues, and the observed intensity increase can be attributed to an increase in hydrophobicity of their local environment and potentially an interaction of HU dimers. Analytical ultracentrifugation and fluorescence anisotropy results had previously demonstrated that two HU dimers bind to this duplex (Wojtuszewski et al. 2001). The increase in hydrophobicity, as suggested by the increase in Phe mode intensity, may result from an interaction between the two HU dimers upon binding to the DNA. In this complex the relative proximity and orientation of the HU dimers with respect to the duplex is not known; however, the HU–DNA crystal structure suggests that there should be at least nine bp in between the Pro residues located on the protein -arms arguing against a linear binding mode (Fig. 1; Swinger et al. 2003).

UVRR measurements of Phe local environment indicate that the dimer interface adapts to the type of complex formed, where complexes with induced bending have a more open interface than linear complexes, such as the HU₂:H1–13 complex, which have a stronger monomer–monomer interaction and tighter interface. The HU₂:H1–13 complex is probably more reminiscent of the high protein:DNA ratio complexes observed by Dame and Goosen (Dame and Goosen 2002), which lead to rigidification and elongation of the DNA instead of bending.

Binding affinity and recognition through the C-terminal helix
Collectively, the amide modes, which arise from the peptide backbone, indicate that an increase in secondary structure occurs upon HU complex formation with DNA. A previous study that evaluated the ability of UVRR to determine secondary structure, demonstrated that the frequency and bandwidth of amide III increased with -helical content (Chi et al. 1998). In the case of the A4–20 and T4–20 complexes, the frequency shifts and intensity increases of the -helical amide modes are attributed to an increase in H-bonding strength in the body of the protein and an increase in -helical structure at the C terminus (Wang et al. 1991; Chi et al. 1998). The increase in -sheet content probably arises from the ordering of the -arms upon interaction with the DNA as has been observed previously by X-ray crystallography and Raman spectroscopy (Rice et al. 1996; Rice 1997; Serban et al. 2003b; Swinger et al. 2003). Although similar spectral changes are observed for the H1–13 complex, the amide I and amide III peptide backbone modes exhibit smaller frequency shifts, indicating that the relative increase in -helical secondary structure for this complex is smaller than for the other two complexes. An intensity increase is detected for the amide II and amide S modes, which is attributed to an increase in -sheet content upon complex formation, because the Raman cross-section of -sheets is greater than that of -helices and the maximum absorbance of -sheets is shifted to a wavelength that is closer to the excitation wavelength of 210 nm (Wang et al. 1991; Austin et al. 1993). The intensity increase observed for the HU₂:H1–13 complex as opposed to the intensity decrease observed for the HU:A4–20 and HU₂:T4–20 complexes further confirms that these complexes and not the HU₂:H1–13 complex experiences an increase in -helical content upon complexation.

The start of the C-terminal helix differs in the NMR solution (Vis et al. 1995) and X-ray crystallographic structures (White et al. 1999) of the protein and this region may be a source of protein flexibility, as are the -arms. We attribute the increase in -helicity of the A4–20 and T4–20 complexes to this region of the protein and speculate that contact between the helix and the DNA facilitates specific binding interactions. The length of the T4–20 and A4–20 duplexes coupled with the induced bending allows them to interact with this region of the protein. In the case of the H1–13 duplex, the shorter 13 bp duplex and the absence of bending prohibits this interaction.

Recognition through this region of the protein stems from a consideration of other members of the HU-like family of proteins, TF1 and IHF. Both of these proteins have longer C-terminal helices than HU and they bind with nanomolar affinity to specific recognition sites (White et al. 1989; Rice 1997). Modification of this region in IHF, either by substitution (Mengeritsky et al. 1993) or deletion (Zulianello et al. 1995) leads to altered binding specificity and affinity. In addition, the IHF and HU cocrystal structures reveal that the DNA interacts with this region of the protein (Rice et al. 1996; Swinger et al. 2003). In the case of the HU–DNA crystal structure, the interaction occurs with another 17 bp DNA duplex in the crystal unit (Fig. 1); however, it is expected that this favorable interaction could occur within one molecule with a longer DNA duplex (Swinger et al. 2003). Thus, the higher binding affinity and reduced protein to DNA stoichiometry observed for the A4–20 duplex is attributed to the formation and interaction of this helix with the DNA and supports the notion that this helix plays an important role in binding and recognition.

