- Email: [email protected]
A Divalent Metal-mediated Switch Controlling Protein-induced DNA Bending
doi:10.1016/j.jmb.2006.09.082
J. Mol. Biol. (2007) 367, 731–740
A Divalent Metal-mediated Switch Controlling Protein-induced DNA Bending† Qiuye Bao 1 , Hu Chen 2 , Yingjie Liu 2 , Jie Yan 2 , Peter Dröge 1 ‡⁎ and Curt A. Davey 3 ‡⁎ 1
Division of Genomics and Genetics, School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, Singapore 2
Department of Physics, National University of Singapore, 2 Science Drive, Singapore 117542, Singapore 3
Division of Structural and Computational Biology, School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, Singapore *Corresponding authors
Architectural proteins that reconfigure the paths of DNA segments are required for the establishment of functional interfaces in many genomic transactions. A single-chain derivative of the DNA architectural protein integration host factor was found to adopt two stable conformational states in complex with a specific DNA target. In the so-called open state, the degree of protein-induced DNA bending is reduced significantly compared with the closed state. The conformational switch between these states is controlled by divalent metal binding in two electronegative zones arising from the lysine-to-glutamate substitution in the protein body proximal to the phosphate backbone of one DNA arm. We show that this switch can be employed to control the efficiency of site-specific recombination catalyzed by λ integrase. Introduction of acidic residues at the protein–DNA interface holds potential for the design of metal-mediated switches for the investigation of functional relationships. © 2006 Elsevier Ltd. All rights reserved.
Keywords: divalent metal; DNA architectural proteins; DNA topology; integration host factor; nucleoprotein complex formation
Introduction The assembly of specialized nucleoprotein structures (snups) inside living cells is a key regulatory step in many DNA transactions, such as site-specific recombination, replication, and transcription.1 In many cases, snup assembly is orchestrated by DNA architectural proteins that alter the trajectory of DNA segments in a more or less defined way through DNA bending. However, very little is known about external factors, such as the topological state of DNA, that could govern the dynamics of functional snup formation in a genome. The integration host factor (IHF) is a key DNA architectural protein in Escherichia coli.2,3 IHF is † P.D. dedicates this work to the memory of Nicholas R. Cozzarelli. ‡ P.D. and C.A.D. contributed equally to this work. Abbreviations used: snup, specialized nucleoprotein structure; IHF, integration host factor; AFM, atomic force microscopy; EMSA, electrophoretic mobility-shift assay. E-mail addresses of the corresponding authors: [email protected]; [email protected]
composed of two homologous subunits, the α and the β-subunit, and displays limited DNA sequence specificity.4 The crystal structure of IHF in complex with the phage λ H′ site, as well as fluorescent resonance energy transfer (FRET) analyses and visualization of IHF–DNA complexes through atomic force microscopy (AFM), revealed strong (120–160°) protein-induced DNA bending.5–7 Bending is achieved mainly by the intercalation of one conserved proline residue from each subunit of the heterodimer into the minor groove. The resulting Uturn-like DNA conformation is stabilized by electrostatic interactions around the protein body.8,9 IHF is involved in the regulation of more than 100 genes in Gram-negative bacteria and it is an essential cofactor in phage λ site-specific recombination, where the protein serves an architectural role during the assembly of snups.10–12 The 240 bp comprising phage attachment (att) site attP is one of the two recombination sequences in the integrative pathway and harbors three specific IHF-binding sites that must be occupied by IHF to generate a functional snup, the so-called integrative intasome. Intasome assembly strictly requires negative DNA supercoiling of attP.13 The intasome then captures
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
732 the protein-free 21 bp attB to form a synaptic complex in which two successive rounds of DNA strand exchange are catalyzed by phage λ integrase (Int).14 Excisive recombination is genetically the exact reversal of integration and employs hybrid attachment sites attL and attR as substrates. Excision also requires IHF, but involves only two binding sites in complex with and bent by IHF. In contrast to the integrative pathway, excision does not depend on negative supercoiling of attL or attR. However, a second DNA-bending co-factor, the phage-encoded excisionase (Xis), is required and binds specifically to two sites on attR. The general function of IHF and Xis in λ recombination is to construct a specific DNA trajectory within each of the three intasomes, which ultimately enables the heterobivalent Int to establish contacts between separate DNA sites simultaneously. Recent intasome models, based on co-crystal structures of Int tetramers in complexes with core and arm-binding sites, further highlighted the crucial role of IHF-induced DNA bending in the formation of functional intasomes.15,16 We recently transferred the phage λ recombination system to mammalian cells and engineered a single-chain IHF, named scIHF2, which is functional in mammalian cells.17 This was achieved by insertion of almost the entire α-subunit into the β-subunit at position 39 using two short peptide linkers.18 Biochemical and functional assays confirmed that scIHF2 behaves like its heteromeric parent. During construction of scIHF2, we identified a variant (scIHF3-K45αE) that carries glutamate instead of lysine at position 45 of the α-subunit. In addition, one of the two linkers is shortened in scIHF3-K45αE. We found that scIHF3-K45αE is nearly inactive in promoting integrative recombination in vitro, while remaining fully active as a co-factor for excisive recombination on negatively supercoiled substrates. The protein also exhibits a defect in its function as an initiation factor for pSC101 replication in vivo.19 In order to analyze this interesting phenotype in more detail, we introduced the αK45E substitution into scIHF2 and performed a detailed functional and structural analysis. This led to the identification of a novel, controllable modular mode of proteininduced DNA bending. In addition, our results obtained with the phage λ site-specific recombination system provide valuable insight into possible dynamics of functional snup formation in general, and how this can be governed by an intricate interplay between DNA architectural proteins and external factors.
Results Biochemical and functional properties of scIHF2-K45αE We introduced the K45αE substitution into scIHF2 (Figure 1(a)) and demonstrated that purified
Metal-mediated Protein-induced DNA Bending
Figure 1. Sequence of scIHF2-K45αE and EMSA. (a) Numbering of residues follows that of the heteromeric IHF. The two short linkers are shaded. The αK45E substitution and corresponding residue in the β-subunit are underlined. (b) EMSA analysis of complexes formed between the 34 bp long, radio-labeled H′ and either scIHF2 or scIHF2-K45αE.
scIHF2-K45αE in complex with a 34 bp oligonucleotide comprising the H′ site migrates more slowly than the scIHF2-H′ complex (Figure 1(b)). Hence, the substitution at position 45 in the α-subunit is solely responsible for the altered complex mobility, which was observed earlier with scIHF3-K45αE.19 We subsequently determined the Kd values of scIHF2-K45αE for all three IHF-specific binding sites on attP as <2 nM at an ionic strength <100 mM NaCl, which are within a factor of 2 of the respective values determined for scIHF2 (data not shown). We then tested scIHF2-K45αE in in vitro recombination assays using protein-to-DNA stoichiometries that, with each variant of IHF tested, lead to at least 80% complex formation on individual IHF target sites. We found that scIHF2-K45αE is marginally active as a co-factor in intramolecular integrative recombination on supercoiled pλIR (Figure 2(a), lane 3), but it triggers excisive recombination on supercoiled pλER that contains attL and attR instead of attB and attP, respectively, to an extent comparable with either scIHF2 or wild-type IHF (data not shown). However, we discovered that the accessory role of scIHF2-K45αE in excisive recombination depends entirely on negative supercoiling of the DNA substrate. No recombination product is detected when scIHF2-K45αE is tested with linearized pλER, which contains directly repeated attR and attL as substrate, under conditions where both scIHF2 and wild-type IHF are fully active (data not shown). We therefore decided to use intermolecular recombination in order to dissect whether attR or attL, or both, requires superhelical tension with scIHF2-K45αE. As shown in Figure 2(b), lanes 5 to 7, wild-type IHF, scIHF2, and scIHF2-K45αE serve as cofactors when supercoiled attR is recombined with linearized attL. However, only wild-type IHF and scIHF2 activate recombination when linearized attR is incubated with supercoiled attL (Figure 2(c), lanes 5 to 7). Since we used a fourfold greater amount of scIHF2-K45αE than scIHF2 in our recombination
733
Metal-mediated Protein-induced DNA Bending
Figure 2. In vitro recombination assays. (a) Intramolecular recombination on supercoiled pλIR. (b) Intermolecular recombination between supercoiled attR and linearized attL. Note that two recombination products are expected, since the supercoiled attR substrate DNA is monomeric and dimeric. (c) Intermolecular recombination between supercoiled attL and linearized attR. Recombination products were analyzed through agarose gel electrophoresis.
assays, and since it is known that large amounts of IHF can inhibit excisive recombination, we successively reduced the amount of scIHF2-K45αE about fivefold but were still unable to detect recombination activity on linearized pλER (data not shown). The absolute dependency of scIHF2-K45αE on topologically underwound attR in excisive recombination led us to analyze the complexes formed on attR through electrophoretic mobility-shift assay (EMSA). We used a DNA fragment in which the two cognate sites present in attR, H1 and H2, are positioned at about the same distance from the ends. We observe two distinct complexes that migrate faster than those formed between scIHF2 and attR (Figure 3). Further, these differences are detectable
regardless of whether scIHF2-K45αE is first bound to supercoiled attR followed by restriction digest, or vice versa (left-hand and right-hand panel, respectively). The opposing differences in electrophoretic mobility of the scIHF2-K45αE-H′ and the scIHF2K45αE-attR complexes in comparison with the corresponding scIHF2–DNA complexes is apparently a consequence of the more intricate influence of intrinsic curvature of longer DNA segments, such as attL, on overall mobility when coupled with protein-induced bending. Nonetheless, these results confirm that scIHF2-K45αE is able to bind to H1 and H2 on linearized attR, and reveal that significant conformational differences must exist between scIHF2-K45αE-attR complexes and those formed by scIHF2 and attR. The crystal structure of IHF in complex with H′ showed that the lysine side-chain at position 45 of the α-subunit interacts with phosphate groups of the A-tract, most likely stabilizing DNA bending.5 Since we substituted this lysine with glutamate and observed altered electrophoretic mobilities of scIHF2-K45αE-DNA complexes, indicative of altered complex conformation, we hypothesized that interactions between the corresponding region of the protein body and the A-tract, which we will subsequently refer to as the left DNA arm, were weakened significantly. This could result in a reduced degree of overall DNA bending that, in turn, affects the electrophoretic mobilities of complexes. In order to obtain more direct evidence for this scenario, we used AFM and determined the degree of DNA bending in images from individual protein–DNA complexes.
