Molecular Dissection of Mycobacterium tuberculosis Integration Host Factor Reveals Novel Insights into the Mode of DNA Binding and Nucleoid Compaction*

Background: mIHF belongs to a subfamily of proteins, distinct from E. coli IHF. Results: Functionally important amino acids of mIHF and the mechanism(s) underlying DNA binding, DNA bending, and site-specific recombination are distinct from that of E. coli IHFαβ. Conclusion: mIHF functions could contribute beyond nucleoid compaction. Significance: Because mIHF is essential for growth, the molecular mechanisms identified here can be exploited in drug screening efforts. The annotated whole-genome sequence of Mycobacterium tuberculosis revealed that Rv1388 (Mtihf) is likely to encode for a putative 20-kDa integration host factor (mIHF). However, very little is known about the functional properties of mIHF or the organization of the mycobacterial nucleoid. Molecular modeling of the mIHF three-dimensional structure, based on the cocrystal structure of Streptomyces coelicolor IHF duplex DNA, a bona fide relative of mIHF, revealed the presence of Arg-170, Arg-171, and Arg-173, which might be involved in DNA binding, and a conserved proline (Pro-150) in the tight turn. The phenotypic sensitivity of Escherichia coli ΔihfA and ΔihfB strains to UV and methyl methanesulfonate could be complemented with the wild-type Mtihf but not its alleles bearing mutations in the DNA-binding residues. Protein-DNA interaction assays revealed that wild-type mIHF, but not its DNA-binding variants, binds with high affinity to fragments containing attB and attP sites and curved DNA. Strikingly, the functionally important amino acid residues of mIHF and the mechanism(s) underlying its binding to DNA, DNA bending, and site-specific recombination are fundamentally different from that of E. coli IHFαβ. Furthermore, we reveal novel insights into IHF-mediated DNA compaction depending on the placement of its preferred binding sites; mIHF promotes DNA compaction into nucleoid-like or higher order filamentous structures. We therefore propose that mIHF is a distinct member of a subfamily of proteins that serve as essential cofactors in site-specific recombination and nucleoid organization and that these findings represent a significant advance in our understanding of the role(s) of nucleoid-associated proteins.

The annotated whole-genome sequence of Mycobacterium tuberculosis revealed that Rv1388 (Mtihf) is likely to encode for a putative 20-kDa integration host factor (mIHF). However, very little is known about the functional properties of mIHF or the organization of the mycobacterial nucleoid. Molecular modeling of the mIHF three-dimensional structure, based on the cocrystal structure of Streptomyces coelicolor IHF duplex DNA, a bona fide relative of mIHF, revealed the presence of Arg-170, Arg-171, and Arg-173, which might be involved in DNA binding, and a conserved proline (Pro-150) in the tight turn. The phenotypic sensitivity of Escherichia coli ⌬ihfA and ⌬ihfB strains to UV and methyl methanesulfonate could be complemented with the wild-type Mtihf but not its alleles bearing mutations in the DNA-binding residues. Protein-DNA interaction assays revealed that wild-type mIHF, but not its DNA-binding variants, binds with high affinity to fragments containing attB and attP sites and curved DNA. Strikingly, the functionally important amino acid residues of mIHF and the mechanism(s) underlying its binding to DNA, DNA bending, and site-specific recombination are fundamentally different from that of E. coli IHF␣␤. Furthermore, we reveal novel insights into IHF-mediated DNA compaction depending on the placement of its preferred binding sites; mIHF promotes DNA compaction into nucleoid-like or higher order filamentous structures. We therefore propose that mIHF is a distinct member of a subfamily of proteins that serve as essential cofactors in site-specific recombination and nucleoid organization and that these findings represent a significant advance in our understanding of the role(s) of nucleoid-associated proteins.
Bacterial nucleoid is a compressed helical structure composed of DNA, RNA, and small nucleoid-binding proteins, whose architecture and protein composition are regulated in a growth phase-dependent manner (1)(2)(3). Nucleoid-associated proteins play crucial roles in many processes such as DNA replication, recombination, repair, and gene expression through DNA bending, bridging, and wrapping (1)(2)(3). Escherichia coli nucleoid contains a diverse set of abundant proteins, collectively known as nucleoid-associated proteins (NAPs), 3 which fulfill both architectural and regulatory roles (2,3). In E. coli, the major NAPs include H-NS, HU, Fis, IHF, and StpA, which influence the topology of bound DNA by bending and bridging of nonadjacent DNA segments (3). Originally discovered in E. coli as an essential host factor for integration/excision of phage (4), IHF links the architecture of the genome to its function inside the cell, influencing replication (5) and transcription (6 -9), and serves as an integral component of several site-specific recombination systems (3,10). E. coli IHF, a member of the DNABII structural family, is composed of two subunits, IHF␣ and IHF␤ (ϳ10 kDa each), each of which is required for full IHF activity (3,8,10,11). E. coli IHF binds with high affinity to a 30 -35-bp DNA having a conserved 3Ј region with a consensus sequence WATCAANNNNTTR (where W is A or T, R is purine, and N is any base), and the 5Ј region is degenerate but is typically AT-rich (8,10,(12)(13)(14). Binding of E. coli IHF causes the DNA to adopt a U-turn, thus bringing the nonadjacent sequences into close proximity (12)(13)(14). However, other studies have shown that the interaction between IHF and DNA is complex, with IHF binding to DNA via different modes that induce different DNA-bending patterns, and these DNAbinding modes are sensitive to various solution conditions (15,16). * This work was supported by European Community Grant CSI_LTB LSHP-CT-Several molecular and genome-scale studies have demonstrated that the members of the IHF-HU superfamily of proteins regulate global and local gene expression in diverse species of Gram-negative and Gram-positive bacteria (17)(18)(19)(20)(21)(22). Mycobacterial IHF was originally discovered in Mycobacterium smegmatis as a factor essential for site-specific recombination promoted by mycobacteriophage L5 integrase (henceforth called phage L5) (23,24). Subsequently, annotation of the whole-genome sequence of M. tuberculosis H37Rv revealed the presence of a putative ihf gene in the pathogen (25). Several lines of evidence suggest that E. coli NAPs share relatively low amino acid identity with their counterparts in a wide variety of microorganisms, including M. tuberculosis (2,3,26,27). For example, M. tuberculosis H37Rv ihf (Mtihf) is predicted to encode a single 20-kDa polypeptide compared with two different protein species in E. coli (25). Consequently, general features of the nucleoid structure and function described for the E. coli paradigm may not be relevant to other bacteria, thus emphasizing the need to understand the identity and roles of NAPs, especially in pathogenic bacteria. Furthermore, unlike wild-type E. coli strains, IHF is essential for the growth and viability of M. smegmatis and Mycobacterium tuberculosis (24, 28 -30). However, despite these considerations, the functional properties of M. tuberculosis IHF (henceforth called mIHF), the mechanism underlying the formation of higher order nucleoprotein filaments and compaction of DNA into nucleoid, largely remains unknown.
