Cyclic di-GMP cyclase SSFG_02181 from Streptomyces ghanaensis ATCC14672 regulates antibiotic biosynthesis and morphological differentiation in streptomycetes

Streptomycetes are filamentous bacteria famous for their ability to produce a vast majority of clinically important secondary metabolites. Both complex morphogenesis and onset of antibiotic biosynthesis are tightly linked in streptomycetes and require series of specific signals for initiation. Cyclic dimeric 3′–5′ guanosine monophosphate, c-di-GMP, one of the well-known bacterial second messengers, has been recently shown to govern morphogenesis and natural product synthesis in Streptomyces by altering the activity of the pleiotropic regulator BldD. Here we report a role of the heme-binding diguanylate cyclase SSFG_02181 from Streptomyces ghanaensis in the regulation of the peptidoglycan glycosyltransferase inhibitor moenomycin A biosynthesis. Deletion of ssfg_02181 reduced the moenomycin A accumulation and led to a precocious sporulation, while the overexpression of the gene blocked sporogenesis and remarkably improved antibiotic titer. We also demonstrate that BldD negatively controls the expression of ssfg_02181, which stems from direct binding of BldD to the ssfg_02181 promoter. Notably, the heterologous expression of ssfg_02181 in model Streptomyces spp. arrested morphological progression at aerial mycelium level and strongly altered the production of secondary metabolites. Altogether, our work underscores the significance of c-di-GMP-mediated signaling in natural product biosynthesis and pointed to extensively applicable approach to increase antibiotic production levels in streptomycetes.


Results and discussion
In silico analysis of SSFG_02181, a putative diguanylate cyclase from S. ghanaensis. In past years, several studies have marked the importance of the second messenger c-di-GMP in Streptomyces life cycle 10,11,[13][14][15] . Interestingly, c-di-GMP metabolizing enzymes were found to be broadly conserved and omnipresent in streptomycetes 15 . Bioinformatic analysis of the S. ghanaensis genome identified ssfg_02181 out of nine genes encoding for putative DGC/PDE enzymes. SSFG_02181 is a 1079-aa protein consisting of a highlyconserved GGDEF domain, followed by a degenerated EAL domain. An N-terminal PAS sensor domain is located upstream the putative DGC active site (A-site), preceded by nine transmembrane-spanning helices (Fig. 1a). Multiple sequences alignment showed SSFG_02181 as a homolog of SCO5511 from S. coelicolor and of SVEN_5187 from S. venezuelae, sharing 80% (sequence coverage 98%) and 72% (sequence coverage 89%) of identity, respectively.
In order to identify conserved residues crucial for the enzymatic activity in SSFG_02181, Clustal Omega software was used to compare the protein sequence with well-characterized homologous proteins in other bacteria. For this purpose, the functional GG(D/E)EF domains of XCC4471 from Xanthomonas campestris 29 , PleD from Caulobacter crescentus 3 and AxDGC2 from Acetobacter xylinum 30 were chosen. As shown in Fig. 1b, SSFG_02181 possesses all five distinctive GGDEF amino acids essential for DGC function. Particularly, two initial glycines are involved in GTP binding, whereas Glu residue at fourth position mediates the ion (Mg 2+ or Mn 2+ ) coordination 3 .  (EcDos) and Acetobacter xylinum (AxPDEA1). Green stars indicate mutated key residues in SSFG_02181 EAL domain. The residues are numbered according to the sequence of SSFG_02181. Alignments were performed using Clustal-Omega (https ://www.ebi.ac.uk/Tools /msa/clust alo/) and the figures were generated using ESPript 2.2 (https ://espri pt.ibcp.fr/ESPri pt/ESPri pt/). www.nature.com/scientificreports/ Asp (in same cases Glu) at the third position is also involved in metal coordination and is crucial for catalysis 3,31,32 . Additionally, SSFG_02181 GGDEF domain contains the canonical RxxD motif (I-site), responsible for binding of dimeric c-di-GMP and thus mediating its own inhibition. Another feature of an active GGDEF domain is the presence of the invariant