In summary, UVRR results indicate that the HU₂:H1–13 complex is distinct from the complexes formed with either the A4–20 duplex or the T4–20 duplex and leads to identification of some structural elements that potentially lead to higher affinity and 1:1 binding stoichiometries. From the spectral comparison of these complexes and the observed differences in binding affinity (5 µM^–1 versus 0.5 µM^–1: Table 1), we propose that induced bending as a consequence of Pro intercalation is a significant element in the HU binding interaction and plays a role in specificity. Bending and untwisting of the DNA leads to interactions with the protein C-terminal helix, which becomes more pronounced in the complex, and these contacts also influence binding and recognition. The non-site-specific interaction, as characterized by the HU₂:H1–13 complex, exhibits greater flexibility in protein structure, stronger monomer–monomer interactions, as evidenced by a tighter hydrophobic core, and little bending or distortion of DNA. Further experiments on discontinuous DNA substrates that form high affinity complexes with HU will help to elucidate the structural elements of specificity in the HU–DNA binding interaction and are currently underway in our laboratory.

	Materials and methods

TOP Abstract Introduction Results Discussion Materials and methods References

 
HU protein
HU protein was isolated from E. coli strain, RLM1078, and purified according to the procedure previously described (Wojtuszewski et al. 2001). The amount of protein used in formation of the HU–DNA complexes for the UVRR experiments was determined based on previously determined binding constants (Wojtuszewski et al. 2001; Wojtuszewski and Mukerji 2003). The concentrations of HU and DNA were chosen such that less than 1% of the total amount of DNA was free. All experiments were conducted in a 10 mM Tris buffer at pH 7.6 containing 0.1 mM EDTA. A4–20 and T4–20 duplexes and complexes were examined in the presence of 100 mM NaCl, while the H1–13 duplex and complex was examined in the presence of 50 mM NaCl. In instances where Na₂SO₄ was used instead of NaCl, 50 mM Na₂SO₄ was used with the A4–20 and T4–20 duplexes and complexes and 25 mM Na₂SO₄ was used with the H1–13 duplex and complex.

Oligonucleotides
Oligonucleotides were purified by UV-shadowing and electro-eluted from the gel (Schleicher and Schuell). Duplex formation was achieved by mixing equal molar amounts of the pure single stranded oligonucleotides, heating at 90°C for 5 min followed by slow cooling to room temperature in a water bath at a rate of approximately 0.2°/min. Concentrations of single strand oligo-nucleotides were determined by measuring the absorbance at 260 nm at 90°C. Extinction coefficients were calculated using the methodology of Richards (1975).

UVRR spectroscopy
The UVRR spectrometer has been previously described (Sokolov and Mukerji 2000; Mukerji and Williams 2002). Absolute frequencies are accurate to ±1 cm^–1 and relative shifts are accurate to ±0.25 cm^–1. HU protein, DNA, and complex samples were continuously stirred in a 3 x 3-mm Spectrosil quartz cuvette (Starna Cells, Inc.) with a sample volume of 80 µL and cooled with a circulating water bath to approximately 6 ± 2°C.

Each sample was acquired for four 15-min cycles in 1 h. A total of 3 h of data for each sample were collected. When applicable, the data were normalized to a factor of unity using the symmetric stretching mode of SO₄^2– occurring at 981 cm^–1. For intensity comparisons, data were normalized to a buffer band at 602 cm^–1. Intensity comparisons made with either the SO₄^2– or the 602 cm^–1 buffer band yielded the same result. For comparisons between complex and added spectra, spectral addition was accomplished using the buffer as intensity standard and correcting for any concentration differences in the complex spectra relative to the component spectra. Data manipulation and analysis were done using Grams/32 and Grams/AI (Galactic Industries).

	Footnotes

 
¹ Present address: Optical Spectroscopy Section, NHLBI/NIH, Laboratory of Cell Biology, Building 10, Room 5D-14, Bethesda, MD 20892-1412, USA.

