Expression Patterns, Genomic Conservation and Input Into Developmental Regulation of the GGDEF/EAL/HD-GYP Domain Proteins in Streptomyces

Actinobacteria constitute one of the largest phyla among bacteria and represent gram-positive bacteria with a high G+C content in their DNA. This bacterial group includes microorganisms exhibiting a wide spectrum of morphologies, from coccoid to fragmenting hyphal forms, as well as possessing highly variable physiological and metabolic properties. Furthermore, Actinobacteria members have adopted different lifestyles, and can be pathogens (e.g., Corynebacterium, Mycobacterium, Nocardia, Tropheryma, and Propionibacterium), soil inhabitants (Streptomyces), plant commensals (Leifsonia), or gastrointestinal commensals (Bifidobacterium). The divergence of Actinobacteria from other bacteria is ancient, making it impossible to identify the phylogenetically closest bacterial group to Actinobacteria. Genome sequence analysis has revolutionized every aspect of bacterial biology by enhancing the understanding of the genetics, physiology, and evolutionary development of bacteria. Various actinobacterial genomes have been sequenced, revealing a wide genomic heterogeneity probably as a reflection of their biodiversity. This review provides an account of the recent explosion of actinobacterial genomics data and an attempt to place this in a biological and evolutionary context.

Cyclic diguanylic acid (c-di-GMP) is a global second messenger controlling motility and adhesion in bacterial cells. Synthesis and degradation of c-di-GMP is catalyzed by diguanylate cyclases (DGC) and c-di-GMP-specific phosphodiesterases (PDE), respectively. Whereas the DGC activity has recently been assigned to the widespread GGDEF domain, the enzymatic activity responsible for c-di-GMP cleavage has been associated with proteins containing an EAL domain. Here we show biochemically that CC3396, a GGDEF-EAL composite protein from Caulobacter crescentus is a soluble PDE. The PDE activity, which rapidly converts c-di-GMP into the linear dinucleotide pGpG, is confined to the C-terminal EAL domain of CC3396, depends on the presence of Mg2+ ions, and is strongly inhibited by Ca2+ ions. Remarkably, the associated GGDEF domain, which contains an altered active site motif (GEDEF), lacks detectable DGC activity. Instead, this domain is able to bind GTP and in response activates the PDE activity in the neighboring EAL domain. PDE activation is specific for GTP (K(D) 4 microM) and operates by lowering the K(m) for c-di-GMP of the EAL domain to a physiologically significant level (420 nM). Mutational analysis suggested that the substrate-binding site (A-site) of the GGDEF domain is involved in the GTP-dependent regulatory function, arguing that a catalytically inactive GGDEF domain has retained the ability to bind GTP and in response can activate the neighboring EAL domain. Based on this we propose that the c-di-GMP-specific PDE activity is confined to the EAL domain, that GGDEF domains can either catalyze the formation of c-di-GMP or can serve as regulatory domains, and that c-di-GMP-specific phosphodiesterase activity is coupled to the cellular GTP level in bacteria.

Introduction In all domains of life, nucleotide-based second messengers allow a rapid integration of external and internal signals into fine-tuned regulatory pathways that control cellular responses to changing conditions. As a unifying theme, a basic second messenger control module consists of two distinct enzymes for synthesis and degradation of the second messenger and a nucleotide sensor that, upon ligand binding, interacts with a target to produce a cellular output (Hengge, 2009). 3′, 5′-cyclic diguanylic acid (c-di-GMP), which is not produced in archaea or eukaryotes, was first discovered as an allosteric effector of cellulose synthase in Gluconacetobacter xylinus and is now recognized as one of the most important and widespread second messengers in bacteria. c-di-GMP is synthesized from two molecules of GTP by diguanylate cyclases (DGCs), which are characterized by active site GGDEF motifs (A-site) (Paul et al., 2004; Chan et al., 2004). The majority of active DGCs also carry a so-called inhibitory or I-site motif, RxxD, which is involved in feedback inhibition (Christen et al., 2006; Schirmer and Jenal, 2009). Specific phosphodiesterases (PDEs), which harbor EAL or HD-GYP domains, degrade the cyclic dinucleotide (Schmidt et al., 2005; Christen et al., 2005; Ryan et al., 2006). The enzymatically active domains involved in c-di-GMP turnover are often associated with diverse sensory domains, thus enabling cells to adjust second messenger levels in response to different environmental stimuli (Hengge, 2009). The binding of c-di-GMP to effector proteins impacts diverse processes such as adhesion, virulence, motility, and biofilm formation in unicellular, flagellated bacteria (Römling et al., 2013). The known c-di-GMP-binding motifs of these proteins are limited but include degenerate GGDEF domain proteins carrying I-site motifs (Duerig et al., 2009; Lee et al., 2007b; Petters et al., 2012), inactive EAL domain receptors (Navarro et al., 2009; Qi et al., 2011; Newell et al., 2009), and PilZ domain-containing proteins (Amikam and Galperin, 2006). Transcription factors that sense c-di-GMP lack these common c-di-GMP-binding motifs and thus must be identified experimentally. The sparse list of known c-di-GMP-responsive transcriptional regulators includes the TetR-like activator LtmA from Mycobacterium smegmatis (Li and He, 2012), the CRP-FNR-like transcription factor Clp from Xanthomonas (Chin et al., 2010; Leduc and Roberts, 2009), Bcam1349 from Burkholderia (Fazli et al., 2011), the NtrC-type protein FleQ from Pseudomonas aeruginosa (Baraquet and Harwood, 2013), and VpsR from Vibrio cholerae (Srivastava et al., 2011). The only c-di-GMP-responsive transcription factor for which structural information is available and hence c-di-GMP binding is understood is VpsT, which is a member of the well-studied FixJ-LuxR-CsgD family of response regulators. The VpsT structure revealed a characteristic response regulator fold and a W(F/L/M)(T/S)R c-di-GMP-binding motif (Krasteva et al., 2010). Notably, in all known structures of c-di-GMP-binding effector proteins or enzymes, the c-di-GMP is bound either as a monomer or intercalated dimer. Biophysical studies suggest the possibility of higher order oligomeric forms of c-di-GMP, but they have yet to be observed in any biological context (Gentner et al., 2012). While the roles played by c-di-GMP in controlling cellular processes in unicellular bacteria are becoming clear, the function(s) of c-di-GMP in multicellular, nonmotile bacteria such as Streptomyces are unknown. The complex Streptomyces life cycle involves two distinct filamentous cell forms: the growing or vegetative hyphae and the reproductive or aerial hyphae, which differentiate into exospores for dispersion through a massive synchronous septation event (Flärdh and Buttner, 2009). In the model species Streptomyces venezuelae, there are three GGDEF proteins, two proteins with HD-GYP domains, and five proteins containing both a GGDEF and an EAL domain (Figure 1A). Altered expression of the GGDEF proteins, CdgA and CdgB, and deletions of the EAL proteins, RmdA and RmdB, have a significant impact on Streptomyces growth progression, suggesting that c-di-GMP plays a role in controlling developmental processes in multicellular bacteria (den Hengst et al., 2010; Tran et al., 2011; Hull et al., 2012). Interestingly, cdgA and cdgB have recently been identified as direct regulatory targets of the developmental master regulator BldD (den Hengst et al., 2010; Tran et al., 2011). Mutations in the bld loci block the formation of aerial hyphae, resulting in a “bald” phenotype, and also affect the production of antibiotics (McCormick and Flärdh, 2012). BldD sits at the top of the regulatory cascade controlling development, serving to repress