Figure 3. EMSA of complex formation on attR. The attR fragment used in this analysis is schematized in the top panel. DNA was either pre-cleaved with EcoRI (right) or the supercoiled plasmid was incubated directly with scIHF2 or scIHF2-K45αE (left), followed by a HindIII/ NotI digest before PAGE.
DNA bending determined by atomic force microscopy A 623 bp attL-carrying DNA fragment was used for AFM imaging. The fragment harbors the H′ site positioned asymmetrically 200 bp from one end.
734 DNA was incubated either with purified wild-type IHF, scIHF2, or scIHF2-K45αE, and the resulting complexes were adsorbed to mica. Naked DNA served as the control. AFM images that showed both a protein signal and DNA bending at the expected DNA region were further analyzed. None of the more than 100 images inspected in the control sample with naked DNA show significant DNA bending together with a protein signal at a corresponding position (not shown). Our analysis of AFM images from wild-type IHF–DNA and scIHF2–DNA complexes reveal mean bending angles of 117(±19)° and 114(±15)°, respectively (Figure 4(a) and (b)). These values are in very good agreement with those from a recent AFM study performed with wild-type IHF on a segment from the TOL plasmid.7 Our analysis using scIHF2K45αE, however, yields a distribution of bending angles with a significantly smaller mean value of 91(±19) °degrees (Figure 4(c)). Together, these results support our hypothesis that the degree of overall DNA bending in scIHF2-K45αE-DNA complexes is reduced significantly, most likely as a result of weakened protein interaction with the left DNA arm that result from the K45αE substitution. In order to investigate this possibility further, we solved the crystal structures of scIHF2 and scIHF2-K45αE in complex with H′ DNA. Crystal structures of scIHF2-DNA and scIHF2-K45αE-DNA complexes The construct used to obtain the crystal structure of the wild-type IHF–DNA complex was composed of a 35 bp DNA containing a nick in one strand adjacent to the proline intercalation site of the βsubunit.5 The presence of the single-strand break was necessary for obtaining well-diffracting crystals, consistent with its involvement in an extensive interparticle contact in the crystal. However, the single-strand nick at this location does not influence the overall DNA bending angle.6 We utilized the same DNA construct and similar crystallization conditions as Rice and co-workers to solve the crystal structure of the scIHF2–DNA complex (Figure 5(a); Table 1). With the exception of the amino acid residues flanking the N and Ctermini, and linker regions where IHF and scIHF2 differ in primary sequence (see Figure 1(a)), the structures of the two complexes appear to be nearly identical (Figure 5(b)). Least-squares superposition of the IHF–DNA and scIHF2–DNA models (2825 atoms: entire DNA + 173 amino acid residues) yielded an r.m.s.d. in atomic position of only 0.67 Å. Importantly, protein–DNA contacts between the two complexes appear to be nearly the same, consistent with the very similar biochemical and functional properties of IHF and scIHF2. The temperature (B) factors in the scIHF2-DNA complex are significantly higher for the flanking regions of the DNA compared to the 13 bp central zone, where extensive protein–base contacts between the sites of intercalation apparently reduce
Metal-mediated Protein-induced DNA Bending
DNA mobility (Table 1; Figure 5(a)). The mean Bfactor for the 11.5 bp left DNA arm is slightly higher than that of the 10.5 bp right arm, which may relate to the relatively more extensive protein–DNA contacts on the right side. At the left arm, the αK45 side-chain takes part in DNA binding by forming direct, as well as water-mediated, hydrogen bonds with two phosphate groups at the mouth of the minor groove (Figure 5(c)). This configuration is further supported through stacking of the αK45 side-chain with that of αQ43, which is in turn hydrogen-bonded to that of αD53. Least-squares superposition of the protein and DNA components of the scIHF2–DNA and the scIHF2-K45αE–DNA models gave an r.m.s.d. in atomic position of 0.32 Å. Thus, the two complexes are essentially identical, with the exception of local alterations in structure due to the presence of glutamate in place of the αK45 side-chain (Figure 5(d) and (e)). Interestingly, the orientation of the αE45 side-chain is very similar to that of αK45. Strikingly, two divalent metal-binding sites appear in the scIHF2-K45αE–DNA complex, which are not observed in the scIHF2–DNA complex. A manganese ion coordinated to the αE45 carboxylate group and the adjacent DNA phosphate group serves as a cation bridge, which apparently compensates for an otherwise unfavorable electrostatic interaction, allowing a native-like, protein-bound DNA configuration to persist. In contrast to our expectations based on EMSAs, the αK45E substitution has a more pronounced effect on local protein as compared to DNA structure. The presence of the negative-charged αE45 side-chain causes repositioning of the adjacent αQ43 side-chain, which engages in a hydrogen bond with the more distant peptide backbone NH group of αL54. This reorientation of the αQ43 side-chain apparently opens up an additional electronegative pocket, formed by the side-chains of αD53 and αE45, and the opposing DNA phosphate groups, which is occupied by a second manganese ion. Thus, divalent metal binding to the electronegative zones proximal to the DNA phosphate backbone, which would appear to otherwise disrupt association of the left arm in the scIHF2-K45αE mutant, stabilizes DNA binding and bending. At the same time, the Bfactor elevation of the left versus the right DNA arms observed in the scIHF2–DNA complex is actually threefold greater in the scIHF2-K45αE–DNA complex (Table 1). Therefore, in spite of the divalent metal-mediated stabilization of left DNA arm association, this region appears to still have increased mobility relative to the scIHF2–DNA complex. Modulation of scIHF2-K45αE–DNA complex conformation by divalent metal ions and supercoiling Our structural analyses, in conjunction with AFM and the biochemical data, indicate that a scIHF2-K45αE–DNA complex can adopt two stable
Metal-mediated Protein-induced DNA Bending
735
Figure 4. AFM analysis. About 20 high-resolution images of attL– protein complexes obtained with (a) wild-type IHF, (b) scIHF2, and (c)– (e) scIHF2-K45αE were analyzed and the degree of DNA bending was determined. In (d) and (e), the concentration of magnesium ions present in the binding reactions with scIHF2-K45αE is indicated. One representative example for each protein –attL complex is presented with tangents used for the measurements indicated. Images are 100 nm × 100 nm in size.
conformational states. In the “open” state, the left DNA arm is mostly detached from the protein body, thus leading to a significantly smaller degree of overall DNA bending. In the “closed” state, two divalent metal ions stabilize left arm interactions with the protein body, which results in the more severe DNA bending observed in the crystal structure. However, the H′-DNA used in our crystallographic studies contained a nick in one strand near a site of proline intercalation. There-
fore, in order to establish that scIHF2-K45αE can also adopt the closed conformational state with covalently closed H′-DNA, we incubated either scIHF2 or scIHF2-K45αE with H′-DNA in the presence of magnesium ions, and analyzed complex formation through EMSA. As indicated by the change in retardation factor (Rf), small amounts of magnesium ions in the binding and electrophoresis buffer almost completely reverse the supershift of the scIHF2-K45αE-H′
736
Metal-mediated Protein-induced DNA Bending
Figure 5. Protein–DNA interactions in the crystal structures of the scIHF2-DNA and scIHF2-K45αE–DNA complexes. (a) The scIHF2–DNA complex, with protein regions corresponding to the α and β-subunits of wild-type IHF colored green and blue, respectively. Regions differing in primary sequence with respect to IHF are colored red, including the linker amino acids (thick tubes) used to fuse the α and β-subunits. The boxed area associated with the left DNA arm corresponds to the section viewed in (c), (d) and (e). (b) Superposition of the scIHF2–DNA (yellow) and IHF–DNA (αsubunit, magenta; β-subunit, cyan; DNA, blue) models. (c) and (d) Simulated annealing, Fo − Fc omit maps (blue), contoured at 4σ and 3σ, superimposed on the (c) scIHF2–DNA and (d) scIHF2-K45αE-DNA models, respectively. Residues αQ43 and αK45 of scIHF2–DNA and αQ43 and αE45 of scIHF2-K45αE–DNA were omitted. An anomalous difference map (brown), contoured at 3σ, denotes the positions of manganese ions (cyan spheres) in (d). Water molecules are shown as red spheres, and hydrogen and coordinate bonds are indicated by broken lines. (e) Structural comparison of the scIHF2-DNA (yellow) and scIHF2-K45αE-DNA (green) complexes in stereo view. Mn2+ associated with the αK45E mutant are shown as cyan spheres.
complex relative to that of scIHF2-H′ analyzed under the same conditions (Figure 6(a)). In order to provide further evidence that the presence of magnesium ions induces the formation of a closed conformational state between scIHF2-K45αE and an intact cognate site, we repeated the AFM analysis with attL in the presence of nanomolar to micromolar concentrations of metal ions and found that, under these conditions, the DNA bending angle obtained with scIHF2-K45αE increased to values observed with scIHF2 (Figure 4(d) and (e) data not shown). For example, the mean bending angles obtained with scIHF2-K45αE in the presence of 200 nM or 20 nM Mg2+ were 114(±15)° and 111(±22)°, respectively, and thus almost identical with that determined for scIHF2 in the absence of magnesium (114(±15)°). It is important to note that the degree of DNA bending induced by scIHF2 was not affected by the presence of magnesium ions (data not shown). These results indicate that a nicked DNA backbone is not a prerequisite for the formation of the closed conformational state in a scIHF2-K45αE– DNA complex. In addition, this finding raised the interesting possibility of modulating the activity of
scIHF2-K45αE in recombination reactions through a metal ion-mediated increase in DNA bending. As shown in Figure 6(b), small amounts of magnesium ions do indeed affect the activity of scIHF2-K45αE in integrative recombination. Quantification of recombination efficiencies revealed an increase from about 20% to 40%, while the activity of scIHF2 (about 80%) is unaffected (Figure 6(c)). Interestingly, this metal ion-mediated activation was restricted to integrative recombination, since we could not observe an effect of metal ions on excisive recombination with topologically relaxed pλER (data not shown).