	Acknowledgments

 
This work was supported by a National Science Foundation Career Development Award (MCB-9507241) and a grant from the Patrick and Catherine Weldon Donaghue Medical Research Foundation. K.W. gratefully acknowledges support from an NIH training grant in Molecular Biophysics (GM08271). We are grateful to Prof. Roger McMacken for the gift of the HU overexpressing E. coli strain RLM1078, and we also thank Phoebe Rice for sharing her results on the Anabena HU–DNA complex prior to publication.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.

	References

TOP Abstract Introduction Results Discussion Materials and methods References

 
Aki, T. and Adhya, S. 1997. Repressor induced site-specific binding of HU for transcriptional regulation. EMBO J. 16: 3666–3674.[CrossRef][Medline]

Austin, J.C., Jordan, T., and Spiro, T.G. 1993. Ultraviolet resonance Raman studies of proteins and related model compounds. In Biomolecular spectroscopy (eds. R.J.H. Clark and R.E. Hester), pp. 55–127. John Wiley and Sons, Ltd., New York.

Burkhoff, A.M. and Tullius, T.D. 1988. Structural details of an adenine tract that does not cause DNA to bend. Nature 331: 455–457.[CrossRef][Medline]

Castaing, B., Zelwer, C., Laval, J., and Boiteux, S. 1995. HU protein of Escherichia coli binds specifically to DNA that contains single-strand breaks or gaps. J. Biol. Chem. 270: 10291–10296.[Abstract/Free Full Text]

Chi, Z., Chen, X.G., Holtz, J.S.W., and Asher, S.A. 1998. UV resonance Raman-selective amide vibrational enhancement: Quantitative methodology for determining protein secondary structure. Biochemistry 37: 2854–2864.[CrossRef][Medline]

Dame, R.T. and Goosen, N. 2002. HU: Promoting or counteracting DNA compaction? FEBS Lett. 529: 151–156.[CrossRef][Medline]

Drlica, K. and Rouviere-Yaniv, J. 1987. Histonelike proteins of bacteria. Microbiol. Rev. 51: 301–319.[Free Full Text]

Fodor, S.P.A. and Spiro, T.G. 1986. Ultraviolet resonance Raman spectroscopy of DNA with 200–266-nm laser excitation. J. Am. Chem. Soc. 108: 3198–3205.[CrossRef]

Fodor, S.P.A., Rava, R.P., Hays, T.R., and Spiro, T.G. 1985. Ultraviolet resonance Raman spectroscopy of the nucleotides with 266-, 240-, 218-, and 200-nm pulsed laser excitation. J. Am. Chem. Soc. 107: 1520–1529.[CrossRef]

Grygon, C.A. and Spiro, T.G. 1990. UV resonance Raman spectroscopy of nucleic acid duplexes containing A–U and A–T base pairs. Biopolymers 29: 707–715.[CrossRef]

Hagerman, P.J. 1986. Sequence-directed curvature of DNA. Nature 321: 449–450.[CrossRef][Medline]

Harada, I. and Takeuchi, H. 1986. Raman and ultraviolet resonance Raman spectra of proteins and related compounds. In Spectroscopy of biological systems (eds. R.J.H. Clark and R.E. Hester), pp. 113–175. Wiley, New York.

Haran, T.E. and Crothers, D.M. 1989. Cooperativity in A-tract structure and bending properties of composite T_nA_n blocks. Biochemistry 28: 2763–2767.[CrossRef][Medline]

Hildebrandt, P.G., Copeland, R.A., Spiro, T.G., Otlewski, J., Laskowski Jr., M., and Prendergast, F.G. 1988. Tyrosine hydrogen-bonding and environmental effects in proteins probed by ultraviolet resonance Raman spectroscopy. Biochemistry 27: 5426–5433.[CrossRef][Medline]

Jordan, T., Mukerji, I., Wang, Y., and Spiro, T.G. 1996. UV resonance Raman spectroscopy of the proline peptide bond. J. Mol. Struct 379: 51–64.[CrossRef]

Kamashev, D. and Rouviere-Yaniv, J. 2000. The histone-like protein HU binds specifically to DNA recombination and repair intermediates. EMBO J. 19: 6527–6535.[CrossRef][Medline]

Kamashev, D., Balandina, A., and Rouviere-Yaniv, J. 1999. The binding motif recognized by HU on both nicked and cruciform DNA. EMBO J. 18: 5434–5444.[CrossRef][Medline]