DxLT motif, which is required for the formation of a stabilizing "wide-turn" structure and represents the starting point of the catalytic site 33 . Other residues in the vicinity of the active site and required for the activity were also found to be highly conserved in SSFG_02181 GGDEF domain 33 , suggesting that this enzyme functions as DGC (Fig. 1b). The predicted EAL domain of SSFG_02181 lacks key residues for the PDE activity and Mg 2+ coordination. In fact, sequence alignment with the enzymatically active PDEs from E. coli (EcDos) 34 , Pseudomonas aeruginosa (RocR) 35 and Acetobacter xylinum (AxPDEA1) 36 showed Glu and Leu residues (of the EAL site) substituted with Cys and Gln, respectively (Fig. 1c). Particularly, Glu was proved to have a crucial role in metal coordination and catalysis, mutations of which result in abolishment of the activity 32,37 . Likewise, a second essential residue involved in Mg 2+ binding (Asn96) 35 is mutated to Arg in SSFG_02181 EAL sequence, further suggesting that this domain is likely enzymatically inactive. Based on these evidences, we predicted SSFG_02181 to work solely as DGC enzyme. SSFG_02181 is an active diguanylate cyclase. To prove an enzymatic activity of SSFG_02181 in vitro, we heterologously expressed ssfg_02181 in E. coli. Despite all our attempts to gain the full protein, it remained insoluble most likely due to the presence of the nine transmembrane helices. A truncated version of the protein (named SSFG_02181 460 ) lacking the first 459 amino acids was purified to homogeneity ( Supplementary Fig. S1) from soluble fraction and used for an in vitro assay. Consequently, SSFG_02181 460 consists of PAS-GGDEF-EAL domains fused with an N-terminal Strep-tag. Strep-tagged SSFG_02181 460 was incubated at 37 °C for 2 h in the presence of 200 µM GTP, leading to the formation of c-di-GMP. As a control, a reaction mixture without the enzyme was used. Furthermore, a mutated version of Strep-tag-SSFG_02181 460 (named SSFG_02181 460 AADEF) carrying both G682A and G683A substitutions in the GGDEF domain was purified and tested for cyclase activity. As shown in Fig. 2, the expected molecular ion of c-di-GMP (m/z = 689.1 [M-H] -) was detected in the reaction sample but not in the negative controls (see MS spectra in Supplementary Fig. S2).
To examine SSFG_02181 for the PDE activity, we incubated SSFG_02181 460 with 200 µM of c-di-GMP. As anticipated from in silico analysis, no hydrolysis of c-di-GMP to pGpG was observed ( Supplementary Fig. S3). In summary, these results confirmed SSFG_02181 to be an active DGC able to convert GTP to c-di-GMP and showed that mutations in the conservative GGDEF domain result in the loss of its catalytic activity. SSFG_02181 is a heme-binding protein. c-di-GMP metabolizing enzymes are often associated with sensor domains, such as PAS or GAF 2 . Likewise, a PAS sensor domain is located at the N-terminal sequence of SSFG_02181. Due to their high plasticity, PAS domains are well-known for their ability to adapt to different ligands, such as heme, FAD, divalent metals or even secondary metabolites 8 . Bioinformatic analysis using BLAST software identified a putative heme-binding pocket in the PAS domain of SSFG_02181 ( Supplementary  Fig. S4). To determine whether SSFG_02181 460 could effectively bind heme, the protein was mixed with hemin in equimolar amounts (10 µM) and UV-vis absorption spectra were recorded (Fig. 3). As a control, the mixture containing no protein was analyzed. As depicted in Fig. 3, the heme spectrum showed a maximum peak at 384 nm and a shoulder in the 550 to 700 nm region of the spectrum. Conversely, the protein-hemin mixture displayed a clear shift of the Soret band (from 384 to 423 nm) and the sharpening of the α-and β-peaks at 557 nm and 528 nm, respectively. Due to the presence of DTT in the buffer, the obtained spectra represent the Fe 2+ , deoxy form of the free heme and the reduced, deoxy form of the heme-protein complex. Collectively, these results demonstrated that SSFG_02181 460 is able to bind heme with an equimolar stoichiometry.