In this study, we address two fundamental questions regarding the identity and function of mIHF. First, is mIHF essential for bacterial growth and DNA compaction? Second, because mIHF is structurally unrelated to E. coli IHF␣␤, what are the amino acid residues and the mechanism(s) involved in DNA binding and nucleoid compaction? Using multiple complementary methods, we show that mIHF alone was necessary and sufficient to restore genetic robustness in both E. coli ⌬ihfA and ⌬ihfB strains, induce DNA compaction, and stimulate site-specific recombination. Strikingly, our work disclosed that functionally relevant amino acid residues and the mechanism underlying mIHF binding to DNA and site-specific recombination are different from that of E. coli IHF␣␤. Overall, our data are consistent with the notion that mIHF is a distinct member of a subfamily of proteins that serve as essential cofactors in site-specific recombination and nucleoid organization and therefore could serve as a potential target for structure-aided drug design.

EXPERIMENTAL PROCEDURES
Homology Modeling and Sequence Alignment-The fulllength protein sequence of M. tuberculosis IHF was retrieved from Uniprot database (UniProtKB ID L0T6Q3) (31). The similarity of mIHF sequence with the experimentally determined protein structures in PDB (32) was analyzed through BLAST (33) (blast.ncbi.nlm.nih.gov). Multiple sequence alignments between different IHF proteins were obtained through T-Coffee webserver (34). The cocrystal structure of Streptomyces coelicolor IHF protein bound to dsDNA (sIHF; PDB code 4ITQ) was the topmost hit, sharing sequence identity of 60% (E-value 1e-20) with the C-terminal end of mIHF(87-190). Conse-quently, the sIHF protein, which lacks an N-terminal fragment of 86 amino acids, was chosen as the template for homology modeling using Modeler 9 version 10. The alignment between sIHF and mIHF was optimized using Promals3D (35). The secondary structures were restrained based on the boundaries derived from PSIPRED predictions (36). The refinement of the model was iteratively carried out using KoBaMin (37). The model with the lowest discrete optimized protein energy score was finally selected for analysis (38,39).
Western Blot Analysis-Polyclonal anti-mIHF antibodies against purified mIHF were prepared in rabbits and characterized according to the standard procedures (40). To investigate the identity of intracellular IHF, M. tuberculosis H37Rv cells were cultured in Middlebrook 7H9 medium (Difco) supplemented with 10% (v/v) albumin/dextrose/catalase enrichment and 0.05% (v/v) Tween 80 in a shaker incubator with a speed of 180 rpm at 37°C (41). Whole-cell lysates were prepared as described previously (42). Briefly, cells equivalent to 1.0 at A 600 at different time points of growth were resuspended in SDS loading buffer and disrupted by sonication using an ultrasonic liquid processor (MISONIX) two to three times for 30 -40 s in the pulse mode at 21% amplitude. The whole-cell lysates were treated with DNase I (1 g/ml) and RNase A (0.2 g/ml) prior to Western blot analysis. Equal amounts of protein from the lysates were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. The blot was blocked for 2 h at 24°C with 1% bovine serum albumin in phosphate-buffered saline. The blot was washed and probed in a solution containing 10 mM sodium phosphate (pH 7.4), 150 mM NaCl plus 0.1% bovine serum albumin, 0.1% ovalbumin, 0.1% Tween 20, and 0.02% sodium azide and anti-mIHF or anti-SigA antibodies for 12 h at 4°C. The blot was subsequently washed three times with PBS and incubated for 1 h at room temperature with alkaline phosphatase-or peroxidase-conjugated (secondary antibody Sigma). Finally, the blot was developed with the Fast 5-bromo-4-chloro-3-indolyl phosphate-nitro blue tetrazolium kit (Sigma) or enhanced chemiluminescence (ECL).
MMS and UV Sensitivity Assays-E. coli wild-type or ⌬ihfA or ⌬ihfB strains harboring pMtihf were grown in LB broth supplemented with 100 g/ml ampicillin. At A 600 nm ϭ 0.4, cells were then treated with 0.5% methylmethane sulfonate (MMS) for 45 min. Cells were collected by centrifugation and resuspended in an equal volume of a solution containing M9 salts. The indicated dilutions were made in M9 salts and were spotted on LB plates supplemented with 100 g/ml ampicillin. In parallel, plates containing serial dilutions of cells were UV-irradiated using an ultraviolet lamp (UVGL-58, 254 nm, G6T5 lamp). Plates were incubated at 37°C in the dark for 22 h.
Preparation of DNA Substrates-Plasmids pMH57 (attB) and pSS19 (attP) were linearized by restriction digestion with HindIII and BamHI, respectively (40). DNA fragments were labeled at the 5Ј-end using [␥-32 P]ATP and T4 polynucleotide kinase (New England Biolabs). Subsequently, labeled DNA was cleaved with EcoRI to release 600-and 546-bp fragments, respectively. The cleaved DNA fragments were electrophoresed through a 5% polyacrylamide gel in 45 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 10 V/cm for 8 h. The bands were excised from the gel and eluted into TE buffer (10 mM Tris-HCl (pH 7.5), 1 mM EDTA).
Curved DNA and 230-bp (pB16) and 220-bp noncurved DNA (pNB10) were excised from the pB16 and pNB10 plasmids (43), respectively, by digestion with HindIII and EcoRI. DNA fragments were labeled at the 5Ј-end using [␥-32 P]ATP and T4 polynucleotide kinase. The labeled DNA was electrophoresed through an 8% polyacrylamide gel in 45 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 10 V/cm for 6 h as described (40). The bands corresponding to 230 and 220 bp was excised from the gel and eluted into TE buffer (10 mM Tris-HCl (pH 7.5), 1 mM EDTA).
Isolation of M. tuberculosis ihf Gene-The coding sequence corresponding to the M. tuberculosis H37Rv ihf (Rv1388) was PCR-amplified from the cosmid DNA (MYTC21B4) using oligodinucleotides (forward primer, 5Ј-GAGGGCCATATGT-TAGGCAACACTATTCATG-3Ј and reverse primer, 5Ј-ATA-CATGGATCCTTAGGCGGAGCCGAA-3Ј) carrying the sites for NdeI and BamHI, respectively. PCR amplification yielded a product of the predicted length (573 bp). The PCR product was gel-purified and digested with restriction enzymes. By using terminal NdeI and BamHI restriction sites that were incorporated into the DNA during amplification, the fragment was directionally ligated into NdeI and BamHI sites of E. coli expression vector pET15b.The resultant plasmid was designated pMtihf. The identity of the recombinant plasmid was ascertained by restriction analysis and DNA sequencing. To confirm the identity of the protein encoded by Mtihf, we raised antibodies against purified mIHF and characterized them according to the standard procedures (40).
