|Home | About | Journals | Submit | Contact Us | Français|
The MogR transcriptional repressor of the intracellular pathogen Listeria monocytogenes recognizes AT-rich binding sites in promoters of flagellar genes to down-regulate flagellar gene expression during infection. We describe here the 1.8Å resolution crystal structure of MogR bound to the recognition sequence 5′ ATTTTTTAAAAAAAT 3′ present within the flaA promoter region. Our structure shows that MogR binds as a dimer. Each half-site is recognized in the major groove by a helix-turn-helix motif and in the minor groove by a loop from the symmetry related molecule, resulting in a ‘cross-over’ binding mode. This oversampling through minor groove interactions is important for specificity. The MogR binding site has structural features of A-tract DNA and is bent by ~52° away from the dimer. The structure explains how MogR achieves binding specificity in the AT-rich genome of L. monocytogenes and explains the evolutionary conservation of A-tract sequence elements within promoter regions of MogR-regulated flagellar genes.
Flagella are surface structures that are required for bacterial motility, adhesion to host cells, invasion and virulence (Macnab, 2004; Ramos et al., 2004; Van Houdt and Michiels, 2005). Given their key role in bacterial infections, flagella are also prone to being recognized by host pattern recognition receptors such as Toll-like receptor 5 (TLR5) (Andersen-Nissen et al., 2005; Feuillet et al., 2006; Hayashi et al., 2001). To evade detection by the host immune response, bacterial pathogens frequently down-regulate production of flagella shortly after colonization (Andersen-Nissen et al., 2005; Ramos et al., 2004). In the Gram-positive, facultative intracellular pathogen Listeria monocytogenes, this down-regulation occurs primarily at mammalian physiological temperature (37°C and above) and during replication within the eukaryotic cell cytosol (Gründling et al., 2004). Repression of flagellar gene expression is mediated by the motility gene repressor, MogR (Gründling et al., 2004). Binding of MogR to operator sequences located within flagellar gene promoter regions directly represses transcription of all flagellar motility genes in a non-hierarchal manner (Shen and Higgins, 2006). In the absence of MogR, all flagellar motility genes are constitutively expressed at high levels in a temperature-independent manner. The resulting over-expression of flagellin (FlaA) by MogR-negative bacteria induces a chaining phenotype. This abnormal cellular physiology compromises the ability of L. monocytogenes to invade host cells (Shen and Higgins, 2006). Thus, MogR is a key mediator of the transition of L. monocytogenes from a ubiquitous extracellular bacterium to an intracellular pathogen.
MogR binding sites were originally identified by examining MogR-regulated flagellar gene promoter regions for conserved sequence motifs (Shen and Higgins, 2006). This analysis revealed that the palindromic sequence 5′ TTTTNNNNNAAAA 3′ occurred repeatedly in flagellar gene promoter regions. The intervening nucleotides (N) typically contained A/T base pairs. For example, the best-characterized MogR binding site within the flaA promoter region contains the sequence 5′ TTTTTTAAAAAAA 3′. The sequence requirements of this binding site were confirmed using gel mobility shift analysis with wild-type and mutant flaA promoter region sequences (Shen and Higgins, 2006). As predicted from the sequence conservation, the flanking thymines and adenines (underlined) were essential for MogR binding, and the intervening sequences were more permissive to nucleotide changes.
The 5′ TTTTNNNNNAAAA 3′ sequence occurs ~5000 times (one mismatch allowed; ~500 times with no mismatch allowed) within the upstream regions of genes in the AT-rich genome of L. monocytogenes (60% AT) (Shen and Higgins, 2006). Although previous microarray analyses revealed that the primary targets of MogR repression are flagellar gene promoters, how MogR discriminates its recognition sequences within flagellar promoter regions over the multitude of possible genomic binding sites remained unknown. Promoter regions of MogR-repressed flagellar genes frequently contain multiple MogR binding sites spaced 1, 2 or 3 helical turns apart, and previous analyses have shown that a minimum of two MogR binding sequences are required to repress flaA transcription in L. monocytogenes (Shen and Higgins, 2006). Therefore, it has been proposed that MogR requires a minimum of two 5′ TTTTNNNNNAAAA 3′ sequences separated by integral helical turns to function as a transcriptional repressor (Shen and Higgins, 2006).