expression of sporulation genes during vegetative growth (den Hengst et al., 2010). In Streptomyces coelicolor, BldD controls the expression of at least 167 genes, including 42 genes (∼25% of the regulon) that encode regulatory proteins (Elliot et al., 2001; den Hengst et al., 2010). Among these BldD targets are many genes known to play critical roles in Streptomyces development, including other bld regulators (e.g., bldA, bldC, bldH/adpA, bldM, and bldN), several whi (white) regulators required for the differentiation of aerial hyphae into spores (e.g., whiG and whiB), and genes encoding critical components of the cell division and chromosome segregation machineries such as FtsZ, SsgA, SsgB, and the DNA translocase SffA (den Hengst et al., 2010; McCormick, 2009). How BldD activity is regulated, however, has been unknown. Here we show that BldD is a c-di-GMP-binding effector protein, thus revealing a link between c-di-GMP signaling and the development of multicellular bacteria. Specifically, structural and biochemical analyses show that the second messenger c-di-GMP activates BldD DNA binding by driving a unique form of dimerization that is mediated by a tetrameric form of c-di-GMP. The c-di-GMP tetramer performs its oligomerization function by adjoining two BldD C-terminal domain (CTD) protomers, the polypeptide chains of which are separated by ∼10 Å. BldD recognizes the c-di-GMP tetramer using a bipartite RXD-X8-RXXD c-di-GMP interaction signature sequence from each subunit. Thus, tetrameric c-di-GMP acts as a small-molecule dimerizing agent that controls the DNA-binding activity of BldD, leading to repression of the BldD regulon of sporulation genes during vegetative growth, thereby controlling the hypha-to-spore transition in multicellular bacteria. Results c-di-GMP Controls Developmental Program Progression in Streptomyces venezuelae To gain insight into the cellular processes controlled by c-di-GMP in streptomycetes, we overexpressed either the active DGC CdgB from S. coelicolor (Tran et al., 2011) or the active PDE YhjH from E. coli (Pesavento et al., 2008). Strikingly, overexpression of both CdgB and YhjH blocked the generation of aerial mycelium by S. venezuelae (Figure 1B). However, scanning electron micrographs (SEMs) revealed that, whereas overexpression of CdgB blocked development, resulting in a classical bald phenotype, overexpression of the PDE YhjH in fact promoted sporulation, but the colonies appeared bald to the naked eye because aerial mycelium formation had been bypassed (Figure 1C). As judged by heat resistance, the spores made by the YhjH overexpression strain were as robust as those of the wild-type (WT) (Figure S1A available online). Moreover, overexpression of catalytically inactive versions of YhjH or CdgB had no effect on S. venezuelae development (Figure S1B). These data suggest that intracellular levels of c-di-GMP influence the timing of development. In particular, they suggest that increased c-di-GMP levels delay differentiation, arresting the colonies in the vegetative growth stage, whereas decreased levels of the second messenger accelerate development, favoring sporulation. BldD Is a c-di-GMP Effector Protein S. venezuelae has no PilZ domain-containing proteins, and no putative c-di-GMP-binding effector proteins have so far been identified in the Streptomyces genus. Thus, to address the mechanism by which S. venezuelae senses c-di-GMP to control sporulation, we sought to identify c-di-GMP effector proteins involved in development. To selectively enrich putative c-di-GMP-binding proteins from S. venezuelae cell extracts, we performed an affinity pull-down assay using a c-di-GMP capture compound (Nesper et al., 2012). Captured proteins were identified by tryptic mass spectrometry fingerprinting. Remarkably, the developmental master regulator BldD was repeatedly recovered in our c-di-GMP-based capture compound experiments. BldD is an 18 kDa DNA-binding protein (Elliot and Leskiw, 1999) consisting of two distinct domains connected by a flexible linker (Figure 2A). The N-terminal domain is the DNA-binding domain (DBD) and has a xenobiotic response element (XRE) helix-turn-helix (HTH) DNA-binding motif (Kim et al., 2006). The BldD CTD harbors a largely helical fold with no known function (Kim et al., 2014). To probe the interaction between BldD and c-di-GMP further and to identify the c-di-GMP-binding domain, we used differential radial capillary action of ligand assays (DRaCALA) (Roelofs et al., 2011). DRaCALA allows the visualization of protein-bound radiolabeled ligand as a concentrated spot after the application of the protein-ligand mixture onto nitrocellulose. Using this assay, we confirmed that full-length (FL) BldD (expressed as an N-terminally His6-tagged protein) from both S. venezuelae (Figures 2B and S2B) and S. coelicolor (data not shown) bind 32P-labeled c-di-GMP. Importantly, the DRaCALA assays demonstrated that the previously uncharacterized CTD of BldD functions as the c-di-GMP-binding domain (Figure 2B). Further, excess unlabeled c-di-GMP, but not GTP, competed with the labeled c-di-GMP for binding to FL BldD and to BldD-CTD. Thus, these data reveal that the CTD of the key developmental regulator BldD is a c-di-GMP-binding domain. Cyclic di-GMP Enhances Binding of BldD to Its Target Promoters In Vitro and In Vivo Using global chromatin immunoprecipitation-microarray analysis (ChIP-chip), we previously identified the complete BldD regulon in S. coelicolor, showing that it encompasses ∼167 transcription units (den Hengst et al., 2010). Through MEME-based sequence analysis of all the promoter regions directly targeted by BldD, we defined a 13 bp pseudo-palindromic sequence, 5′-TNAC(N)5GTNA-3′, designated the BldD box, which functions as a specific binding sequence for BldD (den Hengst et al., 2010). Sequence analysis showed that the BldD box was conserved between S. coelicolor and S. venezuelae for most key BldD target promoters (den Hengst et al., 2010). Further, BldD from S. coelicolor and S. venezuelae contain an identical DBD and differ by only two residues in the DBD-CTD linker and five residues in the CTD, suggesting that BldD function is broadly conserved between the two species. Having shown that BldD binds c-di-GMP, we tested the effect of c-di-GMP on BldD DNA binding. Radiolabeled S. venezuelae DNA fragments encompassing the promoter regions of two well-characterized BldD target genes, bldM and whiG, including the bioinformatically identified BldD box (Figure 2C), were used as target DNAs in electrophoretic mobility shift assays (EMSAs) (Figure 2D). The fixed concentration of BldD used in these assays (0.6 μM) was insufficient to elicit a DNA mobility shift. However, the addition of increasing concentrations of c-di-GMP (0.25–1.75 μM) strongly induced BldD binding to both the tested promoter regions (Figure 2D). To confirm and extend these results into cells, we manipulated the levels of c-di-GMP in S. venezuelae and monitored the effect on BldD binding to the bldM and whiG promoters in vivo using ChIP-sequencing (ChIP-seq). The degree of BldD binding was assayed at a single time point in WT S. venezuelae and the WT overexpressing either the DGC CdgB or the PDE YhjH (the strains whose phenotypes are described above). Consistent with the in vitro EMSA data, overexpression of the DGC enhanced ChIP-seq peak height at the BldD target promoters relative to the WT control (Figure 2E). Conversely, overexpression of the PDE lowered ChIP-seq peak heights at BldD target promoters relative to the WT (Figure 2E). These data demonstrate that c-di-GMP enhances the binding of BldD to the BldD box, stimulating BldD-mediated repression of its target regulon. Thus, it is not BldD, but a BldD-(c-di-GMP) complex, that serves to turn off sporulation genes during vegetative growth. S. venezuelae bldD Mutants Show an Accelerated Sporulation Phenotype that Bypasses Aerial Mycelium Formation The opposing effects of the overexpression