Construction of M. tuberculosis Mutant ihf Expression Plasmids-PCR primers used for site-directed mutagenesis are listed in Table 1. The plasmid pMtihf was mutated using the QuikChange method with PfuTurbo DNA polymerase and DpnI. The arginine at position 170, 171, and 173 was replaced with alanine or aspartate. Similarly, proline at position 150 was substituted with alanine. E. coli DH5␣ was used for plasmid amplification. The mutations were ascertained by restriction analysis and DNA sequencing.
Expression and Purification of M. tuberculosis IHF Wild-type and Mutant Proteins-A culture of E. coli Rosetta2 (DE3)pLysS strain harboring the plasmid pMtihf was grown Luria-Bertani broth containing 100 g/ml ampicillin and 34 g/ml chloramphenicol at 37°C until an A 600 of 0.5. mIHF was induced by the addition of isopropyl 1-thio-␤-D-galactopyranoside (IPTG) to a final concentration of 0.5 mM. The culture was incubated with gentle shaking at 37°C for 6 h. Cells were collected by centrifugation, washed in STE buffer (10 mM Tris-HCl (pH 8), 100 mM NaCl, and 1 mM EDTA), resuspended in buffer A (10 mM Tris-HCl (pH 8), 150 mM NaCl, and 10% (v/v) glycerol), and stored at Ϫ80°C. Cells were thawed and lysed by sonication (Model No. GEX-750, Ultrasonic Processor) on ice at 60% duty cycles in a pulse mode. The suspension was centrifuged in a Beckman Ti-45 rotor at 30,000 rpm for 1 h at 4°C. The supernatant was then loaded onto a Ni 2ϩ -NTA-agarose column equilibrated with buffer A. mIHF was eluted with a 30 -500 mm linear gradient of imidazole in buffer A. Peak fractions were pooled and dialyzed against buffer B (10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM DTT, 100 mM NaCl, and 10% (v/v) glycerol). The dialyzed sample was loaded onto a double-stranded DNA-cellulose column that had been previously equilibrated with buffer B. IHF was eluted with a 0.1-1 M linear gradient of NaCl in buffer B. Peak fractions were pooled and dialyzed against buffer C (10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM DTT, 150 mM NaCl, and 20% (v/v) glycerol). The purity of IHF was assessed by SDS-PAGE and found to be Ͼ98%. Aliquots of mIHF were stored at Ϫ80°C.
M. tuberculosis IHF Protein Variants, i.e. single (P150A) and triple point mutant proteins (R170A, R171A, R173A, and R170D, R171D, R173D) were expressed in E. coli Rosetta2 (DE3)pLysS strain. Cells were grown in LB broth containing 100 g/ml ampicillin and 34 g/ml chloramphenicol at 37°C until an A 600 of 0.5. Mutant proteins were induced by the addition of IPTG to a final concentration of 0.5 mM. The culture was incubated with gentle shaking at 37°C for 6 h. Cells were collected by centrifugation, washed in STE buffer (10 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 1 mM EDTA), resuspended in buffer A (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 10% (v/v) glycerol), and stored at Ϫ80°C. Cells were thawed and lysed by sonication (Model No. GEX-750, Ultrasonic Processor) on ice at 60% duty cycles in a pulse mode. The suspension was centrifuged in a Beckman Ti-45 rotor at 30,000 rpm for 1 h at 4°C. The supernatant was then loaded onto a Ni 2ϩ -NTA column equilibrated with buffer A. IHF was eluted with a 30 -500 mM linear gradient of imidazole in buffer A. Peak fractions were pooled and dialyzed against buffer B (10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM DTT, 150 mM NaCl, and 20% (v/v) glycerol). The purity of IHF mutant proteins was assessed by SDS-PAGE and found to be Ͼ98%. Aliquots of mIHF variants were stored at Ϫ80°C.
Expression and Purification of E. coli IHF␣␤-E. coli IHF␣␤ was purified as described previously (44). The plasmid pET21a IHF, harboring E. coli ihfA and ihfB, was purchased from Addgene (Cambridge, MA) and transformed into E. coli Rosetta2(DE3)pLysS strain. Bacteria were grown in LB broth containing ampicillin (100 g/ml) and chloramphenicol (34 g/ml) at 37°C until an A 600 of 0.6 and induced by the addition of 1 mM IPTG. The culture was incubated with gentle shaking at Microscopy-E. coli ⌬ihfA strain, harboring Mtihf or its variants, was grown in LB broth at 37°C to an A 600 ϭ 0.4. mIHF expression was induced with 0.8 mM IPTG for 2 h at 37°C. The cells from the IPTG-induced and uninduced culture were collected by centrifugation, washed twice with PBS, and then fixed with 70% ethanol. Cells were incubated for 15 min at room temperature and washed twice in PBS. Cells were spread evenly on a poly-L-lysine-coated glass slide and air-dried for 20 min, followed by treatment with DAPI solution (5 ng/l) for 2 min at room temperature in the dark. Slides were gently washed with PBS to remove excess DAPI solution, air-dried, and mounted in ProLong Gold Antifade reagent (Invitrogen). Cells were then visualized using an LSM-710 (Zeiss) confocal laser scanning microscope with a ϫ63 oil objective. Z-stacks were collected at intervals of 3-4 s . Three-dimensional reconstructions were done by weighted back projection of aligned images. Tomographic reconstructions thus obtained were processed as described (45).
Gel Mobility Shift Assay-Reaction mixtures (20 l) contained 40 mM Tris-HCl (pH 8), 100 mM KCl, 1 mM dithiothreitol (DTT), 5 mM potassium phosphate, 5% glycerol, 0.5 mM EDTA, with 2 nM 32 P-labeled double-stranded DNA (dsDNA) and the indicated amounts of mIHF or E. coli IHF␣␤ as specified in the figure legends. After incubation at 37°C for 20 min, 2 l of 10% loading dye was added to each sample. Samples were electrophoresed through polyacrylamide gels as follows: for mIHF, 6% gel, run at 12 V/cm for 6 h; for E. coli IHF␣␤ 4% gel run at 10 V/cm for 4 h; and for pB16 and pNB10 DNA in 0.5ϫ TAE (20 mM Tris acetate buffer (pH 7.4) containing 0.5 mM EDTA at 4°C. Gels were dried, and the bands were visualized and quantified as described previously (26).