In this report, we investigated the molecular mechanism by which MogR recognizes its A-tract DNA binding sequence and thus functions as a master regulator of flagellar gene expression. A-tract binding sites are of particular interest because they are known to exhibit intrinsic DNA curvature, which can play significant roles in transcriptional regulation by affecting promoter geometry (Crothers and Shakked, 1999). First, we determined that MogR contains two functional domains: an N-terminal DNA binding domain and a putative C-terminal leucine zipper oligomerization domain. In the absence of the C-terminal leucine zipper domain, MogR was monomeric and able to bind to its recognition site in vitro, albeit with reduced binding affinity as compared to the full-length protein. We then determined the 1.8Å resolution crystal structure of the N-terminal DNA binding domain of MogR bound to the flaA promoter region binding site 5′ ATTTTTTAAAAAAAT 3′. In the crystal structure, MogR is bound to its palindromic DNA target sequence as a dimer. The structure reveals that the whole 15 bp recognition element, except the central A:T base pair, is recognized and forms a composite MogR dimer binding site. These analyses provide insight into the mechanism of how an A-tract binding protein specifically recognizes its DNA binding sequence in the context of an AT-rich genome.
MogR was originally identified by virtue of its ability to bind to flaA promoter region DNA (Gründling et al., 2004). MogR possesses no apparent homology to any previously characterized protein, and sequence analysis did not reveal a DNA binding domain within MogR. The only apparent feature is a putative leucine zipper domain in the C-terminus (amino acids 218–263, Figure 1A). To further identify functional domains within MogR, we designed C-terminal truncation constructs and assessed their ability to bind to flaA promoter region DNA using gel mobility shift assays and DNAse I footprinting. These experiments showed that a truncated MogR protein containing residues 1–162 allowed DNA binding and that the C-terminal leucine zipper domain is important for overall binding affinity (Figure 2 and Supplemental Figure S1). The oligomerization status of these constructs was further investigated using size exclusion chromatography. We observed that MogR constructs containing the C-terminal leucine zipper region had significantly higher elution volumes, supporting the model that this domain is involved in MogR oligomerization (Figure 1C). The presence of a C-terminal oligomerization domain likely also explains the enhanced DNA binding affinity of constructs containing this domain (Figure 2 and Supplemental Figure S1).
To determine whether the functional domains of MogR identified in vitro play similar roles in L. monocytogenes, we measured the ability of MogR truncation constructs to repress FlaA protein production in MogR-negative L. monocytogenes (ΔmogR). Deletion of the N-terminal 60 amino acids abrogated FlaA repression (Supplemental Figure S2, lanes 11–14). In contrast, deletion of the C-terminal 36 amino acids (MogR1-270), still permitted FlaA repression, even though MogR1-270 was produced at dramatically lower levels than wild-type MogR and also at lower levels than N-terminally truncated MogR (Supplemental Figure S2, lanes 7–10). Further truncation of the C-terminus (MogR1-220) to remove the leucine zipper region destabilized the protein upon expression in L. monocytogenes, as MogR1-220 could not be detected by Western blot analysis. These findings are consistent with our in vitro analysis that localizes the DNA binding activity to the N-terminus of MogR (Figure 2).