of the DGC CdgB and the PDE YhjH suggested that high levels of c-di-GMP retard sporulation and low levels of c-di-GMP accelerate sporulation. Because the BldD-(c-di-GMP) complex serves to keep sporulation genes shut off during vegetative growth, loss of BldD should have a similar effect on Streptomyces development as depletion of c-di-GMP levels. To test this hypothesis, we deleted bldD from the S. venezuelae chromosome. Strikingly, the bldD null mutant formed small colonies lacking aerial hyphae, but—when examined by SEM—even young colonies of the bldD mutant were found to contain spore chains embedded in an excess of extracellular matrix (Figure 3A). Heat resistance tests showed the bldD mutant spores were mildly defective (Figure S1A). By contrast, equivalent young colonies of the WT that were grown and imaged in parallel had not yet developed aerial hyphae or spores (Figure 3B). Thus, loss of BldD mimics the effects of overexpressing the PDE YhjH (compare Figures 1C and 3A). In addition, overexpression of CdgB had no effect on the phenotype of the bldD mutant (Figure S2A), further supporting the idea that c-di-GMP signals through BldD to control the hypha-to-spore transition. Crystal Structures of BldD CTD-(c-di-GMP) Complexes: Small-Molecule-Mediated Protein Dimerization To elucidate the molecular basis for c-di-GMP recognition and binding by the BldD CTD and to gain insight into how this interaction may elicit developmental signaling, we determined structures of the S. venezuelae BldD CTD (residues 80–166) and S. coelicolor BldD CTD (residues 80–167) in complex with c-di-GMP. Three S. venezuelae BldD C-domain-(c-di-GMP) structures and one S. coelicolor BldD C-domain-(c-di-GMP) structure were determined to resolutions of 1.95 Å, 2.33 Å, 1.75 Å, and 2.25 Å, respectively (Figure 4) (see Extended Experimental Procedures and Tables S1 and S2). The BldD CTD structures are composed of two (β-α-α) repeats followed by a short C-terminal helix and are similar to the apo form studied by NMR (Kim et al., 2014) (Figure 4). Database searches revealed that the BldD CTD harbors a new fold, but reduced stringency revealed that it shows limited structural similarity with winged HTH proteins, in particular, the winged HTH domain of eukaryotic heat shock factor 1 (HSF1) (Littlefield and Nelson, 1999). The BldD CTD and HSF1 bind c-di-GMP and DNA, respectively, but the motifs they employ to interact with their nucleotide ligands are completely different (Figure S3). The BldD CTD uses a previously unseen mode of c-di-GMP binding in which two noninteracting CTDs are glued together by a c-di-GMP tetramer composed of two interlocked c-di-GMP dimers (Figure 4A). Thus, in the BldD CTD-(c-di-GMP) complex, c-di-GMP functions as a macromolecular dimerizer. Indeed, the closest approach of any Cα atoms of the two tethered CTDs is ∼10 Å. The BldD CTD interacts with c-di-GMP using two contiguous surface motifs, herein called motif 1 and motif 2, which are located between the two β-α-α repeats. Motif 1 is composed of residues 114–116 (RGD) and motif 2 is composed of residues 125–128 (RQDD) (Figure 5A). These motifs are located on a solvent exposed region at one end of each CTD protomer and not within a pocket or cavity (Figures 4A and 5A). The combined motifs from two CTD subunits provide ideal shape and electrostatic complementarity for encasing the unusual cage-like c-di-GMP tetramer (Figures 4A and 4B). Strikingly, although the BldD CTD is overall acidic (pI∼5.0), the c-di-GMP-binding surface between two BldD protomers is electropositive (Figure 4B). In addition to shape and charge complementary, contacts from the arginine and aspartic acid residues within motifs 1 and 2 provide specificity for recognition of guanine cyclic nucleotides; the multiple interactions effectively exclude binding to adenine cyclic nucleotides. Specifically, motif 2 from each CTD protomer combine to mediate contacts to one intercalated c-di-GMP dimer, Asp116 of motif 1 from each protomer combine to mediate contacts to the other c-di-GMP dimer, while Arg114 sits centrally and anchors both dimers (Figures 5A and 5B). The guanine bases wedged between the two motif 2 regions are specified by contacts from residues Arg125 and Asp128 (Figures 5A and 5B). Residue Asp128 makes two hydrogen bonds to the N1 and the exocyclic N2 atoms of the guanine bases on each end (top and bottom layers) of the intercalated dimer, while Arg125 flanks the guanines in the center (middle layers) of the dimer and makes hydrogen bonds to the guanine O6 and N7 atoms (Figures 5B–5D). The O6 moiety is also specified by the backbone nitrogen of Arg125. Notably, the stacking interactions between the Arg125 side chains are the only direct contacts between the two CTD protomers but are clearly not sufficient to promote BldD dimerization. Arg114 is the only CTD residue that makes contacts to both intercalated c-di-GMP dimers. Arg114 hydrogen bonds to the guanines contacted by Asp128, as well as the O6 atoms of the adjacent guanines of the other c-di-GMP dimer (Figures 5A and 5B). Thus, Arg114 plays a key role in the recognition and stabilization of this unique c-di-GMP tetramer. The c-di-GMP dimer bound between the motif 1 regions has fewer contacts and is more exposed (Figure 5A). In addition to contacts from Arg114, Asp116 of motif 1 hydrogen bonds to the guanine N1 and N2 atoms in a manner analogous to the contacts from Asp128 of motif 2 (Figure 5). Notably, the specific hydrogen bonds from motif 1 and 2 residues to guanine exocyclic O6 and N2 atoms dictate that BldD binds to c-di-GMP but not to c-di-AMP, which is missing an exocyclic atom at the 2 position and harbors a hydrogen bond donor rather than an acceptor at the exocyclic 6 position. BldD motifs 1 (RXD) and 2 (RXXD), although similar in sequence to the inhibitory I-site (RXXD), which is involved in product inhibition feedback control of DGC activity (Christen et al., 2006), are structurally different. Further, mutagenesis of either BldD motif 1 or motif 2 abolishes c-di-GMP binding in DRaCALA assays (Figure S2B), confirming that, unlike I-site c-di-GMP binding, both motifs 1 and 2 in BldD are required to construct the complete binding site for the tetrameric c-di-GMP complex. This dual signature sequence is unlike any previously characterized c-di-GMP-binding motif. Moreover, specific binding of the c-di-GMP tetramer requires encasement by two such bipartite motifs from precisely oriented BldD protomers. While the arginine and aspartic acid residues in BldD motifs 1 and 2 dictate the c-di-GMP-binding arrangement and read the guanine bases, contacts to the c-di-GMP phosphate groups are provided by Lys84 from β1 and Arg130 from α3. Further, Ile110 makes van der Waals interactions and residues Asn118 and Ser123 hydrogen bond with the guanine bases on the top and bottom layers of the c-di-GMP tetramer (Figure 5A). The combination of the optimally positioned bipartite signature sequences from two BldD protomers exquisitely templates binding of the specific and unusual c-di-GMP tetramer structure. The BldD-(c-di-GMP) Structure Reveals a Tetrameric Form of the c-di-GMP Second Messenger c-di-GMP is monomeric in solution at physiological concentrations (Gentner et al., 2012). However, intercalated c-di-GMP dimers have been observed in crystal structures of the nucleotide alone and in complexes with effector proteins. Higher order c-di-GMP structures such as tetramers and octamers have thus far only been inferred from NMR and spectroscopic studies and require very high c-di-GMP concentrations (up to 30 mM) and monovalent cations (Zhang et al., 2006). These higher order structures are characterized by G-quartet interactions with a centrally bound potassium ion coordinated by four guanines. There are minimal base contacts and no base stacking interactions in these structures (Figure S4A) (Zhang et al., 2006; Gentner et al., 2012). By sharp contrast, the BldD-bound tetrameric c-di-GMP is a tightly packed