Discussion The modular mode of DNA bending by scIHF2K45αE can be exploited to study the topography of snups. This is exemplified here for the assembly of λ intasomes, which is required for synapsis of att sites and for subsequent progression through the entire recombination reaction. The paths of the recombining DNA substrates differ in shape in intasomes configured for integration versus excision. A number of elegant studies in which IHF was replaced by
Metal-mediated Protein-induced DNA Bending Table 1. X-ray data collection and refinement statistics scIHF2–DNA scIHF2-K45αE–DNA A. Data collection Resolution (Å) Resolution of last shell (Å) Redundancy (last shell) No. unique hkl Completion % (last shell) Rmerge % (last shell) Space group Unit cell dimensions a (Å) b (Å) c (Å) B. Refinement Resolution (Å) Resolution (last shell Å) R-factor % (last shell) Free R-factor % (last shell) Reflections work/free No. atoms in model (mean B-factor Å2) Total Protein DNA Left arm Right arm Solvent r.m.s.d. from ideality Bond lengths (Å) Bond angles (deg.)
44.5–2.41 2.54–2.41 6.3 (3.9) 18,596 99.0 (94.5) 6.2 (41.6) P212121
45.0–2.72 2.87–2.72 7.7 (7.9) 12,530 97.1 (95.9) 6.5 (47.6) P212121
47.37 55.31 177.83
47.24 54.44 177.87
40.0–2.41 2.52–2.41 23.2 (33.3) 26.8 (34.7) 17,961/587
40.0–2.72 2.84–2.72 23.7 (35.1) 27.4 (39.7) 12,130/381
3318 (48.9) 1610 (48.2) 1426 (50.4) 470 (57.2) 426 (54.6) 282 (45.7)
3207 (52.5) 1610 (53.0) 1426 (52.8) 470 (64.5) 426 (56.7) 171 (44.8)
0.005 1.11
0.004 0.85
other architectural proteins or sequence-directed bends already identified important intasomal DNA architectural requirements. 13,20–24 We focus our discussion here first on excisive recombination, where scIHF2-K45αE functions as an accessory factor only when attR is negatively supercoiled, regardless of the presence or absence of divalent metal ions. The functionally important cognate site for IHF in attR is H2. It is located in close proximity to the two target sites for the second DNA-bending factor, Xis. Earlier reports identified the attR intasome as a more delicate structure than the attL intasome, and an inspection of recent modeling of the DNA path in the attR intasome clearly predicted a need for extreme DNA-bending at H2. 16,22–24 Our data strongly support this model and reveal that an open scIHF2-K45αE-H2 conformation on attR, with overall DNA bending angles significantly smaller than those found in complexes with scIHF2 or wildtype IHF, is not functional. In this case, DNA supercoiling is most likely required to overcome a thermodynamic barrier imposed on the formation of a stable closed scIHF2-K45αE–H2 complex, arising from the combined effects of the stiffness of short DNA segments and weakened interaction between the left DNA arm and the protein body of scIHF2K45αE.25 In the absence of divalent metal ions, but with the aid of supercoiling, the α45E side-chain could maintain an orientation very similar to that observed in the crystal structure, interacting with DNA phosphate groups through solvent bridges, as opposed to metal ions. Glutamate-phosphate water
737 bridges have been found for histone–DNA interactions in the nucleosome core particle.26 Another possibility for the observed dependency of scIHF2K45αE on supercoiling in attR is that its binding to H2 might be sensitive to the periodicity of the helical repeat, which, in turn, is affected by the topological state of DNA.
Figure 6. Divalent metal ion-induced conformational switch in scIHF2-K45αE–DNA complexes. (a) ScIHF2 and scIHF2-K45αE were incubated with radiolabeled H′ DNA in the absence or in the presence of various concentrations of MgCl2, as indicated. EMSA was performed in TB buffers containing the same concentration of MgCl2. The retardation factors (Rf = dc/df) were determined and normalized relative to scIHF2. At the bottom, Rf values obtained through complex formation at identical protein:DNA stoichiometries are indicated. (b) Effect of MgCl2 on intramolecular recombination. Supercoiled pλIR was incubated with Int and either scIHF2 or scIHF2-K45αE in the presence of increasing concentrations of MgCl2 that matched the range in (a), and digested with NcoI before agarose gel electrophoresis. (c) Quantification of recombination efficiencies. The amount of product and remaining substrate DNA shown in (b) was determined and plotted against the actual amount of magnesium ions present in each reaction.
738 The fact that IHF and scIHF2 activate recombination on topologically relaxed attR indicates that wild-type protein–DNA interactions are sufficiently strong in the absence of supercoiling. However, there appears to be no strict requirement for severe IHF-induced bending inside the attL intasome, where an open scIHF2-K45αE–H′ complex seems sufficient to configure a functional snup. Our results also provide insight into snup formation during integrative recombination. It is worth pointing out again that, even with wild-type IHF, negative supercoiling of attP is required to configure a functional intasome. Our finding that scIHF2K45αE is barely active in integrative recombination, but is significantly stimulated by divalent metal ions (Figure 6(b)), indicates that a closed scIHF2-K45αE– DNA complex with a high degree of DNA bending must be a prerequisite to permit necessary Int–DNA interactions within attP. Integrative and excisive recombination differ most notably in their requirement for the IHF cognate site H1 and, again, inspection of recent models for integrative intasomes indicates that a closed scIHF2-K45αE–H1 complex is most likely required there.16 In addition to the proposed effect on the degree of bending at H1 by scIHF2-K45αE, a more complex dependence involving cooperative protein binding effects, including the two other IHF binding sites, is possible. We think that scIHF2-K45αE is a valuable tool to investigate nucleoprotein complex formation in general. It could be used to probe the topography of specific snups involved in DNA transactions in vitro and in vivo, in both prokaryotic and eukaryotic cells. This is already exemplified by our earlier finding that scIHF3-K45αE, which contains one shorter peptide linker but behaves very much like scIHF2-K45αE in biochemical and functional assays, exhibits a defect in its function as an initiation factor for pSC101 replication in vivo.19 It is likely, therefore, that a functional pSC101 origin recognition complex must include a highly bent, closed IHF–DNA complex in vivo. An interesting future application could include the generation of E. coli strains that express either scIHF2 or scIHF2-K45αE instead of wild-type IHF, and perform comparative gene expression profiling in order to probe the topography of IHF-dependent snups involved in gene regulation. Engineering of metal-binding sites at the protein– DNA interface holds promise for the design of switches in other DNA-binding systems. As we have applied the scIHF2 variant in this study, DNA binding or conformational alteration that is dependent on divalent metal concentration can be exploited for probing structure–function relationships in other nucleoprotein assemblies. It would be informative to investigate the general effect of substituting basic amino acid residues that are involved in phosphate group contacts with acidic residues in other DNA-binding proteins. Our results suggest that such changes may give rise to a
Metal-mediated Protein-induced DNA Bending
requirement for a divalent metal-mediated protein–DNA interaction for stabilizing the nativelike, DNA-bound configuration. Such metal-dependent modulation could be explored for the design of nano-sized DNA switch devices.27
Materials and Methods Purification of scIHF2-K45αE Expression vector pETscIHF2-K45αE was derived from pETscIHF3-K45αE, and scIHF2 and scIHF2-K45αE were purified as described.19 Protein concentration was determined using BIO-RAD protein assay kits with IgG as a standard. Gel mobility-shift assays Binding affinities were determined using 32P -labeled oligonucleotides as described.19 Plasmid pWSRGFP contains attR flanked by a set of restriction sites.28 A portion (272 fmol) of supercoiled or EcoRI-cleaved DNA was incubated with scIHF2 and scIHF2-K45αE at 90-fold and 400-fold molar excess, respectively, in 10 mM Tris– HCl (pH 7.9), 50 mM NaCl, 10 mM MgCl2, 1 mM DTT, and 100 μg/ml of bovine serum albumin (BSA). After 30 min at room temperature, DNA was digested with HindIII and NotI. Samples were loaded onto a 4% (w/v) polyacrylamide gel, which was pre-run at 150 V for 1 h at room temperature. Reactions that contained MgCl2 were analyzed on 15% polyacrylamide gels in 0.5 × TBM buffer (45 mM Tris– borate (pH 8.3) plus MgCl2 at 0, 19 μM, 38 μM, or 76 μM). Gels were pre-run in TBM buffer at 210 V at 4 °C for 1 h. The electrophoresis buffer was constantly exchanged during the run. Images were quantified using a Bio-Rad Molecular Imager System. Recombination assays Reactions were performed essentially as described, using pλIR for integrative intramolecular recombination and pλER for excisive recombination.19,28 In 25 μl of buffer containing 84 fmol of att sites, a 90-fold and 400-fold molar excess of scIHF2 and scIHF2-K45αE, respectively, was used plus 1 pmol of Int. We used pWSRGFP and pCMVattL as DNA substrates for intermolecular excisive recombination.18,28 DNA was incubated with a 30, 90, and 400-fold excess of IHF, scIHF2 and scIHF2-K45αE, respectively, and 1 pmol of Int plus 3 pmol of Xis. Recombination efficiency was quantified after staining with ethidium bromide using a Bio-Rad Molecular Imager System. The effect of divalent cations was assayed in a buffer containing 44 mM Tris (pH 7.9), 60 mM KCl, 50 μg/ ml of BSA, 5 mM spermidine, with the indicated amounts of MgCl2. AFM imaging Binding reactions contained 0.246 nM attL and a 10, 30, and 100-fold excess of IHF, scIHF2, and scIHF2-K45αE, respectively, in 0.5 × TBE buffer, and were incubated at room temperature for 40 min. Droplets of 30 μl were spotted onto AP-mica or Glu-mica and incubated for
739
Metal-mediated Protein-induced DNA Bending
15 min. Samples were washed with deionized water and dried under pure nitrogen. Imaging was done with a Veeco Dimension 3000 AFM and Nanoscope IIIa controller (Digital Instruments, Santa Barbara, CA, U.S.A.). PPPNCH silicon tips (NANOSENSORS, Switzerland) were used in tapping-mode scanning (1.0 Hz). Images were analyzed with the Nanoscope software and the free software ImageJ§. AP-mica was prepared using the recently developed APTES solution method.29 Glu-mica was prepared by incubating AP-mica in aqueous glutaraldehyde solution (0.1%–5% v/v) for 10 min, and subsequently washed with deionized water and dried under pure nitrogen.30 Variation in the concentration of glutaraldehyde was used to optimize adherence of DNA to the mica surface. Crystal preparation and structure solution The scIHF–DNA complex was reconstituted by mixing purified scIHF protein and H′ DNA with 1:1 stoichiometry in a buffer containing 10 mM Hepes (pH 7.0), 0.02 M NaCl, 0.1 mM EDTA, and 8% (v/v) glycerol to give a final concentration of 14 mg/ml of the protein– DNA complex. H′ DNA was assembled by reannealing three HPLC-purified oligonucleotides (from SigmaGenosys, Germany): CGGTGCAACAAATTGATAAGCAATGCTTTTTTGGC GGCCAAAAAAGCATT GCTTATCAATTTGTTGCACC
Crystals of the scIHF2–DNA complex were grown via vapor diffusion at 18 °C in droplets made by mixing a 14 mg/ml complex solution with an equal volume of well solution containing 50 mM Tris (pH 7.5), 15% (w/v) PEG5000-MME, 10% glycerol, 50 mM NaCl, 30 mM MnCl2. For crystallization of scIHF2-K45αE–DNA complex, a modified well solution was used containing 50 mM Tris (pH 7.5), 25% PEG5000-MME, 20% glycerol, 50 mM NaCl, 7.5 mM MnCl2. Large crystals of the scIHF2–DNA complex and initial crystals of the scIHF2-K45αE–DNA complex were obtained by microseeding using scIHF2– DNA crystals. For stabilization and improvement of diffraction capability, crystals were transferred into 50 mM Tris (pH 7.5), 15% PEG5000-MME, 10% glycerol, 50 mM NaCl, 30 mM MnCl2, 10% (v/v) 2-methyl-2,4-pentanediol for scIHF2DNA, or 50 mM Tris (pH 7.5), 15% PEG5000-MME, 10% glycerol, 50 mM NaCl, 7.5 mM MnCl2, 15% (v/v) 2methyl-2,4-pentanediol for scIHF2-K45αE–DNA. Crystals were flash-cooled in liquid N2 and transferred into a N2 gas stream at −170 °C for collection of X-ray diffraction intensities on a Rigaku MicroMax 007 X-ray diffractometer equipped with a Rigaku Raxis IV++ image plate. Diffraction data were processed using MOSFLM and SCALA from the CCP4 suite.31,32 The scIHF2–DNA structure was solved by molecular replacement with routines from the CCP4 suite using the wild-type IHF–DNA model.5 Structural refinement was carried out with the programs O and CNS.33,34 In the final stages of refinement, specific B-form sugar pucker restraints present in the CNS DNA parameter file were removed.5 Molecular structure Figures were prepared with PyMOL (DeLano Scientific LLC, San Carlos, CA, USA∥).
