Crystallographic and Functional Study on DNA Binding Proteins

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Three-component ParABS partition systems ensure stable inheritance of many bacterial chromosomes and low-copy-number plasmids. ParA localizes to the nucleoid through its ATP-dependent nonspecific DNA-binding activity, whereas centromere-like parS-DNA and ParB form partition complexes that activate ParA-ATPase to drive the system dynamics. The essential parS sequence arrangements vary among ParABS systems, reflecting the architectural diversity of their partition complexes. Here, we focus on the pSM19035 plasmid partition system that uses a ParBpSM of the ribbon-helix-helix (RHH) family. We show that parSpSM with four or more contiguous ParBpSM-binding sequence repeats is required to assemble a stable ParApSM-ParBpSM complex and efficiently activate the ParApSM-ATPase, stimulating complex disassembly. Disruption of the contiguity of the parSpSM sequence array destabilizes the ParApSM-ParBpSM complex and prevents efficient ATPase activation. Our findings reveal the unique architecture of the pSM19035 partition complex and how it interacts with nucleoid-bound ParApSM-ATP. Editor's evaluation The work by Volante et al. studied a plasmid partition system, in which the authors discovered that four or more contiguous ParS sequence repeats are required to assemble a stable partitioning ParAB complex and activate the ParA ATPase. The work reveals a plasmid partitioning mechanism in which the mechanic property of DNA and its interaction with the partition complex may drive the directional movement of the plasmid. https://doi.org/10.7554/eLife.79480.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Faithful chromosome segregation is essential for the proliferation of bacterial cells, and low-copy-number plasmids also need a robust partition mechanism for their stable inheritance. However, prokaryotes do not possess the mitotic machinery of eukaryotes: instead, alternative active DNA partition systems have evolved, among which ParABS systems (also called class I partition systems) are the most widespread. Basic ParABS systems consist of three components, a partition ATPase (ParA), a ‘centromere’-binding protein (ParB), and a cis-acting centromere-like DNA site (parS). ATP-activated ParA dimers bind nonspecific DNA (nsDNA) and localize to the nucleoid in vivo (Ebersbach and Gerdes, 2001; Ebersbach and Gerdes, 2004; Pratto et al., 2008; Ringgaard et al., 2009). The parS sites, often composed of multiple tandem repeats of binding consensus sequences for ParBs, demark the DNA-cargos that are translocated and positioned into the two halves of the cell before cell division by recruiting ParB molecules to assemble partition complexes (PCs). ParB proteins fall into two structurally unrelated groups, dimeric helix-turn-helix (HTH), and dimeric ribbon-helix-helix (RHH) DNA-binding proteins. HTH-ParBs have been shown to bind not only site-specifically to their cognate parS sequences but also to spread many kilobase pairs into the DNA neighboring the parS sites (Murray et al., 2006; Graham et al., 2014; Soh et al., 2019; Rodionov et al., 1999; Lynch and Wang, 1995; Breier and Grossman, 2007; Jalal et al., 2020; Osorio-Valeriano et al., 2019; Sanchez et al., 2015). Therefore, they form large PCs containing many ParB molecules bound to condensed DNA around parS, and ParB spreading activity is essential for their partition function (Rodionov et al., 1999; Debaugny et al., 2018). In contrast, RHH-ParB does not spread beyond parS site judged by the absence of ParB-mediated silencing of parS-proximal gene expression (J. C. Alonso, unpublished observation), unlike HTH-ParBs (Lynch and Wang, 1995; Rodionov et al., 1999), but like HTH-ParBs, they interact with their cognate ParA proteins via their N-terminus (Radnedge et al., 1998; Figge et al., 2003; Barillà et al., 2007). RHH-ParB proteins also control the expression of the proteins involved in the partition system and plasmid copy number control by binding parS sites, which overlaps promoters of their genes (de la Hoz et al., 2000). ParA–ParB interaction leads to activation of the ParA-ATPase, which is most efficient in the presence of nsDNA and parS DNA (Ah-Seng et al., 2009; Chu et al., 2019; Pratto et al., 2008; Taylor et al., 2021) and leads to dissociation of ParA from nsDNA. Interaction dynamics between the nucleoid-bound ParA and ParB in the PC prior to ATP hydrolysis and ParA dissociation determine the dynamics of the PC relative to the nucleoid. The common results of most systems in vivo appear to be the establishment of equidistant distribution of two or more PCs along the nucleoid(s) so that at cell division each daughter cell inherits at least one copy of the plasmid DNA (Sengupta et al., 2010; Ringgaard et al., 2009; Lioy et al., 2015; McLeod et al., 2017). With biochemical findings and observations from live-cell imaging approaches accumulating in the field, combined with experiments using reconstituted cell-free reaction systems, a diffusion-ratchet mechanism of the ParABS partition was