structure that is not secured by ions. Rather, the c-di-GMP molecules are closely spaced and optimally positioned for interbase pairing, leading to the formation of a multistranded, base-stacked structure with top, middle, and bottom layers (Figures 5D and S4A). There are 12 hydrogen bonds between the two intercalated dimers within the c-di-GMP tetramer, including contacts between the N3 atoms and exocyclic NH2 amides of an adjacent base (Figures 5C and S4B). Such contacts could not be formed with c-di-AMP due to its lack of an exocyclic NH2 atom. Therefore, in addition to contacts from motifs 1 and 2, guanine-guanine base hydrogen bonds serve to specify c-di-GMP tetramer binding to BldD. Notably, formation of the c-di-GMP tetramer buries 24% of the total surface area (buried surface area [BSA]) of the c-di-GMP molecules (Figure S4B). By comparison, in most protein oligomers the BSA between protomers is ∼15% (Wang et al., 2009). Finally, the interface between the intercalated c-di-GMP dimers that forms the tetramer is remarkably complementary in shape (Figure S4B). Thus, the combination of multiple contacts between the c-di-GMP moieties along with its extensive BSA and molecular shape complementarity lead to the creation of a compact and highly specific c-di-GMP tetramer. However, BldD is necessary to stabilize this tetramer and template its formation. c-di-GMP Induces Dimerization of the BldD CTD in Solution The BldD CTD-(c-di-GMP) crystal structures reveal that c-di-GMP acts as a “dimerizer” to link two CTD protomers. To examine the effect of c-di-GMP on the oligomeric state of the BldD CTD in solution, we carried out chemical crosslinking and size exclusion chromatography (SEC) studies. Chemical crosslinking experiments were performed using disuccinimidyl suberate (DSS), which contains amine-reactive N-hydroxysuccinimide esters at both ends of an 11 Å spacer arm. This reagent should therefore be able to crosslink even the distantly anchored CTD protomers observed in our CTD-(c-di-GMP) structures (Figure 4A). In the absence of DSS, the BldD DBD and CTD migrate on SDS-PAGE gels as single bands with the expected monomeric molecular weights of 11 and 12 kDa, respectively (Figure 6A). Upon incubation with DSS, the BldD DBD forms a covalent dimer of ∼21 kDa, consistent with previous biochemical and structural analyses of this domain (Kim et al., 2006; Lee et al., 2007a), and oligomerization is unaffected by the addition of c-di-GMP (Figure 6A). By contrast, the BldD CTD remains monomeric after DSS addition. However, in the presence of c-di-GMP, addition of DSS results in the clear formation of covalent CTD dimers (Figure 6A). To examine c-di-GMP-induced CTD oligomerization further, we performed SEC analyses. As part of this study, we mutagenized the RGD (motif 1)-X8-RQDD (motif 2) c-di-GMP-binding signature of the BldD CTD. To retain the charge of these surface residues, we made the charge-swapped DGR-X8-DQDR CTD mutant, changing the key Arg residues within the two motifs to Asp, and vice versa. DRaCALA assays showed that this mutant CTD was unable to bind c-di-GMP (Figure S2B). SEC experiments in the presence of 3 μM c-di-GMP showed that the WT CTD forms a dimer, while the mutant CTD is monomeric (Figure 6B). Hence, the crosslinking and SEC data support our structural finding that c-di-GMP is required for BldD CTD dimerization. Further, we constructed a bldD mutant allele encoding a protein solely defective in c-di-GMP binding (carrying the DGR-X8-DQDR mutation) and found that it had no ability to complement a bldD mutant (Figure S2A), confirming that the major in vivo functions of BldD are indeed mediated by its binding to c-di-GMP. Affinity, Stoichiometry, and Specificity of c-di-GMP for BldD CTD Our structures reveal a unique interaction between the BldD CTD and a c-di-GMP tetramer. Remarkably, every guanine in this complex is read specifically by either BldD arginines, aspartic acids, and/or other guanine bases, suggesting that BldD binds only c-di-GMP, and not other cyclic nucleotides. To test this hypothesis and further probe the c-di-GMP-binding affinity of BldD, we performed fluorescence polarization (FP) experiments. These studies were performed with BldD CTD that was expressed and purified from Sf9 insect cells to ensure no c-di-GMP was present, as eukaryotes do not produce c-di-GMP (Extended Experimental Procedures). The BldD-(c-di-GMP) structures show that one ribose of each c-di-GMP bound to BldD must be unmodified to permit formation of the BldD-(c-di-GMP) complex (Figure S5A). Hence, for these studies we used the fluoresceinated probe, 2′-Fluo-AHC-c-di-GMP, which harbors the fluorescein dye on only one ribose (Figure S5A). These studies revealed BldD CTD bound to 2′-Fluo-AHC-c-di-GMP, with an apparent Kd of 2.5 μM ± 0.6 (Figure S5B). Consistent with our DRaCALA assays (Figure S2B), the DGR-X8-DQDR CTD mutant failed to bind 2′-Fluo-AHC-c-di-GMP in FP assays (Figure S5B). WT BldD CTD showed no binding to the identically fluoresceinated c-di-AMP tagged molecule, 2′-Fluo-AHC-c-di-AMP (Figure S5B). Next, to ascertain the stoichiometry of c-di-GMP binding in solution, we used an FP-based binding assay. The resulting data (Figure S5C) show a linear increase in fluorescence polarization until saturation of the binding sites. A single inflection point can be fitted at a BldD monomer concentration of ∼12 μM, which equates to a stoichiometry of four c-di-GMP molecules per CTD dimer. However, careful inspection of the data reveals another potential inflection point at a BldD monomer concentration of 6 μM, which would be consistent with an initial binding event of two c-di-GMP molecules per CTD dimer. This putative initial binding event is also apparent in the equilibrium-binding isotherm yielding an apparent Kd of ∼1.7 μM (Figure S5B). Two-step binding is not inconsistent with the structure, but the nearly identical affinities observed for each binding event suggest positive cooperativity. Overall, these data demonstrate unequivocally that c-di-GMP binds the CTD with a stoichiometry of four c-di-GMP molecules per CTD dimer, concordant with our structures. These studies also demonstrate that the BldD CTD binds specifically and with high affinity to c-di-GMP but not c-di-AMP and that both motifs 1 and 2 are essential for this interaction. The Mechanism for (c-di-GMP)-Activated DNA Binding by BldD Dimeric BldD binds pseudo-palindromic DNA sites that contain a 5′-TNAC(N)5GTNA-3′ consensus (den Hengst et al., 2010). Our data and those of others (Lee et al., 2007a) show that the BldD DBD alone can dimerize at higher concentrations (≥10 μM). Consistent with these findings, the crystal structure of the S. coelicolor BldD DBD revealed a dimer with a small contact interface (Kim et al., 2006). We obtained additional views of the BldD DBD by solving the S. venezuelae BldD DBD structure to 2.80 Å resolution. This structure contained two DBD dimers in the asymmetric unit. Comparison of these dimers with the dimer from the S. coelicolor DBD structure showed that, although hydrophobic residues within the C-terminal regions of the DBDs make contacts between the subunits in each case, all three dimers take distinct conformations (Figure S6A). Moreover, there is less than 300 Å2 BSA per subunit in this dimer, which is far less than the 1000 Å2 BSA per subunit typically observed for biologically relevant dimers (Krissinel and Henrick, 2007). These data indicate that the BldD DBD is unlikely to form a stable DNA-binding active dimer at physiologically relevant concentrations. In addition, although the BldD DBD resembles the equivalent DBD of λ repressor, previous modeling studies suggested that the BldD DBDs would not interact specifically with DNA if BldD employed a DNA-binding mechanism similar to that utilized by λ repressor (Kim et al., 2006). Thus, it has been unclear how BldD binds cognate DNA. Our finding that c-di-GMP binding leads to the formation of a c-di-GMP bridged CTD