§ http://rsb.info.nih.gov/ij/ ∥ http://pymol.sourceforge.net/
Protein Data Bank accession codes Atomic coordinates have been deposited in the RCSB Protein Data Bank with accession codes 2IIE (scIHF2DNA) and 2IIF (scIHF2-K45αE-DNA).
Acknowledgements Very special thanks go to A. Segall for purified Int and Xis proteins, and to D. Esposito for purified wild-type IHF. This work was funded through University Research grants provided by NTU (to P.D. and C.A.D.) and NUS (to J.Y.) and an ARC research grant from the Singapore Ministry of Education (to C.A.D. and P.D.).
References 1. Echols, H. (1986). Multiple DNA-protein interactions governing high-precision DNA transactions. Science, 233, 1050–1056. 2. Goosen, N. & van de Putte, P. (1995). The regulation of transcription initiation by integration host factor. Mol. Microbiol. 16, 1–7. 3. Nash, H. A. (1996). The HU and IHF proteins: accessory factors for complex protein-DNA assemblies. In Regulation of Gene Expression in E. coli (Lin, E. C. C. & Lynch, A. S., eds), R.G. Landes Company, Austin, Texas, USA. pp. 149–179. 4. Goodrich, J. A., Schwartz, M. L. & McClure, W. R. (1990). Searching for and predicting the activity of sites for DNA binding proteins: compilation and analysis of the binding sites for Escherichia coli integration host factor (IHF). Nucl. Acids Res. 18, 4993–5000. 5. Rice, P. A., Yang, S.-W., Mizuuchi, K. & Nash, H. A. (1996). Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn. Cell, 87, 1295–1306. 6. Lorenz, M., Hillisch, A., Goodman, S. D. & Diekmann, S. (1999). Global structure similarities of intact and nicked DNA complexed with IHF measured in solution by fluorescence resonance energy transfer. Nucl. Acids Res. 27, 4619–4625. 7. Seong, G. H., Kobatake, E., Miura, K., Nakazawa, A. & Aizawa, M. (2002). Direct atomic force microscopy visualization of integration host factor-induced DNA bending structure of the promoter regulatory region on the Pseudomonas TOL plasmid. Biochem. Biophys. Res. Commun. 291, 361–366. 8. Ellenberger, T. & Landy, A. (1997). A good turn for DNA: the structure of integration host factor bound to DNA. Structure, 5, 153–157. 9. Swinger, K. K. & Rice, P. A. (2004). IHF and HU: flexible architects of bent DNA. Curr. Opin. Struct. Biol. 14, 28–35. 10. Arfin, S. M., Long, A. D., Ito, E. T., Tolleri, L., Riehle, M. M., Paegle, E. S. & Hatfield, G. W. (2000). Global gene expression profiling in Escherichia coli K12. The effects of integration host factor. J. Biol. Chem. 275, 29672–29684. 11. Mangan, M. W., Lucchini, S., Danino, V., Croinin, T. O., Hinton, J. C. D. & Dorman, C. J. (2006). The integration host factor (IHF) integrates stationary-phase and virulence gene expression in Salmonella enterica serovar typhimurium. Mol. Microbiol. 59, 1831–1847.
740
Metal-mediated Protein-induced DNA Bending
12. Azaro, M. A. & Landy, A. (2002). λ Int and the λ Int family. In Mobile DNA II (Craig, N. L., Craigie, R., Gellert, M. & Lambowitz, A., eds), pp. 118–148, ASM Press, Washington, DC. 13. Richet, E., Abcarian, P. & Nash, H. A. (1986). The interaction of recombination proteins with supercoiled DNA: defining the role of supercoiling in lambda integrative recombination. Cell, 46, 1011–1021. 14. Richet, E., Abcarian, P. & Nash, H. A. (1988). Synapsis of attachment sites during lambda integrative recombination involves capture of a naked DNA by a protein-DNA complex. Cell, 52, 9–17. 15. Biswas, T., Aihara, A., Radman-Livaja, M., Filman, D., Landy, A. & Ellenberger, T. (2005). A structural basis for allosteric control of DNA recombination by lambda integrase. Nature, 345, 475–485. 16. Radman-Livaja, M., Biswas, T., Ellenberger, T., Landy, A. & Aihara, H. (2006). DNA arms do the legwork to ensure the directionality of lambda site-specific recombination. Curr. Opin. Struct. Biol. 16, 42–50. 17. Lorbach, E., Christ, N., Schwikardi, M. & Dröge, P. (2000). Site-specific recombination in human cells catalyzed by phage lambda integrase mutants. J. Mol. Biol. 296, 1175–1181. 18. Corona, T., Bao, Q., Christ, N., Schwartz, T., Li, J. & Dröge, P. (2003). Activation of site-specific DNA integration in human cells by a single chain integration host factor. Nucl. Acids Res. 31, 5140–5148. 19. Bao, Q., Christ, N. & Dröge, P. (2004). Single-chain integration host factors as probes for high-precision nucleoprotein complex formation. Gene, 343, 99–106. 20. Snyder, U. K., Thompson, J. F. & Landy, A. (1989). Phasing of protein-induced DNA bends in a recombination complex. Nature, 341, 255–257. 21. Goodman, S. D. & Nash, H. A. (1989). Functional replacement of a protein-induced bend in a DNA recombination site. Nature, 341, 251–254. 22. Goodman, S. D., Nicholson, S. C. & Nash, H. A. (1992). Deformation of DNA during site-specific recombination of bacteriophage lambda: replacement of IHF protein by HU protein or sequence-directed bends. Proc. Natl Acad. Sci. USA, 89, 11910–11914. 23. Goodman, S. D. & Kay, O. (1999). Replacement of
24.
25. 26.
27.
28.
29.
30.
31. 32. 33.
34.
integration host factor protein-induced DNA bending by flexible regions of DNA. J. Biol. Chem. 274, 37004–37011. Segall, A. M., Goodman, S. D. & Nash, H. A. (1994). Architectural elements in nucleoprotein complexes: interchangeability of specific and non-specific DNA binding proteins. EMBO J. 13, 4536–4548. Kahn, J. D., Yun, E. & Crothers, D. M. (1994). Detection of localized DNA flexibility. Nature, 368, 163–166. Davey, C. A., Sargent, D. F., Luger, K., Maeder, A. W. & Richmond, T. J. (2002). Solvent mediated interactions in the structure of the nucleosome core particle at 1.9 Å resolution. J. Mol. Biol. 319, 1097–1113. Shen, W., Bruist, M. F., Goodman, S. D. & Seeman, N. C. (2004). A protein-driven DNA device that measures the excess binding energy of proteins that distort DNA. Angew. Chem. Int. Ed. 43, 4750–4752. Christ, N., Corona, T. & Dröge, P. (2002). Site-specific recombination in eukaryotic cells mediated by mutant lambda integrases: implications for synaptic complex formation and the reactivity of episomal DNA segments. J. Mol. Biol. 319, 305–314. Liu, Z., Li, Z., Zhou, H., Wei, G., Song, Y. & Wang, L. (2005). Immobilization and condensation of DNA with 3-aminopropyltriethoxysilane studied by atomic force microscopy. J. Microsc. 218, 233–239. Wang, H., Bash, R., Yodh, J. G., Hager, G. L., Lohr, D. & Lindsay, S. M. (2002). Glutaraldehyde modified mica: a new surface for atomic force microscopy of chromatin. Biophys. J. 83, 3619–3625. Leslie, A. G. (1999). Integration of macromolecular diffraction data. Acta Crystallog. sect. D, 55, 1696–1702. Collaborative Computational Project Number 4 (1994). The CCP4 suite: programs for protein crystallography. Acta. Crystallog. sect. D, 50, 760–763. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Rystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921.