Next, we repeated the in vitro assays using either Strep-tagged SSFG_02181 460 alone or in complex with heme in both Fe(II) and Fe(III) forms. No significant changes in the DGC activity were observed comparing the reaction products of the apoprotein to the heme-protein complexes ( Supplementary Fig. S5), suggesting that heme alone is not sufficient to affect the enzymatic activity.
Most likely, the activity of SSFG_02181 might be influenced by sensing gaseous ligands. Several DGC/PDE enzymes have been found to vary their activity upon binding of gases such as O 2 , CO and NO. The DGC activity of the heme-binding protein YddV from E. coli is strikingly stimulated when O 2 is bound to the heme Fe(II) complex, whereas the heme Fe(II) and the heme Fe(II)-NO complexes result inactive 38 . Conversely, the PDEA1 from Acetobacter xylinum decreases its enzymatic activity when O 2 is bound to the protein 36 . To date, just one c-di-GMP metabolizing enzyme in Streptomyces was proven to be a hemoprotein. The PDE RmdA from S. coelicolor binds hemin through its PAS9 N-terminal domain and it is able to respond to O 2 and CO, albeit its PDE activity does not change upon the binding 11 . Finally, a truncated version of SSFG_02181 was used for the in vitro assay. In fact, SSFG_02181 460 lacks nine N-terminal transmembrane domains which might also be involved in signals transduction upon sensing environmental stimuli. Whether the activity of SSFG_02181 is affected by gaseous ligands and/or by inputs of the transmembrane domains remains obscure and requires further investigations.
Deletion and overexpression of ssfg_02181 cause impaired morphogenesis and alteration of MmA biosynthesis. In order to investigate a role of SSFG_02181 in S. ghanaensis development and MmA production, we constructed an ssfg_02181 marker-free null-mutant through homologous recombination. Analysis of moenomycin production revealed a 2-fold decrease of antibiotic production in Δssfg_02181 in comparison to the wild-type strain (Fig. 4a). In order to exclude polar effect, complementation experiment was performed using the φC31-based integrative vector pSET152. A native copy of the ssfg_02181 gene, along with its own promoter, was cloned into pSET152 (yielding pSET02181) and integrated into S. ghanaensis Δssfg_02181 chromo-Scientific RepoRtS | (2020) 10:12021 | https://doi.org/10.1038/s41598-020-68856-9 www.nature.com/scientificreports/ some. The Δssfg_02181 mutant carrying the empty vector was used as a control. As shown in Fig. 4a, the effects of ssfg_02181 deletion on MmA production were abolished upon gene complementation. However, a substantial increase of antibiotic accumulation was observed in Δssfg_02181 pSET02181 + in comparison to both mutant and wild-type strains. This phenomenon might be due to the integration of an additional copy of the plasmid into pseudo φC31-sites in the genome of S. ghanaensis Δssfg_02181, as it was already shown for others Streptomyces spp. 39 . Next, we studied the influence of ssfg_02181 deletion on morphological development. The mutant strain was cultivated on SFM agar medium for 5 days at 37 °C. After 2 days of growth, Δssfg_02181 displayed the characteristic green pigment associated with spores formation on its surface, whereas the surface of a control strain still remained white (Fig. 5a). Scanning electron microscopy revealed chains of spores in S. ghanaensis Δssfg_02181 in contrast to long nonsporulating aerial hyphae of the wild-type strain (Fig. 5b). The wild-type phenotype was restored upon complementation with a native copy of the gene (plasmid pSET02181), implying www.nature.com/scientificreports/  www.nature.com/scientificreports/ that the significant decrease of antibiotic production as well as the precocious sporulation were solely due to the ssfg_02181 disruption. Further confirmation of the SSFG_02181 involvement in morphogenesis and secondary metabolite production was achieved upon gene overexpression. For this purpose, the ssfg_02181 coding sequence was placed under control of the strong constitutive promoter ermEp, into the φC31-based integrative vector pTES. The resulting pTES02181 plasmid was transferred into S. ghanaensis. As a control, the wild-type strain carrying an empty copy of the vector was used. Unlike gene deletion, overexpression of ssfg_02181 led to 5-fold increase of MmA production ( Fig. 4b) and delay in morphological development (Fig. 5c). Scanning electron microscopy revealed that after 5 days the morphological progression of S. ghanaensis containing pTES02181 was arrested at the aerial mycelium stage, whereas the control strain, carrying an empty vector, developed mature spores (Fig. 5d).