Kobryn, K., Lavoie, B.D., and Chaconas, G. 1999. Supercoiling-dependent site-specific binding of HU to naked Mu DNA. J. Mol. Biol. 289: 777–784.[CrossRef][Medline]

Lavoie, B.D. and Chaconas, G. 1993. Site-specific HU binding in the Mu transpososome: Conversion of a sequence-independent DNA-binding protein into a chemical nuclease. Genes & Dev. 7: 2510–2519.[Abstract/Free Full Text]

Lavoie, B.D., Shaw, G.S., Millner, A., and Chaconas, G. 1996. Anatomy of a flexer–DNA complex inside a higher-order transposition intermediate. Cell 85: 761–771.[CrossRef][Medline]

Lia, G., Bensimon, D., Croquette, V., Allemand, J.F., Dunlap, D., Lewis, D.E., Adhya, S., and Finzi, L. 2003. Supercoiling and denaturation in Gal repressor/heat unstable nucleoid protein (HU)-mediated DNA looping. Proc. Natl. Acad. Sci. 100: 11373–11377.[Abstract/Free Full Text]

Mengeritsky, G., Goldenberg, D., Mendelson, I., Giladi, H., and Oppenheim, A.B. 1993. Genetic and biochemical analysis of the Integration Host Factor of Escherichia coli. J. Mol. Biol. 231: 646–657.[CrossRef][Medline]

Mukerji, I. and Williams, A.P. 2002. UV resonance Raman and circular dichroism studies of a DNA duplex containing an A₃T₃ tract: Evidence for a premelting transition and three-centered H-bonds. Biochemistry 41: 69–77.[CrossRef][Medline]

Mukerji, I., Shiber, M.C., Spiro, T.G., and Fresco, J.R. 1995. A UV resonance Raman study of d(A⁽⁺⁾-G)₁₀, a single-stranded helix without stacked or paired bases. Biochemistry 34: 14300–14303.[CrossRef][Medline]

Mukerji, I., Shiber, M.C., Fresco, J.R., and Spiro, T.G. 1996. A UV resonance Raman study of hairpin dimer helices of d(A–G)₁₀ at neutral pH containing intercalated dA residues and alternating dG tetrads. Nucleic Acids Res. 24: 5013–5020.[Abstract/Free Full Text]

Mukerji, I., Sokolov, L., and Mihailescu, M.-R. 1998. A UV resonance Raman investigation of poly (rI): Evidence for cation dependent structural perturbations. Biopolymers 46: 475–487.[CrossRef]

Murphy, F.V. and Churchill, M.E.A. 2000. Nonsequence-specific DNA recognition: A structural perspective. Structure 8: R83–R89.[Medline]

Peticolas, W.L. 1995. Raman spectroscopy of DNA and proteins. Methods Enzymol. 246: 389–416.[Medline]

Pinson, V., Takahashi, M., and Rouviere-Yaniv, J. 1999. Differential binding of the Escherichia coli HU, homodimeric forms and heterodimeric form to linear, gapped and cruciform DNA. J. Mol. Biol. 287: 485–497.[CrossRef][Medline]

Rice, P.A. 1997. Making DNA do a U-turn: IHF and related proteins. Curr. Opin. Struct. Biol. 7: 86–93.[CrossRef][Medline]

Rice, P.A., Yang, S.-w., Mizuuchi, K., and Nash, H.A. 1996. Crystal structure of an IHF-DNA complex: A protein-induced DNA U-turn. Cell 87: 1295–1306.[CrossRef][Medline]

Richards, E.G. 1975. Nucleic acids. In Handbook of biochemistry and molecular biology, 3rd ed. (ed. G.D. Fasman). CRC, Cleveland, OH.