Edited by K. Morikawa (Received 3 August 2006; received in revised form 26 September 2006; accepted 27 September 2006) Available online 3 October 2006
J. Mol. Biol. (2007) 367, 731–740
A Divalent Metal-mediated Switch Controlling Protein-induced DNA Bending† Qiuye Bao 1 , Hu Chen 2 , Yingjie Liu 2 , Jie Yan 2 , Peter Dröge 1 ‡⁎ and Curt A. Davey 3 ‡⁎ 1
Division of Genomics and Genetics, School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, Singapore 2
Department of Physics, National University of Singapore, 2 Science Drive, Singapore 117542, Singapore 3
Division of Structural and Computational Biology, School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, Singapore *Corresponding authors
Architectural proteins that reconfigure the paths of DNA segments are required for the establishment of functional interfaces in many genomic transactions. A single-chain derivative of the DNA architectural protein integration host factor was found to adopt two stable conformational states in complex with a specific DNA target. In the so-called open state, the degree of protein-induced DNA bending is reduced significantly compared with the closed state. The conformational switch between these states is controlled by divalent metal binding in two electronegative zones arising from the lysine-to-glutamate substitution in the protein body proximal to the phosphate backbone of one DNA arm. We show that this switch can be employed to control the efficiency of site-specific recombination catalyzed by λ integrase. Introduction of acidic residues at the protein–DNA interface holds potential for the design of metal-mediated switches for the investigation of functional relationships. © 2006 Elsevier Ltd. All rights reserved.
Keywords: divalent metal; DNA architectural proteins; DNA topology; integration host factor; nucleoprotein complex formation
Introduction The assembly of specialized nucleoprotein structures (snups) inside living cells is a key regulatory step in many DNA transactions, such as site-specific recombination, replication, and transcription.1 In many cases, snup assembly is orchestrated by DNA architectural proteins that alter the trajectory of DNA segments in a more or less defined way through DNA bending. However, very little is known about external factors, such as the topological state of DNA, that could govern the dynamics of functional snup formation in a genome. The integration host factor (IHF) is a key DNA architectural protein in Escherichia coli.2,3 IHF is † P.D. dedicates this work to the memory of Nicholas R. Cozzarelli. ‡ P.D. and C.A.D. contributed equally to this work. Abbreviations used: snup, specialized nucleoprotein structure; IHF, integration host factor; AFM, atomic force microscopy; EMSA, electrophoretic mobility-shift assay. E-mail addresses of the corresponding authors: [email protected]; [email protected]
composed of two homologous subunits, the α and the β-subunit, and displays limited DNA sequence specificity.4 The crystal structure of IHF in complex with the phage λ H′ site, as well as fluorescent resonance energy transfer (FRET) analyses and visualization of IHF–DNA complexes through atomic force microscopy (AFM), revealed strong (120–160°) protein-induced DNA bending.5–7 Bending is achieved mainly by the intercalation of one conserved proline residue from each subunit of the heterodimer into the minor groove. The resulting Uturn-like DNA conformation is stabilized by electrostatic interactions around the protein body.8,9 IHF is involved in the regulation of more than 100 genes in Gram-negative bacteria and it is an essential cofactor in phage λ site-specific recombination, where the protein serves an architectural role during the assembly of snups.10–12 The 240 bp comprising phage attachment (att) site attP is one of the two recombination sequences in the integrative pathway and harbors three specific IHF-binding sites that must be occupied by IHF to generate a functional snup, the so-called integrative intasome. Intasome assembly strictly requires negative DNA supercoiling of attP.13 The intasome then captures
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
732 the protein-free 21 bp attB to form a synaptic complex in which two successive rounds of DNA strand exchange are catalyzed by phage λ integrase (Int).14 Excisive recombination is genetically the exact reversal of integration and employs hybrid attachment sites attL and attR as substrates. Excision also requires IHF, but involves only two binding sites in complex with and bent by IHF. In contrast to the integrative pathway, excision does not depend on negative supercoiling of attL or attR. However, a second DNA-bending co-factor, the phage-encoded excisionase (Xis), is required and binds specifically to two sites on attR. The general function of IHF and Xis in λ recombination is to construct a specific DNA trajectory within each of the three intasomes, which ultimately enables the heterobivalent Int to establish contacts between separate DNA sites simultaneously. Recent intasome models, based on co-crystal structures of Int tetramers in complexes with core and arm-binding sites, further highlighted the crucial role of IHF-induced DNA bending in the formation of functional intasomes.15,16 We recently transferred the phage λ recombination system to mammalian cells and engineered a single-chain IHF, named scIHF2, which is functional in mammalian cells.17 This was achieved by insertion of almost the entire α-subunit into the β-subunit at position 39 using two short peptide linkers.18 Biochemical and functional assays confirmed that scIHF2 behaves like its heteromeric parent. During construction of scIHF2, we identified a variant (scIHF3-K45αE) that carries glutamate instead of lysine at position 45 of the α-subunit. In addition, one of the two linkers is shortened in scIHF3-K45αE. We found that scIHF3-K45αE is nearly inactive in promoting integrative recombination in vitro, while remaining fully active as a co-factor for excisive recombination on negatively supercoiled substrates. The protein also exhibits a defect in its function as an initiation factor for pSC101 replication in vivo.19 In order to analyze this interesting phenotype in more detail, we introduced the αK45E substitution into scIHF2 and performed a detailed functional and structural analysis. This led to the identification of a novel, controllable modular mode of proteininduced DNA bending. In addition, our results obtained with the phage λ site-specific recombination system provide valuable insight into possible dynamics of functional snup formation in general, and how this can be governed by an intricate interplay between DNA architectural proteins and external factors.
Results Biochemical and functional properties of scIHF2-K45αE We introduced the K45αE substitution into scIHF2 (Figure 1(a)) and demonstrated that purified
Metal-mediated Protein-induced DNA Bending
Figure 1. Sequence of scIHF2-K45αE and EMSA. (a) Numbering of residues follows that of the heteromeric IHF. The two short linkers are shaded. The αK45E substitution and corresponding residue in the β-subunit are underlined. (b) EMSA analysis of complexes formed between the 34 bp long, radio-labeled H′ and either scIHF2 or scIHF2-K45αE.
scIHF2-K45αE in complex with a 34 bp oligonucleotide comprising the H′ site migrates more slowly than the scIHF2-H′ complex (Figure 1(b)). Hence, the substitution at position 45 in the α-subunit is solely responsible for the altered complex mobility, which was observed earlier with scIHF3-K45αE.19 We subsequently determined the Kd values of scIHF2-K45αE for all three IHF-specific binding sites on attP as <2 nM at an ionic strength <100 mM NaCl, which are within a factor of 2 of the respective values determined for scIHF2 (data not shown). We then tested scIHF2-K45αE in in vitro recombination assays using protein-to-DNA stoichiometries that, with each variant of IHF tested, lead to at least 80% complex formation on individual IHF target sites. We found that scIHF2-K45αE is marginally active as a co-factor in intramolecular integrative recombination on supercoiled pλIR (Figure 2(a), lane 3), but it triggers excisive recombination on supercoiled pλER that contains attL and attR instead of attB and attP, respectively, to an extent comparable with either scIHF2 or wild-type IHF (data not shown). However, we discovered that the accessory role of scIHF2-K45αE in excisive recombination depends entirely on negative supercoiling of the DNA substrate. No recombination product is detected when scIHF2-K45αE is tested with linearized pλER, which contains directly repeated attR and attL as substrate, under conditions where both scIHF2 and wild-type IHF are fully active (data not shown). We therefore decided to use intermolecular recombination in order to dissect whether attR or attL, or both, requires superhelical tension with scIHF2-K45αE. As shown in Figure 2(b), lanes 5 to 7, wild-type IHF, scIHF2, and scIHF2-K45αE serve as cofactors when supercoiled attR is recombined with linearized attL. However, only wild-type IHF and scIHF2 activate recombination when linearized attR is incubated with supercoiled attL (Figure 2(c), lanes 5 to 7). Since we used a fourfold greater amount of scIHF2-K45αE than scIHF2 in our recombination
733
Metal-mediated Protein-induced DNA Bending
Figure 2. In vitro recombination assays. (a) Intramolecular recombination on supercoiled pλIR. (b) Intermolecular recombination between supercoiled attR and linearized attL. Note that two recombination products are expected, since the supercoiled attR substrate DNA is monomeric and dimeric. (c) Intermolecular recombination between supercoiled attL and linearized attR. Recombination products were analyzed through agarose gel electrophoresis.
assays, and since it is known that large amounts of IHF can inhibit excisive recombination, we successively reduced the amount of scIHF2-K45αE about fivefold but were still unable to detect recombination activity on linearized pλER (data not shown). The absolute dependency of scIHF2-K45αE on topologically underwound attR in excisive recombination led us to analyze the complexes formed on attR through electrophoretic mobility-shift assay (EMSA). We used a DNA fragment in which the two cognate sites present in attR, H1 and H2, are positioned at about the same distance from the ends. We observe two distinct complexes that migrate faster than those formed between scIHF2 and attR (Figure 3). Further, these differences are detectable
regardless of whether scIHF2-K45αE is first bound to supercoiled attR followed by restriction digest, or vice versa (left-hand and right-hand panel, respectively). The opposing differences in electrophoretic mobility of the scIHF2-K45αE-H′ and the scIHF2K45αE-attR complexes in comparison with the corresponding scIHF2–DNA complexes is apparently a consequence of the more intricate influence of intrinsic curvature of longer DNA segments, such as attL, on overall mobility when coupled with protein-induced bending. Nonetheless, these results confirm that scIHF2-K45αE is able to bind to H1 and H2 on linearized attR, and reveal that significant conformational differences must exist between scIHF2-K45αE-attR complexes and those formed by scIHF2 and attR. The crystal structure of IHF in complex with H′ showed that the lysine side-chain at position 45 of the α-subunit interacts with phosphate groups of the A-tract, most likely stabilizing DNA bending.5 Since we substituted this lysine with glutamate and observed altered electrophoretic mobilities of scIHF2-K45αE-DNA complexes, indicative of altered complex conformation, we hypothesized that interactions between the corresponding region of the protein body and the A-tract, which we will subsequently refer to as the left DNA arm, were weakened significantly. This could result in a reduced degree of overall DNA bending that, in turn, affects the electrophoretic mobilities of complexes. In order to obtain more direct evidence for this scenario, we used AFM and determined the degree of DNA bending in images from individual protein–DNA complexes.