In Vitro Recombination Assay-The assay was performed as described (46). Phage L5 integrase was purified as described previously (46). Reaction mixtures contained 20 mM Tris-HCl (pH 7.5), 1 mM DTT, 25 mM NaCl, 5 mM spermidine, 0.5 g of phage L5 integrase, 200 ng of negatively superhelical pMH39 DNA (contains attP site), 200 ng of 3.9-kb linear dsDNA containing attB site (generated by digestion of pMH57 with HindIII), and increasing concentrations (0.05 g to 2.5 g) of mIHF or Arg-Ala triple mutant protein. After incubation for 3 h at 24°C, the reaction was terminated by the addition of 1.2 l of a solution containing 5 mM EDTA, 0.1% SDS, and 0.4 mg/ml proteinase K. Samples were incubated for 5 min at 24°C and electrophoresed through a 0.8% agarose gel at 35 V for 10 h. Gels were stained with ethidium bromide, and the products were identified following visualization under UV light.
DNA Circularization Assay-The assay was performed as described previously (47). Reaction mixtures (20 l) contained 20 mM Tris-HCl (pH 8.0), 150 mM KCl, 1 mM DTT, 1 mM potassium phosphate, 5% glycerol, and 0.25 nM 32 P-labeled 140-bp duplex DNA (derived from digestion of pUC19 plasmid DNA with TfiI) and the indicated concentrations of either mIHF or E. coli IHF␣␤. After incubation at 37°C for 30 min, 20 units of T4 DNA ligase (Fermentas) and 1ϫ ligase buffer was added, and incubation was extended for 30 min. In reactions involving exonuclease III, samples were further incubated at 37°C for 20 min with 5 units of Exo III. Reaction was terminated by the addition of 1 l of 20% SDS and 1 l of 10 mg/ml proteinase K, followed by incubation at 37°C for 20 min. DNA was extracted with phenol/chloroform solution and precipitated with ethanol. DNA pellet, obtained after centrifugation, was resuspended in 5 l of 6ϫ DNA gel loading buffer. Samples were subjected to electrophoresis through 5% polyacrylamide gel in 45 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 10 V/cm for 5 h. The gels were dried, and the bands were visualized using a Fuji FLA-9000 phosphorimager.
Atomic Force Microscopy-AFM experiments were performed as described previously (47). DNA fragments were prepared by digestion of closed circular DNA (pB16 and pNB10) with ScaI or HindIII (34). Reaction mixtures contained 20 mM Tris-HCl (pH 7.5), 2 mM MgCl 2 , 150 mM KCl, 5 ng of the indicated DNA, and IHF at 50 nM (Fig. 6B) or 200 nM (Fig. 6C), respectively. After incubation for 30 min, 5-l aliquots were applied to the surface of freshly cleaved mica. Images were acquired using as SNL (silicon-tip on nitride lever) an AFM probe (Agilent Technologies, force constant 21-98 N/m) and Agilent AFM controller operated in tapping mode in air. Imaging was done at a resolution of 512 ϫ 512 pixels. Raw data were selected with the Picoimage software, and the same was used to "flatten" AFM images with second order polynomial fitting.

RESULTS
Sequence Alignments, Homology Modeling, and Identification of Candidate Residues for Mutagenesis-The annotated whole-genome sequence of M. tuberculosis H37Rv identified Rv1388 as the ihf gene. Multiple sequence alignment of the deduced amino acid sequence of Rv1388 revealed the presence of an additional 86 amino acids at the N terminus of mIHF but not in M. smegmatis IHF (Fig. 1A). A pairwise BLAST search of the PDB indicated that mIHF shows 60% sequence identity with S. coelicolor (sIHF) (48). Fig. 1B shows the sequence alignment between sIHF and mIHF derived from PROMALS3D (35). To generate mIHF homology model, the crystal structure of sIHF, determined at a resolution of 2.7 Å in complex with double-stranded DNA (PDB code 4ITQ), was used as a template (48), and the homology model was determined using MODELLER (34). Fig. 1C shows the mIHF structure in complex with duplex DNA colored based upon the conservation of amino acid residues derived from alignment with unique protein entries of UniProt using ConSurf (50). Superposition of the backbone traces of the homology-modeled mIHF structure with the sIHF cocrystal structure yielded a low root mean square deviation value of 0.27 Å, indicating a good match between the two proteins. The DNA-binding site was inferred from the superposition of the crystal structure of sIHF complexed with double-stranded DNA (dsDNA) onto the modeled structure of mIHF-dsDNA. These studies predicted an ensemble of residues in mIHF, Arg-170, Arg-171, and Arg-173, corresponding to Arg-85, Arg-86, and Arg-88 residues in sIHF (48) that might be involved in DNA binding activity. Comparison between the mIHF-dsDNA homology model and the crystal structure of E. coli IHF␣␤ with dsDNA revealed significant differences (Fig. 1D). The central structural element in E. coli IHF␣␤ is a pair of ␤-ribbon arms each with a critical Pro-65 residue that became inserted into the minor groove at the high affinity binding site. Additionally, IHF contacts dsDNA via the phosphodiester backbone, and the dsDNA wrapped around the protein and bent by Ͼ160° (8,14). In mIHF, Pro-150 is the corresponding conserved residue that is embedded in a tight turn (Fig. 1C). Because it is not in contact with DNA, it is pos-sible that it could be involved in maintaining the tertiary structure of mIHF.
Expression Patterns of M. tuberculosis IHF at Various Growth Phases-To assess whether the intracellular levels of mIHF alter during different stages of growth under in vitro conditions, M. tuberculosis H37Rv was grown in Middlebrook 7H9 liquid medium supplemented with 10% (v/v) albumin/dextrose/catalase enrichment and 0.05% (v/v) Tween 80 as described under "Experimental Procedures." Growth was monitored by measuring the absorbance of cells at 600 nm ( Fig. 2A). The growth transition from exponential to stationary phase began after ϳ140 h. To determine the levels of IHF, whole-cell lysates from cells harvested at different time points during the growth cycle were subjected to Western blot analysis using anti-mIHF-specific antibodies (Fig. 2B). In the early phases of exponential growth, the level of mIHF was low; however, its abundance steadily increased by about 4-fold when the cells reached a stationary phase of growth ( Fig. 2A). Our results are in agreement with previous data on the intracellular concentration of E. coli IHF␣␤, which varies from 12,000 molecules in the exponential growth phase to 55,000 molecules (ϳ4-fold) in the stationary phase (51). Western blot analysis of M. tuberculosis SigA, which served as a loading control (52), did not reveal any significant changes in its levels during the bacterial growth cycle. The N-terminal amino acid sequence of the first 10 amino acids of mIHF indicated that indeed it has an additional 86 amino acid residues (data not shown).  ). B, sequence alignment between sIHF and mIHF derived from PROMALS3D. The assigned secondary structures are displayed above the alignment. The coiled-coil structures (H1-H5) represent ␣-helical bundles and the lines ␤-sheet and -turns, respectively. C, homology modeled three-dimensional structure of mIHF double-stranded DNA, using the crystal structure of sIHF double-stranded DNA as a template (PDB code 4ITQ) (48). The mIHF DNA (depicted in a white schematic) binding residues have been inferred from the superposition of the cocrystal structure of sIHF-dsDNA onto the modeled structure of mIHF-dsDNA. The residues (Arg-171, Arg-170, and Arg-173) implicated in DNA binding are depicted as sticks. D, structure of E. coli IHF␣␤ complexed with dsDNA (PDB code 1IHF) (14). Both the structures were colored based upon the conservation of residues derived from Consurf database (38).