proposed. Here, the driving force for the DNA-cargo motion is generated by a propagating nucleoid-bound ParA distribution gradient (Vecchiarelli et al., 2010; Hwang et al., 2013; Vecchiarelli et al., 2013; Vecchiarelli et al., 2014; Sugawara and Kaneko, 2011). Additional models related to the diffusion-ratchet mechanism have also been proposed based on high-resolution imaging observations (Lim et al., 2014; Le Gall et al., 2016). While findings supporting diffusion-ratchet-type models for ParABS systems accumulate, many molecular details required to put the model on quantitatively solid ground are lacking. Large number of HTH-ParB dimers load onto a PC by ParB ‘spreading’ to facilitate partitioning. Resulting high concentration of ParB at the PC assures near saturation ParB-binding to the local nucleoid-bound ParA molecules, leading to efficient sensing of the local ParA distribution gradient by the PC. Combined with relatively short lifetime of the ParA/ParB-mediated cargo-nucleoid bridges and sufficient ParA-ATP reloading rate to the nucleoid would maintain enough cargo-nucleoid bridges needed to significantly suppress thermal diffusion of the cargo without completely blocking the motion of the cargo. Proper balance among these parameters is needed for efficient plasmid partition by diffusion-ratchet mechanism. Insufficient free diffusion suppression results in random diffusion of the cargo, and oversuppression by too many or too stable bridges blocks the cargo motion altogether (Hu et al., 2015; Hu et al., 2017; Taylor et al., 2021). However, it is unclear how a system with RHH-ParBs, which cannot spread ParB beyond the parS sites and thus can load only limited number of ParB dimers to a PC, can fit into the diffusion-ratchet paradigm. Nevertheless, live-cell imaging studies of systems that use RHH-ParB proteins showed an oscillating dynamic ParA distribution pattern on the nucleoid and PC chasing the receding tail of the ParA distribution (Ringgaard et al., 2009; Lioy et al., 2015; McLeod et al., 2017) similar to observations with the F-plasmid partition system and other related systems involving HTH-ParB proteins (Hatano and Niki, 2010; Schofield et al., 2010) for which there is accumulating evidence supporting the diffusion-ratchet mechanism. Therefore, we suspect systems with RHH-ParBs also operate via a diffusion-ratchet-type mechanism. A better understanding of how the PCs are organized for this group and a quantitative understanding of the interaction dynamics of these PCs with nucleoid-bound ParA molecules are critical for advancing our mechanistic understanding of these systems. In this study, we focused on the pSM19035 partition system of Streptococcus pyogenes. This plasmid harbors a ParABS system composed of ParApSM (also called Delta), an RHH ParBpSM (also called Omega), and six parSpSM sites, each comprising 7–10 consecutive non-palindromic 7-bp-long sequence repeats (5′-WATCACW-3′, symbolized by →) that overlap the promoter regions of copS, δ (coding ParApSM) and ω (coding ParBpSM) genes. Each ParBpSM dimer binds one copy of the 7 bp parSpSM consensus sequence. However, the affinity for a single repeat is low, while two dimers bind with high affinity to two direct (→→) or inverted (→←) repeats forming dimers of dimers (de la Hoz et al., 2004; Weihofen et al., 2006; Welfle et al., 2005). Within the structures of these ParBpSM-parSpSM complexes, DNA does not show significant curvature, and although the protein dimers bound to each DNA sequence repeat slightly deviates from dyad symmetry due to the bound DNA sequence asymmetry, full-size parSpSM-ParBpSM complexes could be modeled as nearly straight DNA wrapped by left-handed spiral arrangement of ParBpSM dimers (Weihofen et al., 2006). Atomic force microscopy images of the complex involving seven parSpSM consensus sequence repeats supported its straight arrangement and the lack of spreading (Pratto et al., 2009). ParApSM, unlike most other ParAs, forms dimers in solution in the absence of ATP (Pratto et al., 2008). Like other ParAs, it also undergoes a conformational transition upon binding ATP that increases its affinity for nsDNA (Soberón et al., 2011; Pratto et al., 2008). In the ATP-bound form, ParApSM has been shown to bind nsDNA forming limited size patches containing several ParApSM dimers at random location, instead of individual dimers independently distributed on the nsDNA (Pratto et al., 2009). In the presence of parSpSM DNA and ParBpSM, several parSpSM-ParBpSM mini-filaments and ParApSM-nsDNA patches appeared to bind together to form large protein-DNA complexes bridging multiple DNA molecules (Pratto et al., 2008; Pratto et al., 2009; Soberón et al., 2011; Lioy et al., 2015). Interactions among these components fueled by ATP hydrolysis are thought to drive dynamic oscillations of the nucleoid-bound ParApSM in vivo, which resembles those observed for the TP228 ParABS system (Lioy et al., 2015; McLeod et al., 2017). Despite accumulating information summarized above, how the observed inter-molecular interactions coordinate the in vivo system dynamics resulting in robust