Our results report the involvement of a c-di-GMP metabolizing enzyme in S. ghanaensis life cycle. In a line with our in silico analysis and the enzymatic in vitro assay, we demonstrated that SSFG_02181 is an active DGC. Homologs of SSFG_02181 acting as DGCs have been identified and characterized in other Streptomyces spp. In particular, cdgB and cdgC null-mutants of S. venezuelae 15 exhibit similar phenotype to that obtained after ssfg_02181 deletion (e.g. precocious sporulation). In addition, overexpression of ssfg_02181 led to a delay in morphological development, as it was shown for cdgA, cdgB and cdgD overexpressing mutants in S. coelicolor 9,10,13 .
Recently, we showed that manipulations of c-di-GMP metabolizing enzymes in S. ghanaensis severely affect secondary metabolites production. Indeed, c-di-GMP mediates the dimerization of BldD, which in turn activates adpA gh expression and inhibits wblA gh transcription 12 . AdpA was shown to act as a positive transcriptional regulator of key structural genes in moe cluster 22 , whereas WblA was shown to negatively regulate MmA production 23 . In this study, deletion of ssfg_02181 results in a significant decrease of antibiotic production. Similarly to the www.nature.com/scientificreports/ ΔcdgB gh mutant 12 , we assume that the loss of SSFG_02181 causes a reduction of intracellular c-di-GMP pool and thus the dissociation of BldD dimer from the target promoters (e.g. adpA and wblA).
In the past years, several studies were conducted to fully understand the functions and mechanisms of action of DGCs and PDEs in streptomycetes. Genetic manipulation and biochemical analysis confirmed that these enzymes are responsible for the metabolism of c-di-GMP, which intracellular level is crucially balanced over the entire bacterial life cycle. Pool of this second messenger in turn regulates both morphogenesis and secondary metabolite production by controlling both BldD and the σ WhiG factor 12,14,19 . However, Streptomyces spp. possess more than one c-di-GMP-metabolizing enzymes, suggesting that these proteins might be active at different time points and spatial locations during the cell life. Finally, the majority of these enzymes carry one or more sensor domains whose functions have not yet been fully clarified. Most likely, these domains regulate the catalytic activity by responding to either external or internal stimuli, thus maintaining a well-tuned control over intracellular c-di-GMP levels in the cell.