M. tuberculosis ihf Is Necessary and Sufficient to Maintain
Viability of E. coli ⌬ihfA and ⌬ihfB Strains against Genotoxic Stress-It is known that mutations in ihfA or ihfB in E. coli or Salmonella typhimurium generate a phenotype similar to that of the double mutant (18,20), and ihfA is essential for E. coli growth and efficient colonization (11,(53)(54)(55). However, exactly how IHF contributes to viability has not been defined. Investigating the functionality of ihf in M. smegmatis and M. tuberculosis mutants is difficult because it is essential for cell viability (24, 28 -30). To explore the connection between mIHF and cell viability, we investigated the ability of M. tuberculosis ihf to complement E. coli ⌬ihfA or ⌬ihfB strains for growth and against genotoxic stress. To do so, we used E. coli expression vectors in pET15b containing full-length M. tuberculosis ihf under the control of the T7 promoter. We also constructed similar plasmids in pET15b carrying Mtihf variants bearing triple point mutations in which Arg-170, Arg-171, and Arg-173 residues were replaced with Ala or Asp, predicted to be involved in DNA binding, and a plasmid-bearing point mutation in which Pro-150 was substituted with Ala (see Fig. 1C). Serial dilutions of the indicated wild-type, mutant, and complemented mutant strains, treated with either UV light or MMS, were spread on LB agar plates and incubated as described under "Experimental Procedures." Wild-type E. coli ihfA and ihf␤ strains (WT) showed slight or no sensitivity to the indicated doses of MMS or UV radiation (Fig. 3, A-D). However, ⌬ihfA or ⌬ihfB strains exhibited greater sensitivity at the indicated doses of UV radiation and MMS. In contrast, the survival capabilities of ⌬ihfA and ⌬ihfB strains harboring Mtihf was nearly comparable with those of the wild-type ihfA and ihfB strains, indicating protection against genotoxicity caused by UV and MMS (Fig. 3, A-D). Interestingly, the mIHF Arg-Ala triple mutant (R170A/R171A/R173A) was relatively more sensitive to MMS and UV radiation than Mtihf P150A in both ⌬ihfA and ⌬ihfB strains (Fig. 3, A-D). Similar results were also obtained with mIHF Arg-Asp (R170D/R171D/R173D) triple mutant against genotoxicity caused by UV and MMS in both ihfA and ihfB strains (data not shown). Immunoblot analyses indicated the presence of nearly equivalent amounts of wild-type and Arg-Ala triple mutant (R170A, R171A, R173A) mIHF proteins in ⌬ihfA and ⌬ihfB strains (Fig. 3, E and F). These results indicate that the lack of complementation in the case of Mtihf Arg-Ala triple mutant was not due to the absence of mIHF. Similar results were also seen with the expression of Mtihf Arg-Asp triple mutant and P150A proteins (data not shown). Furthermore, all the mutant proteins were stable in E. coli so that they could be purified in large amounts for biochemical and functional analyses (see below). Altogether, the data indicate that Mtihf alone is necessary and sufficient to restore viability to both ⌬ihfA and ⌬ihfB strains against genotoxic stress.
M. tuberculosis IHF Induces Compaction of DNA into Nucleoids-The previous results suggest that Mtihf helps E. coli to overcome the effects of genotoxic stress. We therefore sought to visualize the nucleoid morphologies of 4Ј,6-diamidino-2-phenylindole (DAPI)-stained E. coli cells bearing Mtihf under uninduced and IPTG-induced conditions using confocal laser scanning microscopy. Analyses of uninduced E. coli ⌬ihfA cells showed a light diffuse and even staining throughout the cytoplasm (98%, n ϭ 210) (Fig. 4a, upper panel).
In induced E. coli ⌬ihfA cells (Ͼ90%, n ϭ 150), we observed a single and highly condensed nucleoid structure in a small volume of the cell (Fig. 4a, lower panel). Furthermore, three-dimensional imaging of E. coli ⌬ihfA cells revealed highly compact nucleoid structures with a well defined shape in induced but not in uninduced cells (Fig. 4b), consistent with the idea that mIHF is a component of E. coli nucleoid and that it plays a role in the compaction of DNA into nucleoids. The expression of the Mtihf variant (R170A, R171A, and R173A) from the same plasmid and promoter in E. coli ⌬ihfA cells failed to induce nucleoid compaction (Ͼ94%, n ϭ 245) (Fig. 4c). Immunoblot analysis of whole-cell lysates from E. coli ⌬ihfA cells expressing Mtihf and its triple mutant showed that they were expressed at similar levels (Fig. 3E). These data clearly establish a direct role for wild-type Mtihf in nucleoid organization.
Ability to Interact with attB and attP Sites Is Conserved between M. tuberculosis and E. coli IHF-The observation that the site-specific mutants failed to complement the UV and MMS sensitivity phenotypes of E. coli ⌬ihfA and ⌬ihfB strains led us to explore the mechanistic basis for their defects. Scrutinizing the sequence identity between M. tuberculosis and E. coli IHF proteins seemed insufficient, a priori, to infer the DNA binding specificity of mIHF. Accordingly, we cloned, expressed, and purified wild-type and variant forms of mIHF and E. coli IHF␣␤ (Fig. 5). To investigate the DNA binding properties of mIHF, we chose dsDNA containing phage L5 attB (600 bp) or attP (546 bp) sites (24,56). Binding reactions were performed with 2 nM 32 P-labeled dsDNA having attP or attB site and increasing concentrations of mIHF. Reaction mixtures were subjected to electrophoretic mobility shift assay (EMSA) as described under "Experimental Procedures." The gel shift assays indicated that mIHF altered the mobility of dsDNA. As shown in Fig. 6, A and B, mIHF bound to attB and attP containing dsDNA to form stable complexes. However, mIHF at lower concentrations formed poorly resolved complexes, thus generating a smear (Fig. 6, A and B, lanes 2-5). With increasing mIHF concentrations, the slower moving band gradually increased in intensity resulting in the formation of a well defined band, indicating that saturation had been achieved (Fig. 6, A and B, lanes  6 -10). However, mIHF bound more efficiently to the attB site containing substrate than to attP site containing substrate. We next tested the effect of mutations in amino acid residues of mIHF that are predicted to make direct contacts with DNA. We observed that substitution of Ala or Asp for Arg-170, Arg-171, and Arg-173 as well as Ala for Pro-150 completely abolished mIHF binding to attB-and attP-containing dsDNA substrates as assessed by EMSA (Fig. 6, C-H). Quantitative analysis of the extent of complex formation between mIHF and various DNA substrates, as a function increasing concentration of mIHF, suggested that mIHF shows ϳ2-fold higher preference for attB containing DNA compared with the attP DNA substrate (Fig. 7I).