plasmid partitioning remains a mystery. To approach this puzzle, here we studied functional requirements of the parSpSM sequence-structure necessary for ParBpSM-mediated activation of the ParApSM ATPase. We found a minimum of four contiguous repeats of the parSpSM heptad consensus sequence without a gap is necessary for full activity. Kinetics of the nsDNA-bound parSpSM-ParBpSM-ParApSM complex formation and disassembly indicated the presence of a complex multistep process involved in ATPase activation. Results ParApSM ATPase is synergistically activated by nsDNA, ParBpSM, and parSpSM-DNA To define the requirements for stimulation of the ParApSM ATPase activity by ParBpSM, the steady-state ParApSM ATP turnover rate was measured with varying concentrations of ParBpSM in the presence or absence of different duplex DNA cofactors. The turnover rate of ParApSM ATPase alone is very low at 37°C (0.9 ± 0.1 ATP/ParA-dimer/h, N = 3; Figure 1A; no DNA). In the presence of a saturating concentration of double-stranded DNA (40 μg/ml pBR322 plasmid DNA plus 23–38 μg/ml double-stranded oligonucleotide with or without parSpSM sequence), to which ATP-ParApSM dimers can bind to support ATPase activation by ParBpSM (see below), no significant rate change was observed (1.0 ± 0.1 h–1, N = 46, Figure 1A, pool of all measurements at [ParBpSM] = 0). Next, effects of the parSpSM-DNA in addition to 40 μg/ml nsDNA and ParBpSM were examined. ParApSM ATP hydrolysis was stimulated up to ~20-fold (kcat = 20.5 ± 2.9 h–1, N = 7) in the presence of ParBpSM and oligonucleotide duplex DNA containing one of the native arrangements containing seven parSpSM heptad-sequence-repeats (7R-parSpSM, →→←→→←←) (Figure 1A). The stimulation approached saturation around 2 μM ParBpSM at varying ParApSM concentrations (Figure 1—figure supplement 1A). In contrast, when the parSpSM DNA fragment was replaced with one having a scrambled sequence, ParBpSM stimulated ParApSM ATPase activity only to 2.4 ± 1.1 h–1 even in the presence of 8 μM ParBpSM that should have allowed non-parSpSM DNA binding (N = 3*) (Figure 1A, scram). Similarly, a low level of ParApSM ATPase stimulation was observed with ParBpSM1-27 peptide, which lacked the DNA-binding and dimerization domains (Figure 1—figure supplement 1B). These results demonstrated that full ParApSM-ATPase stimulation by ParBpSM requires specific parSpSM interactions. Even the low parSpSM-independent ATPase stimulation was not detected in the absence of DNA (Figure 1A, no DNA). Similarly, the low-level ATPase stimulation by non-parSpSM-binding ParBpSM1-27 peptide was not detected without DNA (Figure 1—figure supplement 1B). We conclude ParApSM-nsDNA binding is required for the ATPase activation by ParBpSM, as reported for ParAF ATPase activation by ParBF (Taylor et al., 2021). Consistently, in the absence of nsDNA, excess ParBpSM relative to the concentration of 7R-parSpSM in the reaction inhibited ATPase stimulation, presumably competing with ParApSM for parSpSM DNA binding (Figure 1—figure supplement 1C). Figure 1 with 3 supplements see all Download asset Open asset ParApSM ATPase activation by ParBpSM, parSpSM, and nsDNA. (A) Efficient activation of ParApSM ATPase by ParBpSM exhibits critical dependency on parSpSM heptad-sequence-repeat number. The ATPase reactions contained ParApSM (2 μM), pBR322 DNA (60 μM in bp, unless noted otherwise), ParBpSM (at the concentration indicated), and parSpSM duplex substrates (4.4 μM of the 7 bp consensus sequence repeats) or equal amount of a scrambled sequence duplex (scram). (B) Comparison of parSpSM containing four contiguous repeats with different heptad orientation arrangements. (C) parSpSM containing different heptad arrangements of three contiguous repeats and two contiguous triple heptad repeats with a gap fails to fully activate the ParApSM ATPase. (D) 7R-parSpSM concentration dependence of ParApSM ATPase activity. Reaction mixtures contained ParApSM (2 μM), ParBpSM (2 μM), pBR322 DNA (60 μM in bp), and increasing concentration of 7R-parSpSM duplex (duplex fragment concentration shown, the ratio of parSpSM heptad repeat sequence to ParBpSM dimers are indicated on top). (E) The numbers and arrangement of the heptad repeats of the parSpSM fragments used in this study (also see Figure 1—figure supplement 2). Data points represent means and standard errors of mean (SEM) of N repeated experiments (N* represents repeats for majority of data points, see Figure 1—source data 1 for details). Curves were fitted after subtraction of the background in the absence of ParApSM to an equation v-v0 = (vmax[B]n)/(KAn + [B]n). The maximum turnover rates (vmax) cited in the text represent mean ± 95% confidence intervals (symmetrized to the larger estimated errors from the mean for simplicity). Figure 1—source data 1 https://cdn.elifesciences.org/articles/79480/elife-79480-fig1-data1-v3.xlsx Download elife-79480-fig1-data1-v3.xlsx Formation of short clusters of ParApSM molecules bound to nsDNA might be taken as a hint for cytoskeletal filament treadmilling models for this class of partition systems, rather than diffusion-ratchet