To further understand how MogR recognizes its target sequences, we crystallized the MogR truncation protein comprising residues 1–162 bound to a 15 bp DNA recognition sequence from the flaA promoter region. The MogR1-162 construct was used because it was the minimal domain analyzed sufficient for DNA binding (Figure 1A and Figure 2). The MogR:DNA crystals diffracted to a minimum Bragg spacing of 1.8Å, and the structure was determined using single-wavelength anomalous dispersion (SAD) of a seleneomethionine labeled sample (Supplemental Table 1). The structure was refined using data extending to a minimum Bragg spacing of 1.8 Å and we obtained a Rwork and Rfree of 21.6 and 23.2%, respectively. In our structure, we observed no electron density for MogR region 144–162 indicating that this region is flexible.
The structure contains two MogR1-162 molecules (domain A and B) bound to the 15 bp binding site (Figure 3). The two DNA binding domains are related by a two-fold symmetry axis through the central A:T base pair (in Figure 3A this two-fold axis is indicated by a vertical dashed line). Because of this symmetry, the detailed interactions at the protein-DNA interface are identical for the two domains. The conformation of the 15 bp duplex is regular B-DNA and is bent away from the MogR dimer towards the major groove by ~52°. In the crystal, DNA duplex ends are stacked end-to-end to form pseudocontinuous double helices throughout the crystal. The DNA sequence is pseudopalindromic with a center of symmetry at the central A:T base pair. All base pairs of the 15 bp binding fragment, except the central base pair (A8-T8′), are engaged in MogR binding (Figure 4 and Figure 5). The central base pair is a spacer for the half-operators and does not appear to contribute directly to sequence specificity.
The MogR DNA binding domain contains seven α helices (Figure 3, labeled α1–7 in domain A and α1′–7′ in domain B) connected by short linkers. The first three helices form an antiparallel three-helix bundle. Helix 4 forms a small dimerization interface with a buried surface area of 157A2 (Supplemental Figure S3). Helices 5–7 form a three-helix bundle DNA-binding domain that contains a helix–turn–helix (HTH) motif (α6 and α7), in which α7 is the recognition helix. Helices α5 and α6 are antiparallel, and helix α7 is roughly perpendicular to the axis established by these two helices. The topology of helices 5–7 is most closely related to that seen in homeodomain proteins; a Dali (Holm and Sander, 1996) search shows highest similarity to Pax6 (Xu et al., 1999), a human homeodomain protein, with an RMSD between superimposed Cα atoms from the HTH motif and the minor groove interaction loop of 3.4Å (Z-score 4.3).
One of the most striking features of the structure are the minor groove contacts by residues in the C-terminal loop L3 that add further binding site specificity (Figure 4B). Each half-site is recognized in the major groove by the recognition helix α7 (Figure 4A) and in the minor groove by loop L3 from the symmetry related molecule. Therefore, the whole 15 bp recognition element, except the central A:T base pair, forms a composite binding site for binding of a MogR dimer.
α6 (residues 94–105) and α7 (residues 113–125), separated by a longer than usual seven-residue turn, form a HTH unit. Docking of α7 is stabilized by the phosphate contacts from the N-terminal portion of α6 and from the C-terminal portion of the loop L3 (Figure 3A and Figure 5). Direct base contacts by residues in the recognition helix α7 are responsible for specific DNA sequence recognition. These contacts include (from the N- to the C-terminus): (1) contacts between Ser114 and the N7 of adenine 4′; (2) contacts between Gln117 and the O4 of thymine 3 and thymine 2; (3) bidentate contacts between Asn118 and N7 of adenine 5′ and N6 of adenine 4′; and (4) van der Waals contacts between Tyr121 and the methyl group of thymines 2 and 3 (Figure 4A and Figure 5). In addition, there are a number of water-mediated contacts (Figure 5). These contacts account for conservation of the flanking AT base pairs in MogR binding sites (5′ TTTTNNNNNAAAA 3′) and are consistent with the available biochemical data.