In parallel experiments, we investigated the binding affinity of E. coli IHF␣␤ to phage L5 attB and attP containing dsDNA. In contrast to mIHF, E. coli IHF␣␤ bound attB DNA at levels equivalent to that of attP DNA but to a lesser extent to curved and noncurved DNA (Fig. 8, A-D). Furthermore, quantitative analysis of the formation of protein-DNA complexes as a func-tion increasing the concentration of E. coli IHF␣␤ corroborates with the idea that it binds more efficiently to attB-and attPcontaining substrates compared with curved and noncurved DNA (Fig. 8E).
M. tuberculosis IHF Binds with Greater Affinity to Curved DNA, attP and attB than Noncurved DNA-A striking feature of E. coli IHF␣␤ is its ability to bind phage attB and attP sites and induce bends to facilitate the formation of higher order structures (8,10,14). M. smegmatis IHF appears to promote the integration of phage L5 in a fashion similar to that of E. coli IHF␣␤ (24,28). To gain further insights into mIHF DNA binding activity, we chose to determine the binding affinity of mIHF to 230-bp curved DNA (derived from plasmid pB16) and 220-bp noncurved dsDNA (derived from plasmid pNB10) (43). The curved DNA is characterized by a high (78%) A ϩ T content (43). Binding reactions were performed as described above. We found that mIHF bound with higher affinity to curved dsDNA over noncurved dsDNA substrate (compare Fig. 7, A with E). Furthermore, relatively lower protein concentration of mIHF was required to form complexes with curved DNA than with noncurved DNA. Quantitative analysis indicated that the affinity of mIHF for curved DNA is 2-fold higher compared with noncurved DNA (Fig. 7I and Table 2). Consistent with the data obtained for attB and attP containing DNA substrates, mIHF mutant proteins failed to bind both curved and noncurved DNA even at the highest protein concentrations tested (Fig. 7, B-D and F-H). Likewise, E. coli IHF␣␤ displayed rela- tively higher binding affinity for curved DNA compared with noncurved DNA (Fig. 8E).
To ascertain the binding specificity of mIHF to attP and attB sites, we performed competition experiments using increasing concentrations of noncurved DNA. Fig. 9 shows that unlabeled DNA containing attP and attB sites but not noncurved DNA competed with the formation of 32 P-labeled DNA-mIHF complexes. To further characterize the binding affinity, we have determined the apparent equilibrium dissociation constant (K d ) of mIHF binding to various DNA substrates in comparison with E. coli IHF␣␤. The resulting K d values are summarized in Table 2. We found that the K d values for attB-and attP-containing fragments to be in the range of 0.3 to 0.5 M. These values are similar to those reported for E. coli IHF␣␤ (57). Although the difference in the affinity between specific and nonspecific DNA is smaller, the weaker binding of IHF proteins to nonspecific DNA is likely to be relevant due to the fact that IHF is present at 10 -100 M in the cell (this study and Ref. 51).
Binding of mIHF, but Not E. coli IHF␣␤, to Phage L5 attP and attB DNA Produces a Salt-stable Complex-Elucidation of the thermodynamic parameters of protein-DNA interactions is crucial to the understanding of factors that dictate the function of protein-DNA complexes. The binding affinity and specificity of E. coli IHF␣␤ depend on the salt concentration (58,59). At low salt concentrations below 100 mM, the specificity of IHF is low, and the specificity increases as the salt concentration increases, up to 250 mM. We reasoned that subtle differences between mIHF and E. coli IHF␣␤ in their affinity to dsDNA substrates were not fully pronounced in the above assay. We therefore investigated the stability of protein-DNA complexes formed by mIHF and E. coli IHF␣␤ in the presence of increasing salt concentrations. Interestingly, even 1.5 M NaCl was not sufficient to dissociate the complexes formed with attP or attB containing DNA by mIHF (Fig. 10). However, complete dissociation of complexes formed by E. coli IHF␣␤ with the same substrates occurred in the presence of 150 mM NaCl (Fig. 10). Next, we examined the stability of complexes formed by mIHF or E. coli IHF␣␤ with curved and noncurved DNA. The complexes formed with curved DNA remained stable, even in the presence of 1.5 M NaCl (Fig. 10). In contrast, the stability of complexes formed by mIHF or E. coli IHF␣␤ with noncurved DNA drastically declined in the presence of 200 to 300 mM NaCl. The results are summarized in a graphical form in Fig.  10I. We observed striking differences in the effect of NaCl on the stability of mIHF-attB, mIHF-attP, and mIHF-curved DNA compared with complexes formed with noncurved DNA. Among the potential explanations for this finding, one possibility is that mIHF binding to attP and attB results in encircling of dsDNA, via protein-protein interactions, hence the establishment of a salt-stable complex. This possibility was supported by the observation that binding of mIHf to curved DNA results in the formation of nucleoid-like structures (see below).
M. tuberculosis IHF Mutants Lacking DNA Binding Activity as Well as E. coli IHF Fail to Stimulate Site-specific Recombination Promoted by Phage L5 Integrase-One of the hallmarks of IHF family of proteins is their ability to serve as essential cofactors in the integration of phage genome into or excision from the host chromosome. Previous studies have shown that M. smegmatis IHF is essential for recombination promoted by phage L5 integrase (23,28). To further suggest biological significance, we sought to examine the ability of mIHF and its variants to stimulate site-specific recombination as described previously (28,56). In this assay, recombination between phage L5 attP and attB sites generates a new DNA species with molecular weight of ϳ8 kb. As shown in Fig. 11A, recombination promoted by phage L5 integrase was stimulated by wild-type mIHF in a concentration-dependent manner. In contrast to wild-type mIHF, Arg to Ala triple mIHF mutants failed to stimulate recombination, even at higher concentrations (Fig. 11B). Similar results were obtained in the case of Arg to Asp triple mIHF mutant protein (data not shown). As shown above (Fig. 3), although P150A mIHF mutant partially complemented the UV and MMS sensitivity of E. coli ⌬ihfA and ⌬ihfB strains (Fig. 3), the mutant protein failed to form a stable complex with DNA (Fig. 6). We reasoned that the P150A mIHF-DNA complex might not be sufficiently stable to persist during EMSA. Given this scenario, we examined the ability of the P150A mIHF variant to stimulate integrase-promoted recombination. Interestingly, the P150A mutant protein was active and stimulated site-specific recombination, albeit at somewhat higher concentrations than the wild-type mIHF (Fig. 11C). The product formed by P150A was identical to that formed by wild-type mIHF. Restriction analysis of the recombination products ascertained that the new species of DNA is the product of sitespecific recombination between attP and attB sites (data not shown). Next, we asked whether a heterologous IHF could stimulate site-specific recombination promoted by phage L5 integrase. As shown in Fig. 11D, E. coli IHF␣␤ failed to stimulate the reaction. The failure is perhaps due to the inability of E. coli IHF␣␤ to interact with phage L5 integrase to form a functional intasome (60). The combined data demonstrate that mIHF is structurally (at the primary and three-dimensional level) and functionally distinct from E. coli IHF␣␤.