model. However, the low maximum ATP turnover rate (~0.3 /min) observed would be too slow for a treadmilling filament model. Combined with low abundance of ParApSM molecules inside a cell insufficient to form axial protein filaments, the treadmilling model is highly unlikely to fit this system. Efficient ParApSM-ATPase stimulation requires ParBpSM bound to parSpSM-DNA with at least four contiguous heptad-sequence-repeats We asked whether entire 7R-parSpSM is required for the full stimulation of ParApSM ATPase by ParBpSM. A series of deletions of parSpSM heptad-sequence-repeats were made and their ATPase stimulation activities were tested (Figure 1E). Efficient ATPase stimulation was observed when 6R (→←→→←←), 5R (→→→←←), or 4R (→→←←) was added along ParBpSM (Figure 1A). No significant difference was observed among 5R, 6R, or 7R; both half-saturation concentrations and the apparent kcat were comparable (Figure 1A). ParBpSM-4R-parSpSM also induced similar rates of ParApSM ATP hydrolysis, but a significantly higher concentration was required for full stimulation (Figure 1A). Different heptad orientation arrangements of 4R-parSpSM (→→←→ and →→→→) stimulated ParApSM ATPase to a similar extent (Figure 1B). In contrast, 3R- (→→←), 2R- (→←), and 1R- (→) parSpSM were poor cofactors for ParBpSM-dependent ATPase stimulation (apparent kcat = 3.2–3.6 ± 1.3–2.1 h–1, N = 4–7*, Figure 1A), not significantly different from the scrambled sequence DNA. 3R-parSpSM fragments with different repeat arrangements behaved similarly to each other (Figure 1C). The ATPase stimulation by parSpSM DNA with 1–3 parSpSM repeats also appeared to saturate at around 2 μM ParBpSM, indicating that the affinity of the nsDNA-bound ParApSM to ParBpSM in the presence of truncated parSpSM was not limiting above ~2 μM ParBpSM. Previously it has been shown that ParBpSM bound 3R, 4R DNA fragments of different sequence orientation combinations, or a 10R DNA fragment with roughly similar affinity that was >50-fold higher than nsDNA or 1R DNA (de la Hoz et al., 2004). We confirmed that ParBpSM binding was similarly strong for 4R and 3R duplex DNA (KD ~ 17 nM) and weak for nsDNA (KD ~ 1 μM, Figure 1—figure supplement 3). Therefore, the affinity of ParBpSM for the 3R-parSpSM DNA is not limiting the ParApSM ATPase stimulation. In the above experiment, the length of parSpSM DNA fragments decreased as the number of parSpSM repeats decreased (Figure 1E, Figure 1—figure supplement 2). We tested longer 3R-parSpSM DNA fragments containing an additional non-parSpSM heptamer sequence (non-consensus, nc) (3R-1nc) and confirmed that parSpSM-DNA fragment size was not a significant factor for the ATPase stimulation efficiency (Figure 1C). These results suggested that ParApSM ATP turnover is fine-tuned by the structural arrangement of the ParBpSM and parSpSM within the PC. ParBpSM assembles as a left-handed spiral to wrap parSpSM DNA without significantly distorting the DNA backbone geometry (Weihofen et al., 2006), implying that after approximately four repeats, ParBpSM dimers would make a full turn around parSpSM, positioning themselves on the same face of the nucleoprotein filament. Thus, the position of the fourth repeats relative to that of the first repeat might be functionally important. To test this possibility, we examined the ParApSM ATPase stimulation efficiency of parSpSM DNA fragments containing two copies of 3R sequences separated by 7 bp or 14 bp of non-consensus sequence (3R-1nc-3R and 3R-2nc-3R) (Figure 1E). The 3R-1nc-3R-parSpSM fragments showed only slightly higher stimulation (kcat = 8.3 ± 3.5 h–1, N = 8*) compared to the single 3R-parSpSM fragment, and the 3R-2nc-3R was functionally indistinguishable from the single 3R-parSpSM fragment (Figure 1C). We conclude that disrupted 6R-parSpSM fragments are unable to recover the full stimulation of ParApSM ATPase activity. These findings indicate that the ParBpSM-ParApSM interactions differ when ParBpSM dimers are bound to ≥4 contiguous heptad repeats compared to ParBpSM dimers unbound to the repeats or bound to fewer or a disrupted array of heptad repeats. Substoichiometric concentration of parSpSM relative to ParBpSM is sufficient to fully activate ParApSM ATPase We originally assumed that for efficient ParApSM ATPase activation by the ParBpSM-parSpSM complex all ParBpSM dimers need to be bound to parSpSM and accordingly maintained stoichiometrically excess parSpSM-sequence concentration relative to ParBpSM-dimers in the reaction. To test this assumption, we next changed the concentration of the 7R-parSpSM DNA fragment while keeping the concentrations of ParApSM and ParBpSM both at 2 μM. Contrary to our expectation, significantly lower heptad consensus sequence concentrations of the 7R-parSpSM DNA fragment compared to ParBpSM dimers (B2) were sufficient for ATPase activation (Figure 1D). If the original assumption was correct, 2 μM ParBpSM should have required 1 μM consensus sequence repeats (143 nM 7R-parSpSM) for full activation. According to the results of Figure 1A, at high 7R-parSpSM concentration, half-activation by ParBpSM required ~650 nM ParBpSM, which would have required ~325 nM parSpSM consensus sequence if full binding was needed. Observed half-saturation parSpSM repeat concentration was ~130 nM (~19 nM 7R-parSpSM) or less. Thus, assuming most of the ParBpSM molecules in our preparations are active, it appears that at any given time, less than half of the ParBpSM dimers need to be in complex with 7R-parSpSM to exert full ATPase activation. Stimulation of ParApSM release from nsDNA-carpet by ParBpSM requires parSpSM DNA with four or more heptad repeats Next, we examined how the dissociation of ParApSM-ATP dimers from nsDNA is influenced by ParBpSM in the presence of parSpSM DNA. For these experiments, we used an nsDNA-carpeted two-inlet flow cell observed under TIRF microscopy as described in and (Vecchiarelli et al., 2013; Figure We used a protein and a to to facilitate of protein the Stimulation of ATPase activity by ParBpSM and was comparable to ParApSM (Figure supplement we tested the of individual proteins at 1 the The affinity of to nsDNA was in the absence of ParApSM on the a of ± after of flow (Figure supplement the solution was to a without most ParBpSM with that can be fitted to a function = ± = ± N = 3; all reported in this study represent apparent rate (Figure supplement with ATP and a saturation on the nsDNA-carpet of (N = whereas in the absence of less than of ParApSM bound to the nsDNA-carpet (Figure supplement top). Next, 1 μM with ATP was onto the nsDNA-carpet of saturation at which the flow was to containing ATP without A of than within and the with of (N = Figure Figure 2 with 3 supplements see all Download asset Open asset Kinetics of ParApSM disassembly from the (A) of two-inlet flow cell used for the and dissociation of proteins on and each containing double-stranded DNA fragments or alone as were at different flow rates or from two on the into a flow the of the flow of the flow the observations are of the the the flow rates of the two protein binding to the nsDNA-carpet and dissociation the can be (B) ParBpSM in the presence of parSpSM with at least four contiguous ParBpSM-binding sequence repeats the dissociation from the with 1 ATP was into the nsDNA-carpeted flow cell at while the solution containing the components was at the on the nsDNA-carpet of the saturation = ParApSM the flow rates were to the with solution bp fragment nM in bp), or with ParBpSM without or with different parSpSM fragment μM parSpSM heptad repeat The the to that at = Each represents the mean with to the standard errors of mean (SEM) of N repeated Figure data 1 Download The addition of nsDNA nM scrambled bp duplex in the slightly the of majority of = ± N = in by competing with of to the nsDNA-carpet (Figure 1 μM ParBpSM or was added to the without nsDNA, of from nsDNA-carpet ± (N = Figure dissociation from the nsDNA-carpet was when solution contained nM 7R-parSpSM DNA μM of the consensus sequence and 1 μM ParBpSM = ± after a of dissociation rate (see for (Figure This was or higher compared to ParBpSM or nsDNA Similarly, 4R-parSpSM DNA supported release by ParBpSM, but 3R-parSpSM DNA not (Figure The results those shown in Figure the that a functional parSpSM four copies of ParBpSM sequence repeats. The dependence of ParApSM dissociation on the heptad repeat number of the parSpSM fragment combined with 1 μM ParBpSM in the solution when the was with ParApSM bound to the nsDNA-carpet at a saturating was However, the dissociation were significantly (Figure supplement reflecting the approximately higher on the for which the concentration of ParBpSM used was release of ParApSM from only after of ParBpSM on the in the presence of or 4R-parSpSM The release from the nsDNA-carpet when with or parSpSM complexes after an rate of ParApSM release (Figure This to the binding of ParBpSM to the nsDNA-carpet at different concentrations of ParBpSM. the flow to the solution without parSpSM, ParBpSM bound to the nsDNA-carpet to a of and decreased to a of (Figure The of ParBpSM binding appeared to roughly with the dissociation of the slow dissociation of ParApSM, ParBpSM on the nsDNA-carpet relatively with the protein ratio after several of with 1 μM ParBpSM (Figure Figure 3 Download asset Open asset ParBpSM-parSpSM concentration of ATP ParApSM disassembly from the (A) with 1 ATP was into the nsDNA-carpeted flow cell at and when the on the nsDNA-carpet of the saturation the flow the was = to solution containing ParBpSM or nM) and the concentration of parSpSM fragments indicated in the in on the of each of was to protein and and Each represents mean and standard errors of mean (SEM) of N repeated indicate the ParBpSM on the (B) The of the ratio for the four in the in (A) with ParBpSM concentration with different parSpSM in the (C) ATP hydrolysis is required for ParApSM release from the The experiments shown in (A) were repeated using bound to the nsDNA-carpet and ParBpSM nM) plus parSpSM fragment, 3R, or without parSpSM nM heptad in the was to protein and Each represents mean and of N (D) The of the ratio for the four of (C) Each represents the mean with to the of N repeated Figure data 1