Rouviere-Yaniv, J. and Gros, F. 1975. Characterization of a novel, low-molecular-weight DNA-binding protein from Escherichia coli. Proc. Nat. Acad. Sci. 72: 3428–3432.[Abstract/Free Full Text]

Serban, D., Arcineigas, S.F., Vorgias, C.E., and Thomas Jr., G.J. 2003a. Structure and dynamics of the DNA-binding protein HU of B. stearothermophilus investigated by Raman and ultraviolet-resonance Raman spectroscopy. Protein Sci. 12: 861–870.[Abstract/Free Full Text]

Serban, D., Benevides, J.M., and Thomas Jr., G.J. 2003b. HU protein employs similar mechanisms of minor-groove recognition in binding to different B-DNA sites: Demonstration by Raman spectroscopy. Biochemistry 42: 7390–7399.[CrossRef][Medline]

Sokolov, L. and Mukerji, I. 2000. Structure of sickle cell hemoglobin fibers probed with UV resonance Raman spectroscopy. J. Phys. Chem. 104: 10835–10844.

Sokolov, L., Wojtuszewski, K., Tsukroff, E., and Mukerji, I. 2000. Nucleic acid structure investigated by UV resonance Raman spectroscopy: Protonation effects and A-tract structure. J. Biomol. Struct. Dyn. 327–334.

Swinger, K.K., Lemberg, K.M., Zhang, Y., and Rice, P.A. 2003. Flexible DNA bending in HU–DNA cocrystal structures. EMBO J. 22: 3749–3760.[CrossRef][Medline]

Takeuchi, H. and Harada, I. 1990. Ultraviolet resonance Raman spectroscopy of X-proline bonds: A new marker band of hydrogen bonding at the imide C = O site. J. Raman Spectrosc. 21: 509–515.

Tanaka, I., Appelt, K., Dijk, J., White, S.W., and Wilson, K.S. 1984. 3-Å resolution structure of a protein with histone-like properties in prokaryotes. Nature 310: 376–381.[CrossRef][Medline]

Thomas Jr., G.J. 1999. Raman spectroscopy of protein and nucleic acid assemblies. Annu. Rev. Biophys. Biomol. Struct. 28: 1–27.[CrossRef][Medline]

Thomas Jr., G.J. and Tsuboi, M. 1993. Raman spectroscopy of nucleic acids and their complexes. Adv. Biophys. Chem. 3: 1–70.

Toyama, A., Takeuchi, H., and Harada, I. 1991. Ultraviolet resonance Raman spectra of adenine, uracil and thymine derivatives in several solvents. Correlation between band frequencies and hydrogen-bonding states of the nucleic acid bases. J. Mol. Struct. 242: 87–98.[CrossRef]

Toyama, A., Takino, Y., Takeuchi, H., and Harada, I. 1993. Ultraviolet resonance Raman spectra of ribosyl C(1')-deuterated purine nucleosides: Evidence of vibrational coupling between purine and ribose rings. J. Am. Chem. Soc. 115: 11092–11098.[CrossRef]

Travers, A. 1993. DNA–protein interactions: The three-dimensional architecture of DNA–protein complexes. In DNA–protein interactions, pp. 28–51. Chapman & Hall, London.

Tsuboi, M., Kumakura, A., Aida, M., Kaneko, M., Dupuis, M., Ushizawa, K., and Ueda, T. 1997. Raman scattering tensors in thymine molecule from an ab initio MO calculation. Spectrochim. Acta Part A 53: 409–419.[CrossRef]

Vis, H., Mariani, M., Vorgias, C.E., Wilson, K.S., Kaptein, R., and Boelens, R. 1995. Solution structure of the HU protein from Bacillus stearothermophilus. J. Mol. Biol. 254: 692–703.[CrossRef][Medline]

Wang, Y., Purrello, R., Jordan, T., and Spiro, T.G. 1991. UVRR spectroscopy of the peptide bond. 1. Amide S, a nonhelical structure marker, is a CH bending mode. J. Am. Chem. Soc. 113: 6359–6368.[CrossRef]

White, S.W., Appelt, K., Wilson, K.S., and Tanaka, I. 1989. A protein structural motif that bends DNA. Proteins 5: 281–288.[CrossRef][Medline]

White, S.W., Wilson, K.S., Appelt, K., and Tanaka, I. 1999. The high-resolution structure of DNA-binding protein HU from B. stearothermophilus. Acta Crystallogr. D 55: 801–809.[CrossRef][Medline]

Wojtuszewski, K. and Mukerji, I. 2003. HU binding to bent DNA: A fluorescence anisotropy and resonance energy transfer study. Bioch