The extended polypeptide loop L3 (residues 127–143) lies in the minor groove and makes extensive contacts over a 5 bp region of the DNA. These contacts are on the opposite face of the DNA helix where the recognition helix α7 of the symmetry related molecule makes major groove contacts and include: (1) van der Waals contacts between Pro138 and C2 of adenine 9 and O2 of the paired thymine 9′; (2) a hydrogen bond between the backbone amide of Gly139 and the O2 of thymine 9′; (3) a hydrogen bond between the backbone amide of Arg140 and N3 of adenine 11; (4) a hydrogen bond between the guanidinium side chain of Arg140 and the O2 of thymine 11′; and (5) a van der Waals contact between the guanidinium side chain of Arg140 to the N3 of adenine 12 (Figures 4B and Figure 5). In addition, Arg140 has water-mediated contacts in the minor groove (Figure 5). These minor groove interactions are important for MogR specificity, as they would restrict against the presence of G:C base pairs at these positions. The exocyclic N2 amino group in the minor groove of G:C base pairs would lead to steric repulsion of the minor groove loop. Therefore we propose that the consensus MogR binding site is 5′ TTTTWWNWWAAAA 3′ (W=A or T).
The MogR recognition site contains two A-tract half sites separated by a central TpA base pair step (residues T7A8 in the binding site). We used the programs 3DNA (Lu and Olson, 2003) and Madbend (Strahs and Schlick, 2000) to analyze the DNA helical parameters. The MogR binding site has a relatively standard B-DNA conformation with an average helical twist of 35.8° (10.0 bp/turn). The three base pairs flanking the central TpA step all have high degrees of propeller twist (> −20°) and engage in bifurcated hydrogen bonds. The minor groove is narrow at both ends of the duplex and widens towards the center with a maximum at the central TpA base pair step (Supplemental Figure S4A). The central six base pairs all exhibit positive roll angles with a maximum of ~10° at the central TpA step (Supplemental Figure S4B). This positive roll of consecutive base pairs results in a global bend of ~52° toward the major groove. Although the DNA duplexes stack to form a pseudocontinuous helix in the crystal, we do not believe that the overall DNA conformation is significantly influenced by packing distortions. The HTH motifs of both domains have canonical phosphate backbone contacts and the close apposition of helices α4 dictates their relative orientation. Therefore, the configuration of the two MogR monomers requires the DNA bend that we observe.
The repression of flagellar gene transcription by MogR is an important component of the regulatory network governing adaptation of L. monocytogenes to infection of the host. Previous studies revealed that MogR down-regulates flagellar gene transcription by binding to 5′ TTTTNNNNNAAAA 3′ sequences enriched within flagellar gene promoter regions (Shen and Higgins, 2006). The precise mechanism by which MogR recognizes its AT-rich binding sequences and distinguishes among the many non-specific sites in the AT-rich genome of L. monocytogenes remained unknown. In this study, we defined the structural elements within MogR that permit stringent recognition of target sequences to mediate repression of flagellar gene expression. Amino acid sequence alignments indicate the presence of MogR homologs in other bacterial species (Figure 6). Residues making important DNA backbone and base specific contacts in helix α7 and loop L3 are entirely conserved. Therefore, a similar DNA binding mechanism and mode of flagellar gene regulation through MogR-like master regulators is likely to occur in these bacterial species.
Our domain analysis of MogR indicated that a truncated protein containing amino acids 1–162 was competent for DNA binding, whereas a truncation containing residues 1–140 was not (Figure 2C). The MogR structure indicates that residues 140–143 are part of loop L3 involved in minor groove binding. The failure of MogR1-140 to bind DNA is consistent with an important role of this minor groove binding loop. We crystallized the MogR truncation containing amino acids 1–162 and observed that MogR residues 144–162 are disordered suggesting that this region is a flexible linker. Analytical ultracentrifugation experiments showed that the crystallized MogR1-162 is monomeric in solution (data not shown).