Heterologous expression of ssfg_02181 affects antibiotic production and morphological development in Streptomyces spp.. In order to determine whether overexpression of ssfg_02181 could affect secondary metabolites production and morphological development in other streptomycetes, we introduced an additional copy of the gene controlled by the strong constitutive promoter ermEp (plasmid pTES02181) into two model strains S. coelicolor M145 and S. albus J1074. As a control, strains carrying an empty copy of the vector pTES were used. After 3 days of growth, S. coelicolor pTES02181 exhibited a significant increase of actinorhodin production on SFM agar, in comparison to the control strain (Fig. 6a). Moreover, phenotypic analysis revealed the inability of S. coelicolor pTES02181 to develop aerial mycelium on R2YE agar (Fig. 6b) and to sporulate on SFM agar after 5 days of growth (Fig. 6a). Similarly, the overexpression of ssfg_02181 in S. albus J1074 led to an arrest of morphological development on SFM agar in comparison to the control strain (Fig. 6c). These results demonstrate that SSFG_02181 retains its activity upon heterologous expression, suggesting that similar mechanisms of c-di-GMP-mediated regulatory network operate in other streptomycetes. Indeed, protein sequence alignment of SSFG_02181 and its orthologs from S. coelicolor (SCO5511), S. venezuelae (SVEN_5187), S. griseus (SGR_2001) and S. albus (XNR_1322) showed that these proteins possess a degenerated EAL domain but share the conservative GGDEF domain (Supplementary Fig. S6), indicating that all of them likely function as DGCs. www.nature.com/scientificreports/ Due to the fact that orthologs of this gene are broadly conserved in streptomycetes 15 , the overexpression of ssfg_02181 could be employed as a useful approach to increase antibiotic production titers in other actinobacteria.
The transcription of ssfg_02181 is repressed by the master regulator BldD. The master pleiotropic regulator BldD is one of the most conserved transcriptional factors in streptomycetes, controlling diverse processes in the cell 13,40 . Recently, den Hengst et al. 13 identified the BldD-binding consensus site, consisting of the 15-bp palindromic sequence nTnACnC(A/T)GnGTnAn, named BldD-box. In S. coelicolor, BldD-box was found to be located also in the promoter regions of three genes encoding for c-di-GMP metabolizing enzymes (cdgA, cdgB and sco5511) 13 . Likewise, bioinformatic analysis 12 on ssfg_02181 promoter region identified two sequences (TCT ACG CTC CGT AAC and ATG TCC CTG AGT GAC) (Fig. 7a) resembling the consensus BldDbox. These sequences showed high similarity to the ones uncovered by den Hengst et al. 13 in sco5511 from S. coelicolor. Alignment of BldD from S. ghanaensis and its orthologs from S. coelicolor, S. venezuelae and S. albus revealed an identical DNA binding domain (DBD) and the presence of both c-di-GMP binding motifs RxD-X 8 -RxxD ( Supplementary Fig. S7a), showing that BldD in S. ghanaensis retains all the features required to function as a transcriptional regulator. Indeed, BldD was recently proved to bind to the promoter sequence of cdgB gh in S. ghanaensis 12 .
To determine BldD binding affinity to the ssfg_02181 promoter region, we performed an electrophoretic mobility shift assay (EMSA). An N-terminally His 6 -tagged BldD protein was produced in E. coli and purified. Increasing amounts of BldD were incubated with the 33 P-radiolabeled promoter region of ssfg_02181 at the presence of 1 µM c-di-GMP. As shown in Fig. 7b, BldD was able to bind the target sequence at concentration starting from 0.05 µM. Increasing concentrations of BldD strongly increased binding efficacy, resulted in an almost complete DNA shift at 0.5 µM. To prove the specificity of the BldD binding, excess of unlabeled ssfg_02181-promoter DNA was used to compete with the radiolabeled one for BldD ( Supplementary Fig. S8a). Moreover, we confirmed that the affinity between BldD and ssfg_02181 promoter is affected by the presence of c-di-GMP. In fact, increasing concentration of c-di-GMP greatly enhanced BldD binding to the target promoter region, whereas the elimination of the second messenger from the reaction mixture resulted in no DNA shift ( Supplementary Fig. S8b).