DNA binding proteins play a number of key roles in cells, in processes including transcriptional regulation, recombination, genome rearrangements, and DNA replication, repair, and modification. Of particular interest are the interactions between transcription factors and their DNA binding sites, as they are an integral part of the transcriptional regulatory networks that control gene expression. Despite their importance, the DNA binding specificities of most DNA binding proteins remain unknown, as earlier technologies aimed at characterizing DNA-protein interactions have been time consuming and not highly scalable. We have developed a new DNA microarray-based technology, termed protein binding microarrays (PBMs), that allows rapid, high-throughput characterization of the in vitro DNA binding site sequence specificities of transcription factors in a single day. The resulting DNA binding site data can be used in a number of ways, including for the prediction of the genes regulated by a given transcription factor, annotation of transcription factor function, and functional annotation of the predicted target genes.

The activation of a specific DNA binding protein by neutron irradiation

Although DNA flexibility is known to play an important role in DNA–protein interactions, the importance of protein flexibility is less well understood. Here, we show that protein dynamics are important in DNA recognition using the well-characterized human papillomavirus (HPV) type 6 E2 protein as a model system. We have compared the DNA binding properties of the HPV 6 E2 DNA binding domain (DBD) and a mutant lacking two C-terminal leucine residues that form part of the hydrophobic core of the protein. Deletion of these residues results in increased specific and non-specific DNA binding and an overall decrease in DNA binding specificity. Using 15N NMR relaxation and hydrogen/deuterium exchange, we demonstrate that the mutation results in increased flexibility within the hydrophobic core and loop regions that orient the DNA binding helices. Stopped-flow kinetic studies indicate that increased flexibility alters DNA binding by increasing initial interactions with DNA but has little or no effect on the structural rearrangements that follow this step. Taken together these data demonstrate that subtle changes in protein dynamics have a major influence on protein–DNA interactions.

Polycombing the Genome: PcG, trxG, and Chromatin Silencing

We have recently described the appearance of a specific DNA-binding protein in nuclei from human cells exposed to ionizing radiation which was not detected in nuclear extracts from unperturbed cells (Singh, S. P., and Lavin, M. F. (1990) Mol. Cell. Biol. 10, 5279-5285). We report here a similar activity which is constitutively present in nuclei of both unirradiated and irradiated cells from patients with the human genetic disorder ataxia telangiectasia (A-T). Activity was present in unirradiated nuclear extracts from 3 A-T cell lines of different complementation groups, but was not detected or was present only at a low level in 4 controls. Active protein was detected in the cytoplasm of both cell types. Exposure to ionizing radiation did not change the amount of DNA binding activity in A-T nuclei but led to an increase in nuclei from 4 control cell lines. Purification of the binding activities from A-T nuclei and control cytoplasm was carried out by affinity chromatography, as described previously for control extracts (Teale, B., Singh, S. P., Khanna, K. K., Findik, D., and Lavin, M. F. (1992) J. Biol. Chem. 267, 10295-10301). Southwestern analysis and UV cross-linking confirmed the presence of a major DNA-binding species at 70 kDa in both cases with a minor binding activity at 47 kDa also evident. It was not possible to distinguish between the binding activities from A-T and control cells under different conditions, and phosphorylation was required for binding activity in both cases. Footprint analysis revealed that the same sequence was being recognized by the control and A-T proteins. The constitutive presence of a specific radiation-responsive DNA-binding protein in A-T cells may be indicative of a continuous state of stress in these cells.