The region comprising MogR residues 218–263 contains a putative leucine zipper region (Figure 1). The presence of this region significantly increases the apparent size of MogR by gel filtration analyses, suggesting that it is required for dimerization/oligomerization (Figure 1C). MogR mutants lacking this region still bound flaA promoter region DNA in vitro, but with lower apparent affinity (Figure 2 and Supplemental Figure S1), again implying that this region is involved in dimerization/oligomerization. Alternatively, it is possible that full-length MogR adopts an elongated shape in the presence of the C-terminal region leading to reduced mobility in a gel filtration column. Unfortunately, the solution properties of full-length MogR have not allowed us to measure its oligomerization status more rigorously. A MogR mutant lacking the leucine zipper domain (MogR1-220) was destabilized in L. monocytogenes and thus failed to repress flagellar gene transcription compared to MogR proteins harboring the leucine zipper domain (Figure 1 and Supplemental Figure S2). Although we were unable to rigorously assess the function of the leucine zipper domain in vivo or in vitro, we favor the model that the C-terminal region is required for MogR dimerization and higher order interaction on tandem MogR binding sites.
MogR recognizes its DNA target sequence through a helix-turn-helix (HTH) motif (Figure 3). As in most other HTH-containing proteins, MogR functions as a homodimer with each motif binding one half-site of the symmetry related target DNA site. Major groove interactions of α7 of the HTH motif account for the strict conservation of the flanking T and A base pairs in the binding site 5′ TTTTNNNNNAAAA 3′ (Figure 4A). One of most striking features are the extensive minor groove interactions by loop L3 that partially overlap with the binding site of the symmetry related molecule (Figure 4B). Minor groove contacts by an extended polypeptide chain have been seen in other complexes, such as homeodomains and interferon response factors (Kissinger et al., 1990; Panne et al., 2004; Wolberger et al., 1991; Xu et al., 1999). MogR is unusual because the major and minor groove interactions of each domain are mutually overlapping and therefore the binding site is a composite site for recognition of a MogR dimer. This ‘oversampling’ of each half-site is unusual among HTH-containing proteins and clearly increases site specificity. One possibility is that this ‘oversampling’ has evolved to distinguish among closely related sites in the AT-rich genome. The minor groove interactions also account for evolutionary conservation of the stretch of A:T base pairs in the center of MogR binding sites, as they are additional determinants of site specificity. We propose that the consensus MogR site is 5′ TTTTWWNWWAAAA 3′ (W=A or T). This sequence specificity also explains our previous observation that the central nucleotides are more permissive to mutation provided that the substitutions are conservative (A->T or T->A) (Shen and Higgins, 2006). The refinement of MogR’s recognition specificity reduces the number of possible MogR binding sites in the L. monocytogenes genome from 534 to 92 (within 300 bp upstream of gene start codons, 0 mismatch permitted) and from 4884 to 1404 (1 mismatch permitted).
An emerging theme in protein-DNA recognition is that minor groove interactions can be important to discriminate among closely related binding sites. In the homeodomain protein Scr, a loop inserts into the minor groove only in the functionally relevant target site (Joshi et al., 2007). Specificity is obtained by recognizing a sequence-dependent DNA structure rather than by direct DNA sequence readout. By analogy to Scr, we suggest that (1) the major groove interactions target MogR to AT-rich binding sites, (2) the additional minor groove interactions are important to discriminate among closely related target sites and (3) the intrinsic DNA structure of the target site is an important parameter in site recognition. Further experiments are required to analyze MogR binding site requirements in more detail.