Figure 3. EMSA of complex formation on attR. The attR fragment used in this analysis is schematized in the top panel. DNA was either pre-cleaved with EcoRI (right) or the supercoiled plasmid was incubated directly with scIHF2 or scIHF2-K45αE (left), followed by a HindIII/ NotI digest before PAGE.
DNA bending determined by atomic force microscopy A 623 bp attL-carrying DNA fragment was used for AFM imaging. The fragment harbors the H′ site positioned asymmetrically 200 bp from one end.
734 DNA was incubated either with purified wild-type IHF, scIHF2, or scIHF2-K45αE, and the resulting complexes were adsorbed to mica. Naked DNA served as the control. AFM images that showed both a protein signal and DNA bending at the expected DNA region were further analyzed. None of the more than 100 images inspected in the control sample with naked DNA show significant DNA bending together with a protein signal at a corresponding position (not shown). Our analysis of AFM images from wild-type IHF–DNA and scIHF2–DNA complexes reveal mean bending angles of 117(±19)° and 114(±15)°, respectively (Figure 4(a) and (b)). These values are in very good agreement with those from a recent AFM study performed with wild-type IHF on a segment from the TOL plasmid.7 Our analysis using scIHF2K45αE, however, yields a distribution of bending angles with a significantly smaller mean value of 91(±19) °degrees (Figure 4(c)). Together, these results support our hypothesis that the degree of overall DNA bending in scIHF2-K45αE-DNA complexes is reduced significantly, most likely as a result of weakened protein interaction with the left DNA arm that result from the K45αE substitution. In order to investigate this possibility further, we solved the crystal structures of scIHF2 and scIHF2-K45αE in complex with H′ DNA. Crystal structures of scIHF2-DNA and scIHF2-K45αE-DNA complexes The construct used to obtain the crystal structure of the wild-type IHF–DNA complex was composed of a 35 bp DNA containing a nick in one strand adjacent to the proline intercalation site of the βsubunit.5 The presence of the single-strand break was necessary for obtaining well-diffracting crystals, consistent with its involvement in an extensive interparticle contact in the crystal. However, the single-strand nick at this location does not influence the overall DNA bending angle.6 We utilized the same DNA construct and similar crystallization conditions as Rice and co-workers to solve the crystal structure of the scIHF2–DNA complex (Figure 5(a); Table 1). With the exception of the amino acid residues flanking the N and Ctermini, and linker regions where IHF and scIHF2 differ in primary sequence (see Figure 1(a)), the structures of the two complexes appear to be nearly identical (Figure 5(b)). Least-squares superposition of the IHF–DNA and scIHF2–DNA models (2825 atoms: entire DNA + 173 amino acid residues) yielded an r.m.s.d. in atomic position of only 0.67 Å. Importantly, protein–DNA contacts between the two complexes appear to be nearly the same, consistent with the very similar biochemical and functional properties of IHF and scIHF2. The temperature (B) factors in the scIHF2-DNA complex are significantly higher for the flanking regions of the DNA compared to the 13 bp central zone, where extensive protein–base contacts between the sites of intercalation apparently reduce
Metal-mediated Protein-induced DNA Bending
DNA mobility (Table 1; Figure 5(a)). The mean Bfactor for the 11.5 bp left DNA arm is slightly higher than that of the 10.5 bp right arm, which may relate to the relatively more extensive protein–DNA contacts on the right side. At the left arm, the αK45 side-chain takes part in DNA binding by forming direct, as well as water-mediated, hydrogen bonds with two phosphate groups at the mouth of the minor groove (Figure 5(c)). This configuration is further supported through stacking of the αK45 side-chain with that of αQ43, which is in turn hydrogen-bonded to that of αD53. Least-squares superposition of the protein and DNA components of the scIHF2–DNA and the scIHF2-K45αE–DNA models gave an r.m.s.d. in atomic position of 0.32 Å. Thus, the two complexes are essentially identical, with the exception of local alterations in structure due to the presence of glutamate in place of the αK45 side-chain (Figure 5(d) and (e)). Interestingly, the orientation of the αE45 side-chain is very similar to that of αK45. Strikingly, two divalent metal-binding sites appear in the scIHF2-K45αE–DNA complex, which are not observed in the scIHF2–DNA complex. A manganese ion coordinated to the αE45 carboxylate group and the adjacent DNA phosphate group serves as a cation bridge, which apparently compensates for an otherwise unfavorable electrostatic interaction, allowing a native-like, protein-bound DNA configuration to persist. In contrast to our expectations based on EMSAs, the αK45E substitution has a more pronounced effect on local protein as compared to DNA structure. The presence of the negative-charged αE45 side-chain causes repositioning of the adjacent αQ43 side-chain, which engages in a hydrogen bond with the more distant peptide backbone NH group of αL54. This reorientation of the αQ43 side-chain apparently opens up an additional electronegative pocket, formed by the side-chains of αD53 and αE45, and the opposing DNA phosphate groups, which is occupied by a second manganese ion. Thus, divalent metal binding to the electronegative zones proximal to the DNA phosphate backbone, which would appear to otherwise disrupt association of the left arm in the scIHF2-K45αE mutant, stabilizes DNA binding and bending. At the same time, the Bfactor elevation of the left versus the right DNA arms observed in the scIHF2–DNA complex is actually threefold greater in the scIHF2-K45αE–DNA complex (Table 1). Therefore, in spite of the divalent metal-mediated stabilization of left DNA arm association, this region appears to still have increased mobility relative to the scIHF2–DNA complex. Modulation of scIHF2-K45αE–DNA complex conformation by divalent metal ions and supercoiling Our structural analyses, in conjunction with AFM and the biochemical data, indicate that a scIHF2-K45αE–DNA complex can adopt two stable
Metal-mediated Protein-induced DNA Bending
735
Figure 4. AFM analysis. About 20 high-resolution images of attL– protein complexes obtained with (a) wild-type IHF, (b) scIHF2, and (c)– (e) scIHF2-K45αE were analyzed and the degree of DNA bending was determined. In (d) and (e), the concentration of magnesium ions present in the binding reactions with scIHF2-K45αE is indicated. One representative example for each protein –attL complex is presented with tangents used for the measurements indicated. Images are 100 nm × 100 nm in size.
conformational states. In the “open” state, the left DNA arm is mostly detached from the protein body, thus leading to a significantly smaller degree of overall DNA bending. In the “closed” state, two divalent metal ions stabilize left arm interactions with the protein body, which results in the more severe DNA bending observed in the crystal structure. However, the H′-DNA used in our crystallographic studies contained a nick in one strand near a site of proline intercalation. There-
fore, in order to establish that scIHF2-K45αE can also adopt the closed conformational state with covalently closed H′-DNA, we incubated either scIHF2 or scIHF2-K45αE with H′-DNA in the presence of magnesium ions, and analyzed complex formation through EMSA. As indicated by the change in retardation factor (Rf), small amounts of magnesium ions in the binding and electrophoresis buffer almost completely reverse the supershift of the scIHF2-K45αE-H′
736
Metal-mediated Protein-induced DNA Bending
Figure 5. Protein–DNA interactions in the crystal structures of the scIHF2-DNA and scIHF2-K45αE–DNA complexes. (a) The scIHF2–DNA complex, with protein regions corresponding to the α and β-subunits of wild-type IHF colored green and blue, respectively. Regions differing in primary sequence with respect to IHF are colored red, including the linker amino acids (thick tubes) used to fuse the α and β-subunits. The boxed area associated with the left DNA arm corresponds to the section viewed in (c), (d) and (e). (b) Superposition of the scIHF2–DNA (yellow) and IHF–DNA (αsubunit, magenta; β-subunit, cyan; DNA, blue) models. (c) and (d) Simulated annealing, Fo − Fc omit maps (blue), contoured at 4σ and 3σ, superimposed on the (c) scIHF2–DNA and (d) scIHF2-K45αE-DNA models, respectively. Residues αQ43 and αK45 of scIHF2–DNA and αQ43 and αE45 of scIHF2-K45αE–DNA were omitted. An anomalous difference map (brown), contoured at 3σ, denotes the positions of manganese ions (cyan spheres) in (d). Water molecules are shown as red spheres, and hydrogen and coordinate bonds are indicated by broken lines. (e) Structural comparison of the scIHF2-DNA (yellow) and scIHF2-K45αE-DNA (green) complexes in stereo view. Mn2+ associated with the αK45E mutant are shown as cyan spheres.
complex relative to that of scIHF2-H′ analyzed under the same conditions (Figure 6(a)). In order to provide further evidence that the presence of magnesium ions induces the formation of a closed conformational state between scIHF2-K45αE and an intact cognate site, we repeated the AFM analysis with attL in the presence of nanomolar to micromolar concentrations of metal ions and found that, under these conditions, the DNA bending angle obtained with scIHF2-K45αE increased to values observed with scIHF2 (Figure 4(d) and (e) data not shown). For example, the mean bending angles obtained with scIHF2-K45αE in the presence of 200 nM or 20 nM Mg2+ were 114(±15)° and 111(±22)°, respectively, and thus almost identical with that determined for scIHF2 in the absence of magnesium (114(±15)°). It is important to note that the degree of DNA bending induced by scIHF2 was not affected by the presence of magnesium ions (data not shown). These results indicate that a nicked DNA backbone is not a prerequisite for the formation of the closed conformational state in a scIHF2-K45αE– DNA complex. In addition, this finding raised the interesting possibility of modulating the activity of
scIHF2-K45αE in recombination reactions through a metal ion-mediated increase in DNA bending. As shown in Figure 6(b), small amounts of magnesium ions do indeed affect the activity of scIHF2-K45αE in integrative recombination. Quantification of recombination efficiencies revealed an increase from about 20% to 40%, while the activity of scIHF2 (about 80%) is unaffected (Figure 6(c)). Interestingly, this metal ion-mediated activation was restricted to integrative recombination, since we could not observe an effect of metal ions on excisive recombination with topologically relaxed pλER (data not shown).