MafA is a proto-oncoprotein and is critical for insulin gene expression in pancreatic β-cells. Maf proteins belong to the AP1 superfamily of basic region-leucine zipper (bZIP) transcription factors. Residues in the basic helix and an ancillary N-terminal domain, the Extended Homology Region (EHR), endow maf proteins with unique DNA binding properties: binding a 13 bp consensus site consisting of a core AP1 site (TGACTCA) flanked by TGC sequences and binding DNA stably as monomers. To further characterize maf DNA binding, we determined the structure of a MafA-DNA complex. MafA forms base-specific hydrogen bonds with the flanking G(-5)C(-4) and central C(0)/G(0) bases, but not with the core-TGA bases. However, in vitro binding studies utilizing a pulse-chase electrophoretic mobility shift assay protocol revealed that mutating either the core-TGA or flanking-TGC bases dramatically increases the binding off rate. Comparing the known maf structures, we propose that DNA binding specificity results from positioning the basic helix through unique phosphate contacts. The EHR does not contact DNA directly but stabilizes DNA binding by contacting the basic helix. Collectively, these results suggest a novel multistep DNA binding process involving a conformational change from contacting the core-TGA to contacting the flanking-TGC bases.

The pharmacological actions of morphine and morphine-like drugs such as heroin are mediated primarily through the μ opioid receptor. Previously a single strand DNA element of the mouse μ opioid receptor gene (Oprm1) proximal promoter was found to be important for regulating Oprm1 in neuronal cells. To identify proteins binding to the single strand DNA element as potential regulators for Oprm1, affinity column chromatography with the single strand DNA element was performed using neuroblastoma NS20Y cells followed by two-dimensional gel electrophoresis and MALDI-TOF mass spectrometry. We identified five poly(C)-binding proteins: heterogeneous nuclear ribonucleoprotein (hnRNP) K, α-complex proteins (αCP) αCP1, αCP2, αCP2-KL, and αCP3. Binding of these proteins to the single strand DNA element of Oprm1 was sequence-specific as confirmed by supershift assays. In cotransfection studies, hnRNP K, αCP1, αCP2, and αCP2-KL activated the Oprm1 promoter activity, whereas αCP3 acted as a repressor. Ectopic expression of hnRNP K, αCP1, αCP2, and αCP2-KL also led to activation of the endogenous Oprm1 transcripts, and αCP3 repressed endogenous Oprm1 transcripts. We demonstrate novel roles as transcriptional regulators in Oprm1 regulation for hnRNP K and αCP binding to the single strand DNA element.

Genomic analyses often involve scanning for potential transcription factor (TF) binding sites using models of the sequence specificity of DNA binding proteins. Many approaches have been developed to model and learn a protein's DNA-binding specificity, but these methods have not been systematically compared. Here we applied 26 such approaches to in vitro protein binding microarray data for 66 mouse TFs belonging to various families. For nine TFs, we also scored the resulting motif models on in vivo data, and found that the best in vitro-derived motifs performed similarly to motifs derived from the in vivo data. Our results indicate that simple models based on mononucleotide position weight matrices trained by the best methods perform similarly to more complex models for most TFs examined, but fall short in specific cases (<10% of the TFs examined here). In addition, the best-performing motifs typically have relatively low information content, consistent with widespread degeneracy in eukaryotic TF sequence preferences.

Members of the sterol regulatory element-binding protein (SREBP) family of transcription factors control cholesterol and lipid metabolism and play critical roles during adipocyte differentiation. The transcription factor SREBP1 is degraded by the ubiquitin-proteasome system following phosphorylation of Thr426 and Ser430 in its phosphodegron. We now demonstrate that the glycogen synthase kinase (GSK)-3beta-dependent phosphorylation of these residues in SREBP1 is enhanced in response to specific DNA binding. DNA binding enhances the direct interaction between the C-terminal domain of SREBP1 and GSK-3beta. Accordingly, we demonstrate that GSK-3beta is recruited to the promoters of SREBP target genes in vivo. As a result of the phosphorylation of Thr426 and Ser430, the ubiquitin ligase Fbw7 is recruited to SREBP molecules associated with target promoters. Using a reconstituted ubiquitination system, we demonstrate that Fbw7-mediated ubiquitination of SREBP1 is dependent on its DNA binding activity. Thus, DNA binding could provide a mechanistic link between the phosphorylation, ubiquitination, and degradation of active transcription factors.