Seventy-two genes contain at least one of the 92 consensus 5′ TTTTWWNWWAAAA 3′ MogR binding sites within their promoter regions (Supplemental Table 2) and all of the known MogR-regulated promoters contain this sequence. Strikingly, promoters of MogR-regulated flagellar genes usually contain pairs of MogR binding sites that are spaced one, two or three helical turns apart (Supplemental Table 2; (Shen and Higgins, 2006)). At least two MogR binding sites are required for both high affinity binding in vitro and repressor activity in vivo, another feature of MogR binding that presumably helps prevent spurious repression in the context of the AT-rich L. monocytogenes genome (Gründling et al., 2004; Shen and Higgins, 2006). This arrangement suggests that two MogR dimers further interact on the same face of the DNA helix to form a higher order (tetrameric) assembly. Since a single MogR dimer bends the target site by ~52°, the cumulative bend of two such sites in phase would be at least ~104°. Within the flaA promoter region, two MogR binding sites flank the −35 region, and therefore it is likely that the higher order assembly leads to steric occlusion of the promoter as suggested previously (Shen and Higgins, 2006). In most MogR-regulated promoters, one of the MogR binding sites corresponds to a MogR consensus sequence whereas the others are non-consensus binding sites. We note that graded affinity for tandem operators and cooperative binding to operator sequences are central components of the mechanistic function of well-studied phage and lac repressors (Müller-Hill, 1996; Ptashne, 1986; Stayrook et al., 2008) and thus are likely to also play a key role in MogR-mediated gene regulation.
Consecutive ApA, TpT or ApT base pair steps are known to result in a narrow minor groove due to negative propeller twisting that is stabilized by inter-base pair interactions in the major groove (Crothers and Shakked, 1999). In contrast, TpA base pair steps do not form these interactions and tend to widen the minor groove (Stefl et al., 2004). Hence, A-tract sequences are defined as four or more consecutive A:T base pairs without a TpA step (Hud and Plavec, 2003). By that definition, the flaA promoter region binding site consists of two consecutive A-tract elements disrupted by a central TpA step 5′ ATTTTTTAAAAAAAT 3′ (TpA underlined). Considering directionality, the two A-tracts in the site are oriented ‘head-to-head’ with their 5′ ends in the center of the binding site (at the TpA base pair step). A-tracts in the 5′ -> 3′ direction usually exhibit a narrow minor groove (Crothers and Shakked, 1999; Hud and Plavec, 2003). Consistent with this prediction, we observe that the minor groove is narrow at both ends of the site and symmetrically widens towards the center with a maximum at the central TpA base pair step (Supplemental Figure S4A). In order to distinguish between intrinsic and MogR-induced structural features of the DNA target site, we compared our structure to the DNA duplex containing a T4A4 A-tract (5′CGTTTTAAAACG 3′) as studied by solution NMR (Stefl et al., 2004). In accord with the predictions, the NMR structure also exhibits narrow minor grooves towards the ends of the duplex and a maximum width at the central TpA step (Supplemental Figure S4A). Therefore, we assume that the minor groove geometry as seen in our MogR structure is an intrinsic structural property of the DNA sequence. In the MogR structure, Arg140 inserts at a position where the minor groove is very narrow. It is known that electrostatic focusing in such narrowed grooves generates a negative electrostatic potential, which can provide the anchoring point for the positively charged Arginines (Honig and Nicholls, 1995; Joshi et al., 2007).
The central base pairs exhibit large deviations in local helical parameters, which are important for the overall conformation of the site (Supplemental Figure S4B). The smooth ~52° bend in the MogR DNA is due to consecutive positive roll angles in the central six nucleotide base pair steps with a maximum at the TpA step (Supplemental Figure S4B). We have examined the role of individual base pair steps in the binding site by using quantitative multiple fluorescence relative affinity (QuMFRA) binding studies (Man and Stormo, 2001). We observe that a GC base pair at the central position 8 does not inhibit binding significantly. However, GC base pairs at positions 7 and 9 or at 6 and 10 significantly decrease relative binding affinity (Supplemental Figure 5). Whereas the N2 amide of guanine in the minor groove can be tolerated at the central position 8, at positions 6, 7, 9 and 10, they directly interfere with binding of minor groove loop L3. We also examined the role of TpA base steps in the center of the binding site. We introduced a TpA step at positions 7 and 9. TpA steps are predicted to not interfere directly with minor groove binding but to affect minor groove geometry. We observed that TpA steps at positions 7 and 9 significantly reduced binding affinity. Our interpretation is that sequence-dependent conformation of the minor groove is an important parameter for site recognition and that MogR affinity is modulated by the intrinsic structure of the central nucleotides (positions 6–10). In other protein-DNA complexes, non-contacted bases are frequently important determinants of binding affinity (Anderson et al., 1987; Hizver et al., 2001; Otwinowski et al., 1988). One explanation is that DNA binding proteins not only read the DNA sequence but also sequence-dependent DNA structure. Such sequence-dependent conformational characteristics of DNA binding sites and their ‘indirect readout’ seems to explain in part why there is no simple ‘code’ in protein –DNA recognition (Harrison, 2007).