Discussion The modular mode of DNA bending by scIHF2K45αE can be exploited to study the topography of snups. This is exemplified here for the assembly of λ intasomes, which is required for synapsis of att sites and for subsequent progression through the entire recombination reaction. The paths of the recombining DNA substrates differ in shape in intasomes configured for integration versus excision. A number of elegant studies in which IHF was replaced by
Metal-mediated Protein-induced DNA Bending Table 1. X-ray data collection and refinement statistics scIHF2–DNA scIHF2-K45αE–DNA A. Data collection Resolution (Å) Resolution of last shell (Å) Redundancy (last shell) No. unique hkl Completion % (last shell) Rmerge % (last shell) Space group Unit cell dimensions a (Å) b (Å) c (Å) B. Refinement Resolution (Å) Resolution (last shell Å) R-factor % (last shell) Free R-factor % (last shell) Reflections work/free No. atoms in model (mean B-factor Å2) Total Protein DNA Left arm Right arm Solvent r.m.s.d. from ideality Bond lengths (Å) Bond angles (deg.)
44.5–2.41 2.54–2.41 6.3 (3.9) 18,596 99.0 (94.5) 6.2 (41.6) P212121
45.0–2.72 2.87–2.72 7.7 (7.9) 12,530 97.1 (95.9) 6.5 (47.6) P212121
47.37 55.31 177.83
47.24 54.44 177.87
40.0–2.41 2.52–2.41 23.2 (33.3) 26.8 (34.7) 17,961/587
40.0–2.72 2.84–2.72 23.7 (35.1) 27.4 (39.7) 12,130/381
3318 (48.9) 1610 (48.2) 1426 (50.4) 470 (57.2) 426 (54.6) 282 (45.7)
3207 (52.5) 1610 (53.0) 1426 (52.8) 470 (64.5) 426 (56.7) 171 (44.8)
0.005 1.11
0.004 0.85
other architectural proteins or sequence-directed bends already identified important intasomal DNA architectural requirements. 13,20–24 We focus our discussion here first on excisive recombination, where scIHF2-K45αE functions as an accessory factor only when attR is negatively supercoiled, regardless of the presence or absence of divalent metal ions. The functionally important cognate site for IHF in attR is H2. It is located in close proximity to the two target sites for the second DNA-bending factor, Xis. Earlier reports identified the attR intasome as a more delicate structure than the attL intasome, and an inspection of recent modeling of the DNA path in the attR intasome clearly predicted a need for extreme DNA-bending at H2. 16,22–24 Our data strongly support this model and reveal that an open scIHF2-K45αE-H2 conformation on attR, with overall DNA bending angles significantly smaller than those found in complexes with scIHF2 or wildtype IHF, is not functional. In this case, DNA supercoiling is most likely required to overcome a thermodynamic barrier imposed on the formation of a stable closed scIHF2-K45αE–H2 complex, arising from the combined effects of the stiffness of short DNA segments and weakened interaction between the left DNA arm and the protein body of scIHF2K45αE.25 In the absence of divalent metal ions, but with the aid of supercoiling, the α45E side-chain could maintain an orientation very similar to that observed in the crystal structure, interacting with DNA phosphate groups through solvent bridges, as opposed to metal ions. Glutamate-phosphate water
737 bridges have been found for histone–DNA interactions in the nucleosome core particle.26 Another possibility for the observed dependency of scIHF2K45αE on supercoiling in attR is that its binding to H2 might be sensitive to the periodicity of the helical repeat, which, in turn, is affected by the topological state of DNA.
Figure 6. Divalent metal ion-induced conformational switch in scIHF2-K45αE–DNA complexes. (a) ScIHF2 and scIHF2-K45αE were incubated with radiolabeled H′ DNA in the absence or in the presence of various concentrations of MgCl2, as indicated. EMSA was performed in TB buffers containing the same concentration of MgCl2. The retardation factors (Rf = dc/df) were determined and normalized relative to scIHF2. At the bottom, Rf values obtained through complex formation at identical protein:DNA stoichiometries are indicated. (b) Effect of MgCl2 on intramolecular recombination. Supercoiled pλIR was incubated with Int and either scIHF2 or scIHF2-K45αE in the presence of increasing concentrations of MgCl2 that matched the range in (a), and digested with NcoI before agarose gel electrophoresis. (c) Quantification of recombination efficiencies. The amount of product and remaining substrate DNA shown in (b) was determined and plotted against the actual amount of magnesium ions present in each reaction.
738 The fact that IHF and scIHF2 activate recombination on topologically relaxed attR indicates that wild-type protein–DNA interactions are sufficiently strong in the absence of supercoiling. However, there appears to be no strict requirement for severe IHF-induced bending inside the attL intasome, where an open scIHF2-K45αE–H′ complex seems sufficient to configure a functional snup. Our results also provide insight into snup formation during integrative recombination. It is worth pointing out again that, even with wild-type IHF, negative supercoiling of attP is required to configure a functional intasome. Our finding that scIHF2K45αE is barely active in integrative recombination, but is significantly stimulated by divalent metal ions (Figure 6(b)), indicates that a closed scIHF2-K45αE– DNA complex with a high degree of DNA bending must be a prerequisite to permit necessary Int–DNA interactions within attP. Integrative and excisive recombination differ most notably in their requirement for the IHF cognate site H1 and, again, inspection of recent models for integrative intasomes indicates that a closed scIHF2-K45αE–H1 complex is most likely required there.16 In addition to the proposed effect on the degree of bending at H1 by scIHF2-K45αE, a more complex dependence involving cooperative protein binding effects, including the two other IHF binding sites, is possible. We think that scIHF2-K45αE is a valuable tool to investigate nucleoprotein complex formation in general. It could be used to probe the topography of specific snups involved in DNA transactions in vitro and in vivo, in both prokaryotic and eukaryotic cells. This is already exemplified by our earlier finding that scIHF3-K45αE, which contains one shorter peptide linker but behaves very much like scIHF2-K45αE in biochemical and functional assays, exhibits a defect in its function as an initiation factor for pSC101 replication in vivo.19 It is likely, therefore, that a functional pSC101 origin recognition complex must include a highly bent, closed IHF–DNA complex in vivo. An interesting future application could include the generation of E. coli strains that express either scIHF2 or scIHF2-K45αE instead of wild-type IHF, and perform comparative gene expression profiling in order to probe the topography of IHF-dependent snups involved in gene regulation. Engineering of metal-binding sites at the protein– DNA interface holds promise for the design of switches in other DNA-binding systems. As we have applied the scIHF2 variant in this study, DNA binding or conformational alteration that is dependent on divalent metal concentration can be exploited for probing structure–function relationships in other nucleoprotein assemblies. It would be informative to investigate the general effect of substituting basic amino acid residues that are involved in phosphate group contacts with acidic residues in other DNA-binding proteins. Our results suggest that such changes may give rise to a
Metal-mediated Protein-induced DNA Bending
requirement for a divalent metal-mediated protein–DNA interaction for stabilizing the nativelike, DNA-bound configuration. Such metal-dependent modulation could be explored for the design of nano-sized DNA switch devices.27
Materials and Methods Purification of scIHF2-K45αE Expression vector pETscIHF2-K45αE was derived from pETscIHF3-K45αE, and scIHF2 and scIHF2-K45αE were purified as described.19 Protein concentration was determined using BIO-RAD protein assay kits with IgG as a standard. Gel mobility-shift assays Binding affinities were determined using 32P -labeled oligonucleotides as described.19 Plasmid pWSRGFP contains attR flanked by a set of restriction sites.28 A portion (272 fmol) of supercoiled or EcoRI-cleaved DNA was incubated with scIHF2 and scIHF2-K45αE at 90-fold and 400-fold molar excess, respectively, in 10 mM Tris– HCl (pH 7.9), 50 mM NaCl, 10 mM MgCl2, 1 mM DTT, and 100 μg/ml of bovine serum albumin (BSA). After 30 min at room temperature, DNA was digested with HindIII and NotI. Samples were loaded onto a 4% (w/v) polyacrylamide gel, which was pre-run at 150 V for 1 h at room temperature. Reactions that contained MgCl2 were analyzed on 15% polyacrylamide gels in 0.5 × TBM buffer (45 mM Tris– borate (pH 8.3) plus MgCl2 at 0, 19 μM, 38 μM, or 76 μM). Gels were pre-run in TBM buffer at 210 V at 4 °C for 1 h. The electrophoresis buffer was constantly exchanged during the run. Images were quantified using a Bio-Rad Molecular Imager System. Recombination assays Reactions were performed essentially as described, using pλIR for integrative intramolecular recombination and pλER for excisive recombination.19,28 In 25 μl of buffer containing 84 fmol of att sites, a 90-fold and 400-fold molar excess of scIHF2 and scIHF2-K45αE, respectively, was used plus 1 pmol of Int. We used pWSRGFP and pCMVattL as DNA substrates for intermolecular excisive recombination.18,28 DNA was incubated with a 30, 90, and 400-fold excess of IHF, scIHF2 and scIHF2-K45αE, respectively, and 1 pmol of Int plus 3 pmol of Xis. Recombination efficiency was quantified after staining with ethidium bromide using a Bio-Rad Molecular Imager System. The effect of divalent cations was assayed in a buffer containing 44 mM Tris (pH 7.9), 60 mM KCl, 50 μg/ ml of BSA, 5 mM spermidine, with the indicated amounts of MgCl2. AFM imaging Binding reactions contained 0.246 nM attL and a 10, 30, and 100-fold excess of IHF, scIHF2, and scIHF2-K45αE, respectively, in 0.5 × TBE buffer, and were incubated at room temperature for 40 min. Droplets of 30 μl were spotted onto AP-mica or Glu-mica and incubated for
739
Metal-mediated Protein-induced DNA Bending
15 min. Samples were washed with deionized water and dried under pure nitrogen. Imaging was done with a Veeco Dimension 3000 AFM and Nanoscope IIIa controller (Digital Instruments, Santa Barbara, CA, U.S.A.). PPPNCH silicon tips (NANOSENSORS, Switzerland) were used in tapping-mode scanning (1.0 Hz). Images were analyzed with the Nanoscope software and the free software ImageJ§. AP-mica was prepared using the recently developed APTES solution method.29 Glu-mica was prepared by incubating AP-mica in aqueous glutaraldehyde solution (0.1%–5% v/v) for 10 min, and subsequently washed with deionized water and dried under pure nitrogen.30 Variation in the concentration of glutaraldehyde was used to optimize adherence of DNA to the mica surface. Crystal preparation and structure solution The scIHF–DNA complex was reconstituted by mixing purified scIHF protein and H′ DNA with 1:1 stoichiometry in a buffer containing 10 mM Hepes (pH 7.0), 0.02 M NaCl, 0.1 mM EDTA, and 8% (v/v) glycerol to give a final concentration of 14 mg/ml of the protein– DNA complex. H′ DNA was assembled by reannealing three HPLC-purified oligonucleotides (from SigmaGenosys, Germany): CGGTGCAACAAATTGATAAGCAATGCTTTTTTGGC GGCCAAAAAAGCATT GCTTATCAATTTGTTGCACC
Crystals of the scIHF2–DNA complex were grown via vapor diffusion at 18 °C in droplets made by mixing a 14 mg/ml complex solution with an equal volume of well solution containing 50 mM Tris (pH 7.5), 15% (w/v) PEG5000-MME, 10% glycerol, 50 mM NaCl, 30 mM MnCl2. For crystallization of scIHF2-K45αE–DNA complex, a modified well solution was used containing 50 mM Tris (pH 7.5), 25% PEG5000-MME, 20% glycerol, 50 mM NaCl, 7.5 mM MnCl2. Large crystals of the scIHF2–DNA complex and initial crystals of the scIHF2-K45αE–DNA complex were obtained by microseeding using scIHF2– DNA crystals. For stabilization and improvement of diffraction capability, crystals were transferred into 50 mM Tris (pH 7.5), 15% PEG5000-MME, 10% glycerol, 50 mM NaCl, 30 mM MnCl2, 10% (v/v) 2-methyl-2,4-pentanediol for scIHF2DNA, or 50 mM Tris (pH 7.5), 15% PEG5000-MME, 10% glycerol, 50 mM NaCl, 7.5 mM MnCl2, 15% (v/v) 2methyl-2,4-pentanediol for scIHF2-K45αE–DNA. Crystals were flash-cooled in liquid N2 and transferred into a N2 gas stream at −170 °C for collection of X-ray diffraction intensities on a Rigaku MicroMax 007 X-ray diffractometer equipped with a Rigaku Raxis IV++ image plate. Diffraction data were processed using MOSFLM and SCALA from the CCP4 suite.31,32 The scIHF2–DNA structure was solved by molecular replacement with routines from the CCP4 suite using the wild-type IHF–DNA model.5 Structural refinement was carried out with the programs O and CNS.33,34 In the final stages of refinement, specific B-form sugar pucker restraints present in the CNS DNA parameter file were removed.5 Molecular structure Figures were prepared with PyMOL (DeLano Scientific LLC, San Carlos, CA, USA∥).