Next, we aimed to investigate which of two identified consensus sequences TCT ACG CTC CGT AAC (here named BldD-box I) and ATG TCC CTG AGT GAC (here named BldD-box II) effectively corresponds to the major BldD-box in the ssfg_02181 promoter region. To do this, we designed a set of oligonucleotide sequences. One of them carried a nonsense mutation in the BldD-box I, another in the BldD-box II and the last in both simultaneously. As a control, a nucleotide sequence carrying both native BldD-boxes was used. From the results of EMSA, depicted in Fig. 7c, BldD could bind to the native oligonucleotide fragment already at concentration of 0.05 µM, whereas the mutation in BldD-box I completely inhibited the binding (Fig. 7d). Conversely, the mutation in BldD-box II caused an alteration of the binding affinity, resulting in a higher concentration of BldD required for the binding (0.15 µM) and less amount of shifted DNA (Fig. 7e). Finally, the lack of both sites led to analogous effects observed with the mutation of BldD-box I (Supplementary Fig. S8c).
Next, we analyzed the influence of BldD on the transcription of ssfg_02181 by semiquantitative (sq)RT-PCR and GusA reporter assay 41 . For sqRT-PCR, the cDNA from S. ghanaensis wild-type and S. ghanaensis ΔbldD 12 were used as template. As depicted in Fig. 7f, the transcriptional levels of ssfg_02181 were found to increase in S. ghanaensis ΔbldD, in comparison to the wild-type strain. This observation was further confirmed by the GusA reporter system, showing a 1.7-fold increase in the transcriptional activity of the ssfg_02181 promoter in the bldD mutant (Fig. 7g). Likewise, deletion of ssfg_02181 in S. ghanaensis caused an increase in the transcriptional activity of its own promoter (Fig. 7g). We propose that the lack of SSFG_02181 causes a drop in c-di-GMP pool, which favors the dissociation of BldD dimer from the ssfg_02181 promoter. Taken together, these data confirmed that BldD effectively binds to the promoter of ssfg_02181 and represses its transcription. Based on our results, we could demonstrate that BldD-box I represents the primary binding site of BldD on ssfg_02181 promoter. Furthermore, we showed that BldD-box II is likely to be required for enhancing the binding efficacy between BldD and ssfg_02181 promoter. These consensus sites were also identified in the promoter of sco5511 in S. coelicolor 13 and multiple sequence alignment revealed that both BldD-boxes are shared and highly-similar in ssfg_02181 promoter and its orthologs (Supplementary Fig. S7b). We speculate that the presence of BldD-box II might reinforce the binding affinity of BldD to the target promoter, regulating the ssfg_02181 transcription at different BldD/c-di-GMP concentrations. Nevertheless, the biological significance of BldD-box II in vivo remains obscure and needs to be investigated in the future.
In addition, we hypothesize that intracellular c-di-GMP levels undergo a fine-tuned regulation through a negative feedback mechanism. Notably, the expression of the DGC-encoding genes (e.g. cdgA, cdgB and orthologs of ssfg_02181) is under control of BldD. These genes were found to be omnipresent among streptomycetes 15 , indicating that such mechanism might be universally present across Streptomyces. Likely, the transcription of these genes is blocked by the BldD dimer when c-di-GMP concentrations reach a certain threshold. As result, the homeostasis of the second messenger is retained during the entire bacterial life cycle. However, it can not be excluded that the binding of BldD might be also required to temporally control the transcription of these genes. Studies in S. venezuelae revealed that cdgB is expressed in all developmental stages, whereas expression of cdgC (ortholog of ssfg_02181) increases over the time and reaches the highest levels during sporulation. Finally, cdgA is the least transcribed among all DGCs 15 . Therefore, all these findings point to the conclusion that regulatory mechanisms governing the c-di-GMP levels are immensely intricate and well-tuned to permit a coordinate morphological progression and antibiotic synthesis in actinobacteria. www.nature.com/scientificreports/

Conclusions
In the past few years, c-di-GMP has emerged as a crucial molecule involved in streptomycetes development, controlling both morphogenesis and secondary metabolite production [9][10][11][12]14,15 . The intracellular c-di-GMP turnover is modulated by the coordinated activities of DGCs and PDEs, encoded by streptomycetes genome. In this study, we demonstrated that the DGC SSFG_02181 affects MmA biosynthesis and morphogenesis in S. ghanaensis.