Details regarding the construction of bacterial strains, Western blot, gel mobility shift, and DNAse I footprinting analyses used in this study are described in Supplemental Materials.
Sixteen to twenty hour cultures of pET29b-MogRx variants in E. coli BL21(DE3) (Novagen) were diluted 1:1000 into 250 ml of 2YT media and grown at 37°C shaking until an OD~0.6 was reached (approximately 3.5 hr). IPTG was added at 350 μM, and cultures were induced for another 4 hr at 25°C. Cultures were pelleted and resuspended in 10 ml of LIB buffer (100 mM Tris [pH 7.5], 15 mM imidazole, 500 mM NaCl, 10% glycerol) supplemented with 2 mM β-mercaptoethanol, 1 mg/ml lysozyme, and Complete-EDTA protease inhibitor mixture (Roche) and lysed by sonication. The lysate was cleared by pelleting at 15 500 × g for a minimum of 30 min. Two-hundred-fifty microliters of Ni-NTA agarose beads (Qiagen) was added to the cleared lysate and incubated with rotation for a minimum of 2 hr at 4°C. Samples were pelleted at 1 800 × g for 1 min and washed 3 times in 4 ml of LIB buffer. His6-tagged MogR variants were eluted by adding 600μl of HIB buffer (100 mM Tris [pH 7.5], 350 mM imidazole, 500 mM NaCl, 10% glycerol) and incubating with rotation for 15 min at 4°C.
For purification of His6-tagged MogR1-162 used in crystallization studies, 16–20 hr cultures of the appropriate E. coli strain were diluted 1:1000 in 5 L 2YT media and grown shaking at 37°C. When an OD600 of 0.6 was reached, IPTG was added at 350 μM, and cultures were grown for 4 hr at a reduced temperature of 30°C. Cultures were pelleted, resuspended in 25 ml lysis buffer [500 mM NaCl, 100 mM Tris, pH 7.5, 15 mM imidazole, 10% glycerol] and flash frozen in liquid nitrogen. Lysates were thawed then lysed by sonication, and cleared by pelleting at 15 000 × g for 30 min. His6-tagged MogR1-162 was batch affinity purified by incubating the lysates with 1.5 ml Ni-NTA agarose beads (Qiagen) shaking for 1 hr at 4°C. Selenomethionine-substituted MogR1-162 was produced by metabolic labeling in E. coli BL21(DE3) grown in minimal M9 medium supplemented with essential vitamins, nucleotide bases, and amino acids, with Met replaced by SeMet [procedure modified from that of (Van Duyne et al., 1993)]. Selenomethionine incorporation was verified by MALDI TOF/TOF mass-spectroscopic analysis after trypsin digestion.