dimer provides the missing link to this puzzle. However, how c-di-GMP binding to the CTDs is signaled to the DBDs to bring about DNA binding remained unclear. To deduce the mechanism by which c-di-GMP activates BldD to bind DNA, we determined the structure of a BldD-(c-di-GMP)-21-mer DNA complex to 4.5 Å resolution (Extended Experimental Procedures; Figure S6B). While the low resolution of the structure precludes a detailed analysis, the electron density maps show the overall arrangement of the domains and how the DBDs dock onto the DNA. Critically, the structure reveals that BldD binding to cognate DNA is more similar to the DNA binding mode of the XRE protein SinR (Lewis et al., 1998; Newman et al., 2013) than to that of the λ repressor, as the two BldD DBDs are juxtaposed when BldD is bound to DNA (Figures 7A and 7B). The DBD-DBD interacting surfaces observed in the BldD-DNA complex correspond to the hydrophobic regions near the DBD C terminus that interact in the apo DBD structures. However, the interfaces are yet again different, supporting the notion that DBD dimerization is weak and malleable. Such malleability is critical to allow the dimeric BldD HTH elements to bind the DNA, which is bent by ∼30°. The BldD DBDs are tethered to the CTD via a linker that was previously shown to be highly flexible (Kim et al., 2006; Lee et al., 2007a). Not surprisingly, this linker (PGTTPGGAAEPPP; residues 71–84) is disordered in the BldD-(c-di-GMP)-21-mer structure. Its flexibility is underscored by the different orientation of the two CTDs in the structure relative to the DBD-DNA complex (Figure 7A). The CTD subunits in the dimer make different interactions that help anchor them in the BldD-(c-di-GMP)-DNA crystal; one of the CTDs interacts with a hydrophobic patch on its cognate DBD, while the other CTD makes crystal contacts with a symmetry mate (Figures 7A and S6C). However, the CTDs in the FL BldD-(c-di-GMP)-DNA structure are dimerized in a manner identical to that observed in our CTD-(c-di-GMP) structures (Figures 4D, 5A, 7A, and S6C). Additional evidence that the linker region between the DBD and CTD is flexible was provided by proteolysis experiments employing Endoproteinase Glu-C, which cleaves exposed peptide bonds on the carboxyl side of glutamic or aspartic acid residues. Hence, if the BldD linker region is unstructured, Glu-C should selectively cleave after BldD residue Glu80. Glu-C proteolysis experiments were carried out on the FL BldD protein, the FL protein with the 21-mer DNA present, and the FL protein in the presence of both c-di-GMP and the 21-mer DNA. In all three cases BldD was readily cleaved into two bands, corresponding to the DBD and the CTD (Figure S6D). Thus, neither the presence of DNA, nor of c-di-GMP and DNA, protected BldD from proteolysis, indicating the linker is exposed even in the presence of cognate ligands. Therefore, the combined data suggest a molecular model for c-di-GMP activation of BldD DNA binding in which binding of c-di-GMP leads to the formation of a c-di-GMP-linked BldD CTD dimer (Figure 7C). Such CTD dimerization effectively brings the two DBDs into proximity, thereby increasing their local concentration to allow germane DBD dimerization on cognate DNA (Figure 7C). The inherent flexibility afforded by the DBD-CTD linker allows the DBDs to adjust for optimal binding to multiple pseudo-palindromic BldD DNA boxes. Discussion The role of c-di-GMP has been studied extensively in unicellular Gram-negative bacteria, in which most c-di-GMP-dependent signaling pathways control the transition from a planktonic, motile lifestyle to a surface-associated, sessile lifestyle (“stick or swim”). Here we show that the activity of the Streptomyces master regulator BldD is also controlled by c-di-GMP, thus bringing the regulatory role of this key second messenger into a new physiological arena, that of differentiation in Gram-positive multicellular bacteria. Our studies indicate that c-di-GMP binding to BldD controls the developmental switch between vegetative growth and sporulation. Specifically, we demonstrate that c-di-GMP binding to BldD activates its DNA-binding activity, which results in repression of sporulation genes during vegetative growth. Consistent with this, genetic studies revealed that bldD null mutants sporulate precociously, mimicking the effect of overexpressing a c-di-GMP phosphodiesterase. Thus, c-di-GMP signals through BldD to control the hypha-to-spore developmental transition in Streptomyces. Few c-di-GMP effector-binding motifs have been identified to date. These include I-site motifs (Duerig et al., 2009; Lee et al., 2007b; Petters et al., 2012), inactive EAL domains (Navarro et al., 2009; Qi et al., 2011; Newell et al., 2009), and PilZ domains (Amikam and Galperin, 2006). Structures have shown that c-di-GMP interacts with these motifs as either a monomer or intercalated dimer. The BldD protein does not contain any previously characterized c-di-GMP effector-binding motifs. Thus, to elucidate the mechanism by which c-di-GMP acts as a switch to turn on the DNA-binding activity of BldD, we determined several structures of the BldD CTD complexed to c-di-GMP as well as a 4.5 Å structure of the BldD-(c-di-GMP)-DNA complex. These structures revealed that BldD interacts with c-di-GMP using a heretofore unseen c-di-GMP binding mode involving a unique c-di-GMP-binding signature sequence consisting of two proximal arginine and aspartic acid containing motifs, motif 1 (RXD) and motif 2 (RXXD), separated by eight residues. Remarkably, in this binding mode, a tetrameric form of the c-di-GMP functions as a small-molecule dimerizer to adjoin two noninteracting BldD protomers. Notably, the identical CTD dimer-(c-di-GMP) tetramer structure was seen in multiple crystal forms. Finally, binding studies confirmed that the BldD CTD binds c-di-GMP with a stoichiometry of four c-di-GMP molecules to one BldD CTD dimer. The c-di-GMP tetramer revealed in these structures represents a previously unknown form of this nucleotide second messenger. Indeed, we do not know of any other example in which a signaling molecule can assume different oligomeric states to effect its function. BldD is present throughout the sporulating actinomycetes (den Hengst et al., 2010), including, for example, nitrogen-fixing Frankia that live in symbiosis within the root nodules of alder trees, and members of the marine genus Salinospora, which have recently emerged as an important source of antibiotics and other medically significant compounds. Outside of the genus Streptomyces, the only actinomycete in which BldD has been investigated is Saccharopolyspora erythraea, where BldD directly controls expression of the biosynthetic cluster of the clinically important antibiotic erythromycin (Chng et al., 2008). Homologs from across the sporulating actinomycetes share 77%–99% sequence identity with S. venezuelae BldD. The main region of conservation between these proteins is the N-terminal DNA-binding domain, which shares 95%–100% identity. Although our BldD-(c-di-GMP)-DNA structure is too low resolution to ascribe specific protein-DNA contacts, it reveals the location of the HTH motif and residues that likely contact the DNA. Notably, these amino acids are the most conserved among BldD homologs (essentially 100%; Figure S7). By contrast, the CTD regions of BldD proteins are less well conserved (as low as 48% identity). Hence, it is striking, given this low conservation, that the residues that interact with c-di-GMP are strictly conserved (Figure S7). The only exception is Lys84, which contacts c-di-GMP phosphate groups. However, in all BldD homologs this residue is either a lysine or arginine and thus able to make the same electrostatic interaction. Of particular note, residues R114, D116, R125, and D128 (from motifs 1 and 2), which mediate essential specifying contacts with c-di-GMP, are conserved in all homologs (Figure S7). Further, all of the actinomycetes that encode an