Recently, two more c-di-GMP metabolizing enzymes in S. ghanaensis, CdgB gh and RmdB gh , were found to be involved in secondary metabolite production and morphological development 12 . Altogether, these data demonstrate the decisive role of c-di-GMP in S. ghanaensis life cycle. Additionally, we found that the transcription of ssfg_02181 is directly controlled by the pleiotropic regulator BldD, providing an evidence that the intracellular c-di-GMP pool is strictly regulated to maintain its homeostasis during the bacterial life cycle. Furthermore, we showed that heterologous expression of ssfg_02181 affects antibiotic production and morphological development in model Streptomyces spp. Finally, we found that SSFG_02181 is able to bind heme, albeit the correlation between the enzymatic activity and the ligand binding remains to be further investigated. Collectively, our study provides an effective approach to improve the production of secondary metabolites in various actinobacteria, coupling genetic manipulation of c-di-GMP metabolizing enzymes to their functional analysis.

Materials and methods
Strains, plasmids and growth conditions. The bacteria strains and plasmids used in this work are listed in Supplementary Table S1. Luria-Bertani (LB) medium was used to grow E. coli strains. S. ghanaensis strains were grown on soya flour mannitol agar (SFM) and oatmeal agar media and in TSB liquid medium at 37 °C. S. coelicolor M145 was grown on SFM and R2YE agar media at 28 °C. S. albus J1074 was cultivated on SFM at 28 °C. Where necessary, appropriate antibiotics were added to the media.

DNA manipulation and intergeneric conjugation.
Polymerase chain reactions (PCRs) were performed using Phusion polymerase (ThermoFisher). All primers used in this work are listed in Supplementary  Table S2. Plasmid/chromosomal DNA isolations were carried out using standard procedures 42 . All plasmids were verified by enzymatic digestion, PCR or sequencing. Conjugal transfer of plasmids from E. coli to Streptomyces spp. was achieved using the dam dcm mutant strain E. coli ET12567, carrying the helper plasmid pUZ8002 and performed as described previously 43 .
Scanning electron microscopy. For morphological evaluation, a slice of S. ghanaensis strains was recovered from the SFM agar plates and placed on a metal stub. Each sample was introduced into a Quanta 250 environmental scanning electron microscope (SEM) and images were captured at magnification 12,800 ×.
Bioinformatic analysis. SSFG_02181 protein sequence was analyzed by the SMART domain database (https ://smart .embl-heide lberg .de/). Multiple sequence alignments of BldD and the DGC and PDE domains were constructed using the Clustal Omega software (https ://www.ebi.ac.uk/Tools /msa/clust alo/) and the figures were generated using the ESPript 2.2 software (https ://espri pt.ibcp.fr/ESPri pt/ESPri pt/). BLAST software was employed to identify a putative heme-binding site in SSFG_02181 sequence. Finally, the identification of a BldD-binding site on ssfg_02181 promoter was done as described elsewhere 12 and the alignment of ssfg_02181 promoter and its orthologs was performed using Clustal Omega software.

Construction of S. ghanaensis ssfg_02181 null mutant and plasmid for complementation experiment.