M. tuberculosis IHF Promotes Bending of Duplex DNA-It is known that DNA architectural proteins, such as E. coli IHF␣␤, have the capacity to drastically modify DNA structure by bending (3). To investigate whether mIHF has the ability to promote DNA bending, we used T4 DNA ligase-catalyzed circularization of 140-bp DNA fragment excised from pUC19 plasmid DNA, which was devoid of intrinsically bent sequence. The principle underlying this assay is based on the fact that bending of DNA would bring the two ends into close proximity to facilitate ligation. In this assay, we used E. coli IHF␣␤ as a positive control. We incubated the 32 P-labeled DNA fragment first with increasing concentrations of E. coli IHF␣␤ or mIHF followed by T4 DNA ligase. Subsequently, reaction mixtures were treated with exonuclease III to remove linear DNA molecules. The DNA products in the reaction mixture were separated by PAGE and visualized by phosphorimaging of the dried gels. In the presence of E. coli IHF␣␤, and in agreement with previous studies (3,10), we observed the formation covalently closed circular DNA, which was resistant to exonuclease III digestion (Fig.  12A). This species was absent where ligation was performed in the absence of E. coli IHF␣␤ (Fig. 12A, lane 2). We next examined the ability of mIHF to generate covalently closed DNA molecules. We found that mIHF, over the same range of concentrations, not only exhibited the ability to generate covalently closed DNA molecules but also had the capacity to bridge DNA molecules into linear multimers, albeit to a much lesser extent (Fig. 12B, lanes 4 -8). We considered that the formation of a linear multimer seen with mIHF posits a role in the overall organization and compactness of the nucleoid.
Different Modes of mIHF Binding to DNA-A number of studies have shown that E. coli IHF␣␤ binds with high affinity to 30 -35 bp having a conserved 3Ј region with a consensus sequence WATCAANNNNTTR (where W is A or T, R is purine, and N is any base), and the 5Ј region is degenerate but is typically AT-rich (8,10,(12)(13)(14). To gain further insights into the structural features of mIHF-DNA complexes, we generated DNA fragments that had A ϩ T tracts embedded either in the center or near the end (Fig. 13A). Reactions were performed under conditions similar to those used for mobility shift assays. The products of the reactions were visualized using AFM. In the absence of mIHF, dsDNA fragments were devoid of bent structures and tangles (Fig. 13B, panels i and v). Reactions per-formed with DNA fragments containing A ϩ T tract at the center and limiting amounts of mIHF, we observed that mIHF binding was predominantly at the center so that the two arms extended to either side of the complex (Ͼ90%, n ϭ 120) (Fig.  13B, panels ii-iv). Furthermore, this mode of binding resulted in sharp bends around the protein core, and the estimated contour lengths were in agreement with mIHF binding at the center (Fig. 13B, panels ii-iv). Under identical conditions, with a DNA fragment containing an A ϩ T tract near the end, binding was primarily at the ends with no bending (Fig. 13B, panels vi-viii). Interestingly, we observed globular structures at the ends, indicating the formation of higher order nucleoprotein structures. A further increase in the amount of mIHF led to the formation of two distinct types of structures, one with the DNA fragment having A ϩ T tracts near the end; mIHF first bound to the end and then to the end lacking the A ϩ T-tract (n ϭ 160) (Fig. 13C, panels i and ii). However, for DNA fragment that had A tracts embedded at the center, mIHF binding resulted in the formation of nucleoid-like structures in which DNA is fully wrapped around the protein core (Fig. 13C, panels iii-iv). The two modes of mIHF binding have important implications for the compaction of DNA into nucleoids and the formation of higher order nucleoprotein structures.

DISCUSSION
A great deal of structural and biochemical work has been devoted to understanding the functional properties of E. coli IHF␣␤. Very little is known about the components of nucleoid, and their function in mycobacteria, but such knowledge is crucial for understanding of the changes that M. tuberculosis nucleoid experiences in response to the frequently altering milieu of the host. Multiple sequence alignment and homology modeling of the mIHF three-dimensional structure indicated that it is unrelated in sequence and structure to the prototype E. coli IHF␣␤, but it is similar to that of sIHF (Fig. 1). Notwithstanding the structural differences, mIHF alone was necessary and sufficient for the restoration of viability in both E. coli ⌬ihfA and ⌬ihfB strains against genotoxic stress, and it could induce DNA compaction and catalyze site-specific recombination. The discovery that functionally relevant amino acid residues and the mechanism that governs mIHF binding to DNA as well as DNA bending are different from that of both E. coli IHF␣␤ and sIHF, and provides protection from genotoxic stress, is striking. Overall, our data are consistent with the notion that mIHF is a distinct member of the family of proteins that serve as essential cofactors in site-specific recombination   and nucleoid organization. Furthermore, we believe that our results represent an important contribution and provide insights into those functions of mIHF that are likely to be essential for growth and viability of M. tuberculosis. Studies of different bacterial systems have revealed the existence of complexities in the number and intracellular concen-trations of NAPs (1)(2)(3). In many bacteria, the intracellular levels of IHF are dependent on the growth conditions and influence the patterns of gene expression in a temporal fashion (2,3). Very little is known on the expression patterns and biological significance of NAPs in mycobacteria. Here, we show that the expression of mIHF follows a dynamic pattern as a function of the growth cycle (Fig. 2). Like E. coli IHF␣␤ (3), expression of mIHF manifests during the early exponential phase, increases during the mid-exponential phase, and then reaches a plateau in stationary phase. The growth phenotypes of E. coli mutants lacking both IHF and HU are more severe than the phenotypes of single mutants (53)(54)(55). However, it remains possible that IHF mutants may display mild growth attenuation that is not severe enough to be identified in the experiments described in the literature. In contrast to E. coli, null mutations of ihf in both M. tuberculosis and M. smegmatis are lethal (24, 28 -30). Therefore, we used E. coli ⌬ihfA and ⌬ihfB strains to explore the biological roles of mIHF under certain stringent environmental conditions. Strikingly, we found that Mtihf alone could complement the growth and MMS and UV sensitivity phenotypes of E. coli ⌬ihfA and ⌬ihfB strains as well as induce DNA compaction, raising the possibility that it may function as a barrier in the defense against genotoxic stress and/or in the regulation of multiple DNA repair pathways.