ortholog of BldD also encode GGDEF domain-containing DGCs. These combined findings indicate that BldD-(c-di-GMP) is likely to control key developmental processes throughout the sporulating actinomycetes, using tetrameric c-di-GMP as a second messenger. Experimental Procedures For a full explanation of the experimental protocols, see Extended Experimental Procedures in Supplemental Information. Bacterial Strains, Plasmids, and bldD Null Mutant Construction Strains and plasmids used are shown in Table S3, and oligonucleotides used are shown in Table S4. Plasmids were constructed as described in Extended Experimental Procedures. A bldD null mutant (SV77) was constructed by Redirect PCR targeting, and the bldD::apr mutant allele was moved into a new WT background by generalized transduction using the S. venezuelae-specific phage SV1. Transduction of the bldD::apr allele was confirmed by PCR and the strain was named SV74. Capture of c-di-GMP-Binding Proteins and Differential Radial Capillary Action of Ligand Assays Cyclic di-GMP capture compound experiments were performed as described previously (Nesper et al., 2012) but with the minor modifications described in Extended Experimental Procedures. Briefly, the c-di-GMP capture compound was added to the soluble lysates and, following UV irradiation in the caproBox, magnetic streptavidin beads were added to the reaction. After incubation the beads were collected, washed, and boiled in sample buffer, the proteins released were run on SDS-PA gels and cut out for mass spectrometry analysis. The DRaCALA assays used His6-BldD or N-terminally His-tagged domains, which were incubated with ∼11 nM 32P-c-di-GMP. The competition experiments had 266 μM cold c-di-GMP or GTP added to the reaction. Samples were spotted onto nitrocellulose membranes and analyzed using Phosphorimaging. DNA-Binding Assays: EMSA and ChIP-Seq Experiments DNA fragments spanning the bldM (158 bp) and whiG (151 bp) promoter regions of S. venezuelae were generated by PCR and 5′ end labeled using [γ32-P]-ATP and T4 polynucleotide kinase. The binding reactions were performed using 0.6 μM His6-BldD and radiolabeled DNA (∼8,000 cpm) as well as 0.5 μg poly[d(I-C)] as nonspecific competitor DNA. When appropriate, increasing amounts of c-di-GMP (0.25–1.75 μM) were added to the mixture. The reaction samples were incubated for 20 min at room temperature and then run on 5% polyacrylamide gels. Chromatin immunoprecipitations were performed as described (Bush et al., 2013) using an anti-BldD polyclonal antibody. Biochemical Studies on BldD Oligomeric State The oligomeric states of BldD and its domains were analyzed via chemical crosslinking using DSS in the presence and absence of c-di-GMP (see the Extended Experimental Procedures) and visualized on SDS-PA gels. Molecular weight analyses using SEC experiments were performed with a HiLoad 16/600 Superdex 75 pg column. Crystallization and Structures Determination of BldD Complexes For detailed descriptions of the protein expression, purification, crystallization, structure determination, and refinement protocols, see the Extended Experimental Procedures. Extended Experimental Procedures Bacterial Strains, Growth Conditions, and Conjugations All E. coli strains used in this study (Table S3) were grown in LB medium under aeration at 37°C. E. coli DH5α was used for plasmid and cosmid propagation and BL21(DE3)pLysS for protein overexpression. BW25113 (Datsenko and Wanner, 2000) containing a λ RED plasmid, pIJ790, was used to create the bldD disruption cosmid and ET12567 containing pUZ8002 (Paget et al., 1999) was used for conjugation experiments. S. venezuelae strains (Table S3) were grown at 30°C on maltose-yeast extract-malt extract (MYM) medium (Stuttard, 1982) containing 50% tap water (MYM-TAP) and 200 μl trace element solution (Kieser et al., 2000) per 100 ml. Liquid cultures were grown under aeration at 250 rpm. Conjugations between E. coli and S. venezuelae were carried out as previously described (Bibb et al., 2012). Construction of Plasmids The oligonucleotides used for plasmid constructions are listed in Table S4. For protein overexpression and purification, bldD and its individual domains were cloned into pET15b resulting in N-terminally His-tagged FL (Full-length) BldD (amino acid residues 1-166), BldD-DBD (amino acid residues 1-79) and BldD-CTD (amino acid residues 80-166). For overexpression of yhjH in S. venezuelae, an N-terminally codon optimized variant of yhjH was cloned downstream of the ermEp ∗ promoter in the ΦBT1 attB site-specific integrative vector pIJ10257 (Hong et al., 2005). Point mutations in bldD were introduced by following the four-primer/two-step PCR protocol (Germer et al., 2001). For complementation analysis the bldD gene carrying its native promoter and the R114D, D116R, R125D and D128R mutations was expressed from the integrative vector pMS82. Construction of a bldD Null Mutant Derivative of S. venezuelae and Phage Transduction The bldD mutant was generated according to the Redirect PCR targeting protocol (Gust et al., 2003, 2004). The Sv-4-H05 cosmid was introduced into E. coli BW25113 and bldD was replaced with the apramycin-resistance (apr) cassette containing oriT, which was amplified from pIJ773 using primers with bldD-specific extensions (Table S4). The disrupted cosmid was confirmed by restriction and PCR analyses and introduced into E. coli ET12567/pUZ8002 for conjugation into S. venezuelae. A null mutant generated by double crossing over was identified by its apramycin-resistant and kanamycin-sensitive phenotype and named SV77 after confirmation by PCR using test primers listed in Table S4. The mutant allele bldD::apr was moved into a new WT background by generalized transduction using the S. venezuelae-specific phage SV1 (Stuttard, 1979). To prepare SV1-lysate, 104 phage were added to 106 SV77 donor spores in 800 μl pre-warmed (45°C) soft nutrient agar (SNA) and poured onto Difco nutrient agar plates containing 0.5% glucose, 10 mM MgSO4 and 10 mM Ca(NO3)2. The plates were incubated at 30°C overnight, then flooded with 2.5 ml Difco nutrient broth (DNB) and incubated for 3-4 hr at room temperature. The phage-containing DNB soak-out was harvested and filtered through a 0.45 μm filter to eliminate bacterial contamination. For transduction of the bldD::apr allele, 109 phage particles harvested from the bldD::apr mutant strain SV77, were mixed with 107-108 WT spores and incubated overnight on MYM agar at room temperature before overlaying with apramycin for selection. Plates spread with the recipient strain or the phage alone were used as controls. Transduction of the bldD::apr allele was confirmed by PCR using test primers listed in Table S4, and the strain was named SV74. c-di-GMP Protein Capture Experiments Cyclic di-GMP capture compound experiments were performed as described previously (Nesper et al., 2012) with minor modifications. S. venezuelae cultures were grown in MYM-TAP supplemented with trace element solution until late transition phase for 24 hr at 30°C and then pelleted by centrifugation for 10 min at 6,000 rpm. The pellet was resuspended in lysis buffer (6.7 mM MES, 6.7 mM HEPES, 200 mM NaCl, 6.7 mM potassium acetate (KAc), pH 7.5) containing protease inhibitor and DNase I. Cells were put through a French Press four times at 18,000 psi and then centrifuged at 100,000 x g for 1 hr. For the capture experiments, the protein concentration of the soluble fraction was determined using a UV/Vis Nanodrop spectrophotometer and 400 μg protein were mixed with 10 μM c-di-GMP capture compound and with 20 μl 5 x capture buffer (100 mM HEPES, 250 mM KAc, 50 mM magnesium acetate (MgAc), 50% glycerol, pH 7.5). 