Understanding the basis of the DNA-binding specificity of transcription factors (TFs) has been of long-standing interest. Despite extensive efforts to map millions of putative TF binding sequences, identifying the critical determinants for DNA binding specificity remains a major challenge. The coevolution of residues in proteins occurs due to a shared evolutionary history. However, it is unclear how coevolving residues in TFs contribute to DNA binding specificity. Here, we systematically collected publicly available data sets from multiple large-scale high-throughput TF-DNA interaction screening experiments for the major TF families with large numbers of TF members. These families included the Homeobox, HLH, bZIP_1, Ets, HMG_box, ZF-C4, and Zn_clus TFs. We detected TF subclass-determining sites (TSDSs) and showed that the TSDSs were more likely to coevolve with other TSDSs than with non-TSDSs, particularly for the Homeobox, HLH, Ets, bZIP_1, and HMG_box TF families. By in silico modeling, we showed that mutation of the highly coevolving residues could significantly reduce the stability of the TF-DNA complex. The distant residues from the DNA interface also contributed to TF-DNA binding activity. Overall, our study gave evidence that coevolved residues relate to transcriptional regulation and provided insights into the potential application of engineered DNA-binding domains and proteins. IMPORTANCE While unraveling DNA-binding specificity of TFs is the key to understanding the basis and molecular mechanism of gene expression regulation, identifying the critical determinants that contribute to DNA binding specificity remains a major challenge. In this study, we provided evidence showing that coevolving residues in TF domains contributed to DNA binding specificity. We demonstrated that the TSDSs were more likely to coevolve with other TSDSs than with non-TSDSs. Mutation of the coevolving residue pairs (CRPs) could significantly reduce the stability of THE TF-DNA complex, and even the distant residues from the DNA interface contribute to TF-DNA binding activity. Collectively, our study expands our knowledge of the interactions among coevolved residues in TFs, tertiary contacting, and functional importance in refined transcriptional regulation. Understanding the impact of coevolving residues in TFs will help understand the details of transcription of gene regulation and advance the application of engineered DNA-binding domains and protein.

The energetic contributions of individual DNA-contacting side chains to specific DNA recognition in the human papillomavirus 16 E2C-DNA complex is small (less than 1.0 kcal mol(-1)), independent of the physical and chemical nature of the interaction, and is strictly additive. The sum of the individual contributions differs 1.0 kcal mol(-1) from the binding energy of the wild-type protein. This difference corresponds to the contribution from the deformability of the DNA, known as "indirect readout." Thus, we can dissect the energetic contribution to DNA binding into 90% direct and 10% indirect readout components. The lack of high energy interactions indicates the absence of "hot spots," such as those found in protein-protein interfaces. These results are compatible with a highly dynamic and "wet" protein-DNA interface, yet highly specific and tight, where individual interactions are constantly being formed and broken.

DNase I footprinting of the apoA-II promoter using sterol regulatory element binding protein-2 [(SREBP-2 (1-458)] expressed in bacteria identified four protected regions, designated AIIAB (-64 to -48), AIICD (-178 to -154), AIIDE (-352 to -332) and AIIK (-760 to -743), which bind SREBP-2 and contain either palindromic or direct repeat motifs. Potassium permanganate and dimethyl sulfate interference experiments using the AIIAB region as probe showed that the nucleotides of a decameric palindromic repeat RTCAMVTGMY and two 5' T residues participate in DNA-protein interactions. SREBP-2 transactivated the intact (-911/+29) apoA-II promoter 1.7-fold and truncated apoA-II promoter segments which contain one, two or three SREBP-2 sites 11- to 17-fold in HepG2 cells. Transactivation of a promoter construct containing the binding site AIIAB and the apoA-II enhancer, which includes the binding site AIIK, was abolished by mutations in element AIIAB. An SREBP-2 mutant defective in DNA binding caused a dose-dependent repression of the apoA-II promoter activity. Repression was also caused by an SREBP-2 mutant which lacks the N-terminal activation domain (residues 1-93) but binds normally to its cognate sites. In contrast, a double SREBP-2 mutant which lacks both the DNA binding and the activation domains has no effect on the apoA-II promoter activity. Overall, the findings suggest that SREBP-2 can transactivate the apoA-II promoter by binding to multiple sites. Furthermore, the repression caused by the DNA binding deficient mutants results from squelching of positive activator(s) which appear to recognize the activation domain of SREBP-2.

Nuclear hormone receptors belong to a class of transcription factors that recognize specific DNA sequences either as monomers, homodimers, or heterodimers with the common partner retinoic X receptor. In vitro mutagenesis studies, as well as determination of the crystal structure of several complexes formed by the DNA-binding domain of receptors bound to their cognate response elements, have begun to explain the molecular basis for protein-DNA and protein-protein interactions essential for high-affinity and specific DNA binding by nuclear receptors. In this study, we have used the related orphan nuclear receptors, RORalpha and RevErbAalpha, to study the molecular determinants involved in the transition from monomeric to homodimeric modes of DNA binding by nuclear receptors. While both receptors bind DNA as monomers to a response element containing a core AGGTCA half-site preceded by a 5'-A/T-rich flanking sequence, RevErbAalpha also binds as a homodimer to an extended DR2 element. Gain-of-function experiments using point mutations and subdomain swaps between RORalpha and RevErbAalpha identify four amino acids within RevErbAalpha sufficient to confer RORalpha with the ability to form cooperative homodimer complexes on an extended DR2. This study reveals how the transition from monomer to homodimer DNA binding by members of the nuclear receptor superfamily could be achieved from relatively few amino acid substitutions.

TDP-43: A DNA and RNA binding protein with roles in neurodegenerative diseases