Our comprehensive biochemical analyses suggest that wildtype mIHF, in contrast to its mutant variants, binds attB and attP sites, forms stable nucleoprotein complexes, and stimulates site-specific recombination. Furthermore, by using mutant proteins, we assessed the impact of the amino acid residues identified from our modeling efforts on the biological and biochemical properties of mIHF. We found that the alleles affecting DNA binding failed to complement E. coli ⌬ihfA and ⌬ihfB strains, and the variant proteins lacked detectable DNA binding activity. Molecular modeling studies showed that the Pro-150 is not in contact with DNA and therefore is unlikely to be essential for mIHF DNA binding activity. Intriguingly, we found that the P150A mutant protein failed to display significant DNA binding activity as assessed by EMSA. We speculated that the inability of Pro-150 mutant to bind DNA could be due to the instability of DNA-protein complex during gel mobility assays. This premise is supported by the observation that the Pro-150 mutant protein was able to catalyze site-specific recombination under solution conditions. The similarity in sequence and structure of mIHF and sIHF raises the question of how these two closely related proteins perform their functions with requirements for binding specificity. Although mIHF can bind nonspecific sequences, it forms especially "stable" high affinity complexes with cognate DNA that are not disrupted by the addition of high salt and competitor DNA. However, sIHF interacts nonspecifically with double-stranded DNA with K d in the micromolar range (48). The ability to interact with cognate DNA as well as nonspecific sequences at varying affinities is a common characteristic of the IHF-HU DNABII superfamily of proteins (3,8,10). In the co-crystal structure of sIHF with 19-bp DNA, the interface between sIHF and DNA is less extensive and showed no significant DNA bending (48). Although the amino acid residues of sIHF involved in DNA binding have not been identified, it is devoid of the conserved proline residue present   in E. coli IHF that becomes inserted into the minor grove and introduces DNA bends (8). However, the lack of DNA bending in the cocrystal structure of sIHF-DNA may not be due to lack of proline, possibly due to short stretch of duplex DNA.
Our studies reveal that mIHF causes wrapping and bending of short DNA fragments, suggesting a probable mechanism underlying the formation of nucleoid-like structures. By analyzing both target and nontarget DNA substrates, we observed that mIHF displays relatively higher affinity for phage L5 attP and attB sites and DNA rich in A/T sequences. These findings are in agreement with earlier studies for E. coli IHF␣␤ (10,12,13,61,62). Comparison of over 170 known E. coli IHF␣␤-binding sites has led to the identification of a consensus DNA-binding motif (10,62,63). Among these, the two most highly conserved elements are the sequence WATCAA starting near the center of the site and the second sequence TTR located 4 bp in the 3Ј direction from WATCAA (62)(63)(64). Other high affinity binding sites of E. coli IHF␣␤ contain poly(A)-tract containing 4 -6 adenines (65,66). Although mIHF exhibits relatively high affinity for phage L5 attP and attB sites, further studies are necessary to identify the consensus binding motif(s) in these sites. Since its discovery (4), E. coli IHF␣␤ has attracted considerable attention because of its role in various DNA transactions, wherein IHF-induced sharp bends or DNA loops coordinate the cooperative assembly of multicomponent nucleoprotein complexes (2,3,8). The biochemical data, combined with the crystal structures, have shown that E. coli IHF␣␤ binds to the minor groove of DNA and bends the double helix by Ͼ160° (14,67,68). Other studies have shown that E. coli IHF ␣␤ binding to DNA, and subsequently DNA bending, occurs in a concerted fashion (67)(68)(69).
Nucleoid-associated proteins are bifunctional molecular entities. First, the ensemble and single molecule measurements have shown that NAPs contribute to DNA architecture in the organization of bacterial chromatin by folding and packaging of DNA into nucleoids. Second, NAPs play key roles in the regulation of many important genes by interaction with specific sequences in the target gene promoters. One of the traditional views in regard to roles of NAPs is that, due to their high abundance and promiscuity with respect to DNA binding, they play key roles in the regulation of transcription by modulating the state of DNA structure (1-3). An array of mechanisms has been described by which NAPs mediate compaction of DNA into nucleoids and regulate gene expression. As the affinity of NAPs to DNA is generally weak, cooperative binding is therefore essential to increase the local concentrations of NAPs near the target sites. Other mechanisms include introduction and constraining of DNA supercoiling and bridging of adjacent segments of DNA. Collectively, these processes are thought to influence the gene order along the genome and likely to regulate the temporal order of gene expression (1)(2)(3). Despite these advances, our knowledge with respect to the factors and the molecular mechanism that determines the choice between the architectural and regulatory roles of NAPs remains poorly understood. Our study provides insights into the determinants involved in the formation of filamentous and nucleoid-like structures. Altogether, our findings are consistent with the notion that the dual roles by IHF may be determined by its binding position relative to the genes it controls.
While considering the capacity of mIHF to engage in genome organization and as a global transcriptional silencer, it is useful to understand its physical impact on the DNA to which it binds. Our study suggests that the sequence determinants and their context might influence the choice between the above-mentioned processes. We also note that this is a fundamental mechanism pertinent to nucleoid organization and gene expression, which may be generally applicable to all the IHF-regulated genes and binding sites in bacterial cells (70). Given the fact that The scale bars are as follows: 100 nm (B, panels i and ii) and 300 nm (B, panels ii-iv and panels vi-viii). C, higher concentrations of mIHF (200 nM) promote the formation of higher order filamentous and nucleoid-like structures. Panels i and ii, randomly selected images showing the formation of filamentous structures; panels iii and iv, randomly selected images depicting the wrapping of DNA into nucleoid-like structures. The closed bars on the left-hand side indicate DNA fragments with the IHF-binding site (red box) at the end or center. White arrow denotes higher order filamentous structures; green and blue arrows denote mIHF binding at the center and nucleoid-like structures, respectively. Scale bars are as follows: 300 nm (panels i-iv) or and 400 nm (panel ii).
sIHF and mIHF are structurally related, interestingly, sIHF is not essential for growth but is required for normal chromosome compaction (48). However, because IHF is essential for growth of M. tuberculosis, it is likely to be involved in multiple essential biological functions as well as in the maintenance of genome integrity; as such, this can be exploited in drug screening efforts.