Analytical gel filtration was carried out in an AKTA FPLC unit using a HiPrep 16/60 Sephacryl S-300 HR (GE Healthcare). Five-hundred microliters of affinity purified protein was injected into a column equilibrated with 500 mM NaCl, 100 mM Tris pH 7.5, 10% glycerol with 0.5 mM TCEP. Elution was tracked by absorbance at 280 nm at a flow rate of 0.5 ml/min. The column was calibrated using the following standards (Biorad): thyroglobulin (670 kDa), gamma-globulin (158 kDa), ovalbumin (44 kDa), myoglobin (17 kDa), vitamin B12 (1.35 kDa). The calibration curve was log Mr = 7.46−0.04Ve (correlation coefficient > 0.99). The elution volume (Ve) was assigned using AKTA PrimeViewer software (GE Healthcare).
Commercially synthesized DNA oligonucleotides were resuspended in 10 mM NH4CO3, pH 8.0 at a final concentration of 50 μM and annealed. A series of 6 different duplexes ranging in size from 15 to 25 bp, all containing a single A/T overhang was tested in crystallization trials with the 15mer yielding well diffracting crystals.
The complex of MogR and DNA was prepared by mixing MogR1-162 with DNA at a final concentration of 190 μM:200 μM in water. Crystals suitable for X-ray structural analysis were obtained with the hanging drop vapor diffusion method. A volume of 1 μl of the MogR:DNA complex solution was mixed with 1 μl precipitant solution containing 100 mM 2-(N-morpholino)ethanesulfonic acid (MES), pH 6.0, 10–15% (w/v) PEG 4000, and equilibrated against 1 ml well solution. The first crystals appeared ~20 min after setup of the drops. The largest crystals grew as rods to a maximum size of 0.2 × 0.05 × 0.05 mm3 at 21°C. Crystals were stable in a cryoprotectant buffer containing 50% (v/v) well solution and 50% (v/v) 100 mM MES, pH 6.0, 11% (w/v) PEG 4000, 50% glycerol. The native and derivative data sets were obtained by flash freezing the crystals in liquid nitrogen, and data were collected under cryogenic conditions (100 K). We collected native and derivative diffraction data at the Argonne National Laboratory Advanced Photon Source, beamline ID24, under a nitrogen gas stream at 100 K, using a wavelength of 0.979 Å. We processed the data with HKL2000 (Otwinowski and Minor, 1997) (Supplemental Table 1).
The crystals belong to the space group P21 with the unit cell dimensions a=48.34 Å, b=93.59 Å, c=60.76Å and b= 113.40°. Initial phases were calculated by using the SHELX-suite as implemented in HKL2MAP (Pape and Schneider, 2004). Eight high-occupancy sites were found resulting in an initial correlation coefficient of 0.41 for the best solution. After density modification, the map was of sufficient quality for automated model building by ARP/wARP (Perrakis et al., 1999), as implemented in the CCP4 package (CCP4, 1994). The initial model contained 98 residues of MogR (68% of the final model). DNA was built in O (Jones et al., 1991). Structure refinement involved the rebuilding of several loops and extension of both polypeptides towards the N and C termini and was performed with Coot (Emsley and Cowtan, 2004). The structure was refined with multiple rounds of CNS 1.1 (Brunger et al., 1998) and model building in Coot. We used NCS restraints between domains A and B of MogR that were released in the final cycle of refinement. The Ramachandran plot for each domain indicated 93% of residues in the most favored region (119/8/1/0; favored/allowed/generously allowed/disallowed). The local DNA helical parameters were calculated using 3DNA (Lu and Olson, 2003) and the global bend angles obtained using Madbend (Strahs and Schlick, 2000).
We would like to thank Stephen C. Harrison for support and comments on the manuscript; Ann Hochschild for comments on the manuscript, E. Settembre and A. Schmidt for help with data collection. Data was collected at beamline ID-24 at the Advanced Photon Source (APS) of Argonne National Laboratory. Coordinates and structure factors have been deposited in the RSCB Protein Data Bank with accession code 3FDQ. This work was supported by U.S. Public Health Service grant AI53669 from the National Institutes of Health (DEH). AS was a recipient of a Howard Hughes Medical Institute predoctoral fellowship award.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.