1 mM c-di-GMP was added to the control reaction and incubated for 30 min prior to capture compound addition. The reaction volume was adjusted with H2O to 100 μl and incubated for 2 hr at 4°C in the dark on a rotary wheel. After UV irradiation for 4 min in a caproBox, 50 μl magnetic streptavidin beads and 25 μl 5 x wash buffer (250 mM Tris pH 7.5, 5 M NaCl, 0.1% n-octyl-β-glucopyranoside) were added to the reaction and the mixture was incubated for 45 min at 4°C on a rotary wheel. The beads were then collected with a magnet and washed 6 times with 200 μl wash buffer. The beads were resuspended in 20 μl sample buffer and run for ∼10 min on a 15% SDS polyacrylamide gel after 10 min incubation at 95°C. The gel was stained using InstantBlue Coomassie Stain solution, and a 1×1 cm gel slice containing all captured proteins was excised for analysis by mass spectrometry. Electrophoretic Mobility Shift Assays DNA fragments spanning the bldM (158 bp) and whiG (151 bp) promoter regions of S. venezuelae were generated by PCR using oligonucleotides listed in Table S4 and then 5′ end-labeled using [γ32-P]-ATP and T4 polynucleotide kinase. The binding reactions were performed in bandshift buffer (10 mM Tris pH 7.5, 1 mM EDTA, 5% glycerol, 10 mM NaCl, 1 mM MgCl2) in 20 μl reaction mixture containing 0.6 μM BldD and radiolabeled DNA (∼8,000 c.p.m.) as well as 0.5 μg poly[d(I-C)] as nonspecific competitor DNA. When appropriate, increasing amounts of c-di-GMP (0.25 – 1.75 μM) were added to the mixture. The reaction samples were incubated for 20 min at room temperature followed by electrophoresis on a 5% polyacrylamide gel in 0.5 x TBE (Tris-Borate-EDTA) buffer at 80V for 105 min. The gels were dried before being analyzed on a Phosphorimager. BldD ChIP-Seq Experiments S. venezuelae strains for ChIP-seq were grown in MYM-TAP. Chromatin immunoprecipitations were performed as previously described (Bush et al., 2013), except that an anti-BldD polyclonal antibody was used and pulled down with protein A-sepharose beads. Library construction, sequencing and ChIP-seq data analyses were all carried out as previously described (Bush et al., 2013). Chemical Crosslinking and SDS Polyacrylamide Gel Electrophoresis The His6-BldD-DBD and His6-BldD-CTD proteins were dialyzed into crosslinking buffer (100 mM NaH2PO4, 150 mM NaCl, pH 8) and then incubated at room temperature for 30 min in 20 μl reaction samples containing 10 μM protein, 1 mM disuccinimidyl suberate (DSS) in dimethylsulfoxide (DMSO), and c-di-GMP as indicated. The reaction was stopped by adding 50 mM Tris pH 8 and incubation for 15 min followed by addition of SDS sample buffer and heating to 95°C for 10 min. Samples were separated on a 15% SDS polyacrylamide gel and visualized by Coomassie staining. Determination of c-di-GMP Binding to Proteins by Differential Radial Capillary Action of Ligand Assay Radiolabeled c-di-GMP was synthesized in vitro using [γ32-P]-GTP and the purified diguanylate cyclase PleD∗ as described (Paul et al., 2004). The DRaCALA assays (Roelofs et al., 2011) were performed using 2 μg of His6-BldD or its N-terminally His-tagged domains that were incubated with ∼11 nM 32P-c-di-GMP in DGC buffer (250 mM NaCl, 25 mM Tris pH 8, 10 mM MgCl2, 5 mM β-mercaptoethanol). For competition experiments, 266 μM cold c-di-GMP or GTP were added to the reaction. After a 5 min incubation at room temperature, 5 μl of the binding sample were spotted onto nitrocellulose membrane and the dried membranes were analyzed using a Phosphorimager. Purification, Crystallization, and Structure Determination of S. venezuelae and S. coelicolor BldD CTD-(c-di-GMP) Complexes For structural studies on the CTD, the regions encoding residues 80-166 (S. venezuelae BldD) and 80-167 (S. coelicolor BldD) were cloned into the pET15b vector and the proteins induced at 37°C and purified via Ni-NTA chromatography. The His-tags were removed from the proteins used for structural studies by thrombin cleavage. Crystals of the S. venezuelae BldD CTD-(c-di-GMP) complex, which assumed the trigonal space group, P32, were obtained using protein at 30 mg/mL and 1 mM c-di-GMP. Crystals were produced via the hanging drop vapor diffusion method and mixing the complex 1:1 with 28% PEG 1500, 100 mM sodium acetate, pH 5.5. The S. venezuelae BldD CTD contains no methionines and hence for phasing, Leu92 and Ile135 were substituted with methionines. Semet(L92M/I135M) BldD CTD was expressed using the methionine inhibitory pathway and the protein purified and crystallized with c-di-GMP as per the WT CTD. The selenomethionine-substituted L92M/I135M protein crystallized in the wild-type protein P32 space group. A second crystal form of the S. venezuelae BldD CTD-(c-di-GMP) complex was grown with protein at 20-40 mg/mL and 1 mM c-di-GMP using 20% PEG 2000 monomethyl ether, 100 mM MES, pH 6.0, as a crystallization reagent and took the orthorhombic space group P21212. The third crystal form of the S. venezuelae BldD CTD-(c-di-GMP) complex was produced with protein at 10 mg/mL and 1 mM c-di-GMP using 1.2 M sodium/potassium phosphate, 50 mM citrate pH 5.6. These crystals take the orthorhombic space group C2221. Crystals were obtained for the S. coelicolor BldD CTD-(c-di-GMP) complex using 25 mg/mL protein, 1 mM c-di-GMP and mixing the complex 1:1 with a reservoir comprised of 1.4 M sodium/potassium phosphate, 100 mM HEPES pH 7.5. These crystals take the C2221 space group. Multiple wavelength anomalous diffraction (MAD) data were collected for a Semet(L92M/I135M) S. venezuelae BldD CTD-(c-di-GMP) crystal to 2.28 Å resolution at ALS (Advanced Light Source, Berkeley, CA, USA) beamline 8.3.1 (Table S1). The data were processed using MOSFLM and the heavy atom substructure was obtained via SOLVE (Terwilliger and Berendzen, 1999). The figure of merit (FOM) for the solution was 0.65. Phenix was used for final phasing and density modification (Adams et al., 2010). The crystal contains 12 protein molecules in the asymmetric unit (ASU) and each of the six dimers is glued together by four c-di-GMP molecules. The six c-di-GMP complexed dimers are essentially identical (Figure 4D). Final refinement was done using a data set collected to 1.95 Å resolution for a Semet(BldDL92M) CTD-(c-di-GMP) crystal. A WT data set was also collected to 2.2 Å resolution and the structure was identical to the L92M and L92M/I135M structures. The final 1.95 Å resolution-structure contains residues 84-161 for each of the 12 subunits and 24 c-di-GMP molecules (Table S2). Data were collected to 1.75 Å, 2.25 Å and 2.33 Å resolution for the S. venezuelae and S. coelicolor BldD CTD-(c-di-GMP) C2221 forms and the S. venezuelae BldD CTD-(c-di-GMP) P21212 crystal form, respectively, and the structures solved by molecular replacement (MR). The S. coelicolor BldD CTD and S. venezuelae BldD CTD C2221 crystal forms contain a CTD dimer and four c-di-GMP molecules in the ASU and the S. venezuelae BldD C-domain-(c-di-GMP) P21212 crystal form contains 10 subunits (five dimers), and 20 c-di-GMP molecules. The structures were solved by MR using the program Phaser (McCoy et al., 2007). Final refinement statistics are provided in Table S2. The topology of every BldD CTD is β1 (residues 84-88)-α1 (residues 89-93)-α2 (residues 98-113)-β2 (residues 120-124)-α3 (residues 128-136)-α4 (residues 140-149)-α5 (residues 154-160). Crystallization and Structure Determination of the S. venezuelae BldD DBD The BldD DNA binding domain (BldD DBD), encoding residues 1-79, was cloned into pET15b, expressed in E. coli BL21(DE3) and the protein purified via Ni-NTA chromatography. Prior to crystallization, the hexa-His tag was removed via thrombin cleavage and the protein further purified by size exclusion chromatography. Crystals were grown by mixing the protein (40 mg/mL) 1:1 with a reservoir consisting of 35% PEG 400, 0.1 M MgCl2 and 0.1 M Tris pH 7.5. The crystals take the hexagonal space group P6122. X-ray intensity data were collected to 2.8 Å resolution at ALS beamline 8.3.1 and processed with MOSFLM. The Rsym and I/σ(I) for the data are 11.3% (38.4%) and 11.4 (4.3), respectively, where the values in parentheses indicate data from the highest resolution shell. The structure, which contains 3 subunits in the ASU (2 subunits form a dimer and