§ http://rsb.info.nih.gov/ij/ ∥ http://pymol.sourceforge.net/
Protein Data Bank accession codes Atomic coordinates have been deposited in the RCSB Protein Data Bank with accession codes 2IIE (scIHF2DNA) and 2IIF (scIHF2-K45αE-DNA).
Acknowledgements Very special thanks go to A. Segall for purified Int and Xis proteins, and to D. Esposito for purified wild-type IHF. This work was funded through University Research grants provided by NTU (to P.D. and C.A.D.) and NUS (to J.Y.) and an ARC research grant from the Singapore Ministry of Education (to C.A.D. and P.D.).
References 1. Echols, H. (1986). Multiple DNA-protein interactions governing high-precision DNA transactions. Science, 233, 1050–1056. 2. Goosen, N. & van de Putte, P. (1995). The regulation of transcription initiation by integration host factor. Mol. Microbiol. 16, 1–7. 3. Nash, H. A. (1996). The HU and IHF proteins: accessory factors for complex protein-DNA assemblies. In Regulation of Gene Expression in E. coli (Lin, E. C. C. & Lynch, A. S., eds), R.G. Landes Company, Austin, Texas, USA. pp. 149–179. 4. Goodrich, J. A., Schwartz, M. L. & McClure, W. R. (1990). Searching for and predicting the activity of sites for DNA binding proteins: compilation and analysis of the binding sites for Escherichia coli integration host factor (IHF). Nucl. Acids Res. 18, 4993–5000. 5. Rice, P. A., Yang, S.-W., Mizuuchi, K. & Nash, H. A. (1996). Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn. Cell, 87, 1295–1306. 6. Lorenz, M., Hillisch, A., Goodman, S. D. & Diekmann, S. (1999). Global structure similarities of intact and nicked DNA complexed with IHF measured in solution by fluorescence resonance energy transfer. Nucl. Acids Res. 27, 4619–4625. 7. Seong, G. H., Kobatake, E., Miura, K., Nakazawa, A. & Aizawa, M. (2002). Direct atomic force microscopy visualization of integration host factor-induced DNA bending structure of the promoter regulatory region on the Pseudomonas TOL plasmid. Biochem. Biophys. Res. Commun. 291, 361–366. 8. Ellenberger, T. & Landy, A. (1997). A good turn for DNA: the structure of integration host factor bound to DNA. Structure, 5, 153–157. 9. Swinger, K. K. & Rice, P. A. (2004). IHF and HU: flexible architects of bent DNA. Curr. Opin. Struct. Biol. 14, 28–35. 10. Arfin, S. M., Long, A. D., Ito, E. T., Tolleri, L., Riehle, M. M., Paegle, E. S. & Hatfield, G. W. (2000). Global gene expression profiling in Escherichia coli K12. The effects of integration host factor. J. Biol. Chem. 275, 29672–29684. 11. Mangan, M. W., Lucchini, S., Danino, V., Croinin, T. O., Hinton, J. C. D. & Dorman, C. J. (2006). The integration host factor (IHF) integrates stationary-phase and virulence gene expression in Salmonella enterica serovar typhimurium. Mol. Microbiol. 59, 1831–1847.
740
Metal-mediated Protein-induced DNA Bending
12. Azaro, M. A. & Landy, A. (2002). λ Int and the λ Int family. In Mobile DNA II (Craig, N. L., Craigie, R., Gellert, M. & Lambowitz, A., eds), pp. 118–148, ASM Press, Washington, DC. 13. Richet, E., Abcarian, P. & Nash, H. A. (1986). The interaction of recombination proteins with supercoiled DNA: defining the role of supercoiling in lambda integrative recombination. Cell, 46, 1011–1021. 14. Richet, E., Abcarian, P. & Nash, H. A. (1988). Synapsis of attachment sites during lambda integrative recombination involves capture of a naked DNA by a protein-DNA complex. Cell, 52, 9–17. 15. Biswas, T., Aihara, A., Radman-Livaja, M., Filman, D., Landy, A. & Ellenberger, T. (2005). A structural basis for allosteric control of DNA recombination by lambda integrase. Nature, 345, 475–485. 16. Radman-Livaja, M., Biswas, T., Ellenberger, T., Landy, A. & Aihara, H. (2006). DNA arms do the legwork to ensure the directionality of lambda site-specific recombination. Curr. Opin. Struct. Biol. 16, 42–50. 17. Lorbach, E., Christ, N., Schwikardi, M. & Dröge, P. (2000). Site-specific recombination in human cells catalyzed by phage lambda integrase mutants. J. Mol. Biol. 296, 1175–1181. 18. Corona, T., Bao, Q., Christ, N., Schwartz, T., Li, J. & Dröge, P. (2003). Activation of site-specific DNA integration in human cells by a single chain integration host factor. Nucl. Acids Res. 31, 5140–5148. 19. Bao, Q., Christ, N. & Dröge, P. (2004). Single-chain integration host factors as probes for high-precision nucleoprotein complex formation. Gene, 343, 99–106. 20. Snyder, U. K., Thompson, J. F. & Landy, A. (1989). Phasing of protein-induced DNA bends in a recombination complex. Nature, 341, 255–257. 21. Goodman, S. D. & Nash, H. A. (1989). Functional replacement of a protein-induced bend in a DNA recombination site. Nature, 341, 251–254. 22. Goodman, S. D., Nicholson, S. C. & Nash, H. A. (1992). Deformation of DNA during site-specific recombination of bacteriophage lambda: replacement of IHF protein by HU protein or sequence-directed bends. Proc. Natl Acad. Sci. USA, 89, 11910–11914. 23. Goodman, S. D. & Kay, O. (1999). Replacement of
24.
25. 26.
27.
28.
29.
30.
31. 32. 33.
34.
integration host factor protein-induced DNA bending by flexible regions of DNA. J. Biol. Chem. 274, 37004–37011. Segall, A. M., Goodman, S. D. & Nash, H. A. (1994). Architectural elements in nucleoprotein complexes: interchangeability of specific and non-specific DNA binding proteins. EMBO J. 13, 4536–4548. Kahn, J. D., Yun, E. & Crothers, D. M. (1994). Detection of localized DNA flexibility. Nature, 368, 163–166. Davey, C. A., Sargent, D. F., Luger, K., Maeder, A. W. & Richmond, T. J. (2002). Solvent mediated interactions in the structure of the nucleosome core particle at 1.9 Å resolution. J. Mol. Biol. 319, 1097–1113. Shen, W., Bruist, M. F., Goodman, S. D. & Seeman, N. C. (2004). A protein-driven DNA device that measures the excess binding energy of proteins that distort DNA. Angew. Chem. Int. Ed. 43, 4750–4752. Christ, N., Corona, T. & Dröge, P. (2002). Site-specific recombination in eukaryotic cells mediated by mutant lambda integrases: implications for synaptic complex formation and the reactivity of episomal DNA segments. J. Mol. Biol. 319, 305–314. Liu, Z., Li, Z., Zhou, H., Wei, G., Song, Y. & Wang, L. (2005). Immobilization and condensation of DNA with 3-aminopropyltriethoxysilane studied by atomic force microscopy. J. Microsc. 218, 233–239. Wang, H., Bash, R., Yodh, J. G., Hager, G. L., Lohr, D. & Lindsay, S. M. (2002). Glutaraldehyde modified mica: a new surface for atomic force microscopy of chromatin. Biophys. J. 83, 3619–3625. Leslie, A. G. (1999). Integration of macromolecular diffraction data. Acta Crystallog. sect. D, 55, 1696–1702. Collaborative Computational Project Number 4 (1994). The CCP4 suite: programs for protein crystallography. Acta. Crystallog. sect. D, 50, 760–763. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Rystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921.
Edited by K. Morikawa (Received 3 August 2006; received in revised form 26 September 2006; accepted 27 September 2006) Available online 3 October 2006