crystal symmetry generates a second dimer), was solved with Phaser using a single S. coelicolor DBD subunit (PDB code 2EWT) as the search model. The final model contains residues 3-71 of each subunit and was refined using Phenix to final Rwork/Rfree values of 23.7%/28.9%, respectively (Adams et al., 2010). Crystallization and Structure Determination of the S. venezuelae BldD-(c-di-GMP)-21-Mer Complex Crystals of the Full-Length (FL) S. venezuelae BldD-(c-di-GMP)-21-mer complex were grown by using protein in which the N-terminal hexa-His tag had been cleaved and incubated with 1 mM c-di-GMP (final concentration). This BldD-(c-di-GMP) solution was mixed in a 1:1 molar ratio of BldD dimer to 21-mer DNA duplex (5′-CCCCTCACGCTGCGTGACGGG-3′, with the canonical BldD box underlined, annealed to its complementary oligodeoxynucleotide) for crystallization trials. Crystals were grown by mixing the protein-DNA complex 1:1 with the crystallization solution composed of 100 mM sodium citrate tribasic/citric acid pH 4.0 and 200 mM ammonium sulfate. The crystals take the trigonal space group, P3221, with a = b = 114.0 Å, c = 95.2 Å and contain a BldD dimer-(c-di-GMP)-21-mer duplex in the ASU. Data were collected to the limiting resolution of 4.5 Å. The Rsym = 9.0% (86.0%) and I/σ(I) = 5.8 (1.8), where the values in parentheses are for the highest resolution shell. The structure was solved by MR in stages. First, BldD DNA binding domain-DNA complex models were constructed based on the SinR-DNA (PDB code 3ZKC) or λ repressor-DNA (PDB code 1LMB) complex structures and used in the MR program Phaser (McCoy et al., 2007). A clear solution was obtained for the SinR-DNA based model. Packing revealed that the DNA forms a pseudocontinuous helix in the crystal. This solution was then used as a static model with the CTD-(c-di-GMP) dimeric structure as a search model. A clear solution was obtained with MolRep. Due to its low resolution, the structure was subjected to rigid body refinement only (Rwork/Rfree = 33.0%/38.9%, respectively). Proteolysis of FL BldD by Endoproteinase Glu-C To examine the flexibility of the linker region that connects the BldD DBD and CTD, limited proteolysis experiments were carried out. Specifically, the accessibility of residue Glu80, which is the only acidic residue in the BldD linker region, was determined by the ability of the Glu-C protease to cleave after this residue. In these experiments, Endo-Glu-C (100 Units/mL) was a diluted 50 fold into samples of 1) FL BldD (1 mg/mL), 2) FL BldD-21-mer DNA (1 mg/ml protein with 100 μM DNA) or 3) FL BldD-(c-di-GMP)-21-mer (1 mg/mL protein with 100 μM 21-mer DNA and 1 mM c-di-GMP). The proteolysis buffer was 50 mM sodium phosphate, pH 7.5 and the final protein concentration of each sample was 4 mg/mL. The samples shown in Figure S6D were taken at a time point of 4 hr. Notably, proteolysis of each sample was identical, demonstrating that ligand binding of neither cognate DNA nor c-di-GMP by BldD affect the accessibility of the linker region, which remains flexible and unstructured. Size Exclusion Chromatographic Analyses of Wild-Type BldD CTD and the BldD CTD Quadruple Mutant Size exclusion chromatography experiments were carried out using a HiLoad 16/600 Superdex 75 pg column. 5 mg of either the wild-type BldD CTD or the quadruple mutant were loaded onto a column that had been pre-equilibrated with 150 mM NaCl, 5% glycerol, 20 mM Tris HCl pH 7.5 and 3 μM c-di-GMP. The samples were run and eluted with the same c-di-GMP containing buffer. The elution volume was plotted against a standard curve to determine the relative molecular weights of the samples. The standard curve was determined using cytochrome C (12 kDa), carbonic anhydrase (29 kDa) and albumin (66 kDa). Cryo-Scanning Electron Microscopy Cryo-SEM was performed as previously described (Bush et al., 2013). BldD CTD Expression and Purification from Sf9 Cells When induced in E. coli BL21(DE3) at 37°C, BldD CTD samples contained little to no c-di-GMP contamination, which was supported by A280/A260 values of the purified protein, which were approximately 1.6. However, to ensure that there was no endogenous c-di-GMP present in samples used for binding affinity measurements, the BldD CTD was expressed and purified in Sf9 insect cells. For these studies, a gene encoding the same S. venezuelae BldD CTD region that was expressed in E. coli was codon optimized for expression in insect cells (Genscript, Piscataway, NJ, USA; http://www.genscript.com) and subcloned into the expression vector F1, which was transfected into the DH10Bac strain for generation of the recombinant bacmid. Positive bldD CTD containing clones were identified by PCR. rBacmids were then transfected in Sf9 insect cells with Cellfectin II and the cells incubated in Sf-900 II SFM for 56 days before harvest. The supernatant was collected for the P1 viral stock and P2 was amplified for later infection. Sf9 cells expressing BldD CTD were harvested at 72 hr post infection and cells were lysed into 25 mM Tris pH 7.5, 300 mM NaCl and 5% glycerol with protease inhibitors. The protein was purified from the supernatant by Ni-NTA followed by size exclusion chromatography and was > 90% pure as assessed by SDS-PAGE analysis. Determination of the Affinity, Stoichiometry, and Specificity of c-di-GMP for Sf9-Purified BldD CTD by Fluorescence Polarization To measure binding, 2′-O-(6-[Fluoresceinyl]aminohexylcarbamoyl)-cyclic diguanosine monophosphate (2′-Fluo-AHC-c-di-GMP), was used as the fluoresceinated ligand. This molecule is conjugated via a 9 atom spacer to one of the 2′ hydroxyl groups of the c-di-GMP, hence meeting the structural requirement for BldD binding that only one 2′ hydroxyl group be unmodified and the other available for the interactions observed in the BldD-(c-di-GMP) structures (Figure 5). Binding was carried out at 25°C in a buffer of 150 mM NaCl and 25 mM Tris-HCl pH 7.5, which contained 1 nM 2′-Fluo-AHC-c-di-GMP. Increasing concentrations of BldD CTD were titrated into the reaction mixture to obtain the binding isotherms. After each addition of protein the reaction sample was incubated for 30 min to ensure equilibrium had been reached. c-di-GMP binding to BldD appears to have the characteristics of a high affinity/slow binding ligand. The resulting data were plotted using KaleidaGraph and curves were fitted to deduce binding affinities. To determine the binding stoichiometry of the BldD CTD-(c-di-GMP) complex, the same FP binding conditions were used but the total concentration of c-di-GMP (c-di-GMP + 2′-Fluo-AHC-c-di-GMP) was 25 μM, ∼10-fold above the Kd of 2.5 μM ensuring stoichiometric binding. Close inspection of the binding isotherm revealed the possibility that there are two nearly identical binding events (likely corresponding to the binding of each intercalated c-di-GMP dimer). The graph of the resulting data shows a linear increase in the observed mPs until saturation of the binding sites, after which the line is flat. The inflection point(s) are shown in Figure S5C. Importantly, the final inflection occurs at a BldD monomer concentrations of 12 μM, which, when divided by the concentrations of c-di-GMP (25 μM), indicates a stoichiometry of two CTD protomers per four c-di-GMPs. As anticipated from a purely chemical complementarity-orientated argument, FP studies carried out in the same buffer also revealed that wild-type BldD CTD does not bind to c-di-AMP. In addition, the BldD DGR-X8-DQDR CTD mutant did not bind the c-di-GMP probe. Author Contributions N.T. designed, performed, and interpreted experiments, created figures, and wrote the paper. M.A.S. designed, performed, and interpreted experiments, created figures, and wrote the paper. S.S. designed, performed, and interpreted experiments and created figures. N.B.C. performed experiments. K.C.F. performed experiments. R.G.B. designed and interpreted experiments and wrote the paper. M.J.B. designed and interpreted experiments and wrote the paper.