The DNA fragment containing ssfg_02181 coding sequence and its flanking regions was amplified from the genomic DNA of S. ghanaensis by PCR (primers 02181_del_for and 02181_del_rev) and cloned into the EcoRV-digested pBluescriptKS vector, yielding pBlue02181. The apramycin resistance cassette (aac(3)IV), flanked by loxP sites, was amplified from pLERECJ using 02181_kn_for and 02181_kn_rev primers. pBlue02181 and aac(3)IV were introduced into E. coli BW25113/pIJ790 and ssfg_02181 was replaced with the gene marker by REDIRECT technology 44 . The resulting fragment 02181::aac(3)IV was amplified by PCR www.nature.com/scientificreports/ using the same primer pair and cloned into the suicide EcoRV-digested pKGLP2 vector (hygromycin resistance), yielding pKG02181::aac(3)IV. The latter was transferred into S. ghanaensis by conjugation. Double-crossover mutants were selected for apramycin resistance and hygromycin sensitivity. The replacement of ssfg_02181 with the aac(3)IV gene marker in the genome of S. ghanaensis Δ02181::aac(3)IV was confirmed by PCR. In order to excise the apramycin cassette and yield a marker-free mutant, the Cre-expressing helper plasmid pUWLCre was introduced into S. ghanaensis Δ02181::aac(3)IV by conjugation 45 . The exconjugants were selected for resistance to thiostrepton followed by screening for sensitivity to apramycin, yielding S. ghanaensis Δssfg_02181. The deletion of apramycin cassette from the genome of S. ghanaensis Δssfg_02181 was confirmed by PCR. For complementation experiment, the ssfg_02181 coding sequence, along with its own promoter region, was amplified by PCR (primers 02181_compl_for and 02181_compl_rev) and digested with XbaI and EcoRI. The fragment was cloned into the XbaI-EcoRI digested pSET152 vector, to gain pSET02181. The latter was then transferred into S. ghanaensis Δssfg_02181 by conjugation.
Overexpression of ssfg_02181 in S. ghanaensis, S. coelicolor and S. albus. In order to create the ssfg_02181-overexpressed S. ghanaensis strain, the DNA fragment comprising only the coding sequence of ssfg_02181 was amplified by PCR (primers 02181_exp_for and 02181_exp_rev) and digested with XbaI and EcoRI. Next, the resulting fragment was cloned under the control of the strong, constitutive promoter ermEp in the integrative pTES vector, linearized by XbaI and EcoRI. The resulting plasmid pTES02181 was transferred into S. ghanaensis, S. coelicolor M145 and S. albus J1074 strains by conjugation.
Analysis of moenomycin production. The quantification of moenomycin production levels was done as described previously 46,47 . Briefly, S. ghanaensis strains were grown in 50 ml TSB medium for 4 days, in triplicate. The cellular pellet was collected by centrifugation and mixed with 10 ml of methanol overnight. The resulting extracts were concentrated in vacuo and dissolved in methanol prior analysis by HPLC-MS. The analysis was performed as established previously 12  Proteins production and purification. The truncated N-terminal Strep-tagged version of SSFG_02181 was produced as following. DNA sequence encoding the PAS-GGDEF-EAL domains of SSFG_02181 (here named SSFG_02181 460 ) was amplified by PCR (primers 02181_460_for and 02181_460_rev) and digested with BamHI and HindIII. The digested fragment was cloned into BamHI-HindIII linearized pET51b vector, yielding pET02181 460 . To generate a mutated version of SSFG_02181 460 (named SSFG_02181 460 AADEF), PCR mutagenesis was applied to amplify a 7 kb DNA fragment from pET02181 460 using primers designed to bring G682A and G683A substitutions (02181_460_aadef_for and 02181_460_aadef_rev). The resulting amplicon was treated with T4 Polynucleotide kinase and then self-ligated, yielding pET02181 460 _AADEF. pET02181 460 and pET02181 460 _AADEF were individually introduced into E. coli BL21 Star (DE3) pLysS. The strain was grown at 37 °C to OD 600 of 0.5 and the protein production was induced with 0.2 mM IPTG. Following 5 h of incubation at 22 °C, cells were harvested by centrifugation and the pellet was resuspended in Step-tag equilibration buffer (50 mM Tris-HCl, 0.5 M NaCl, 1 mM MgCl 2 , 5 mM β-mercaptoethanol and 5% glycerol, pH 8.0). Cell lysis was achieved using three passages through French-press and the lysate was centrifuged at 14.000 rpm for 35 min at 4 °C. The soluble cell extract was loaded onto a column containing 4 ml of pre-equilibrated Strep-Tactin resin (IBA). Fractions were eluted using Strep-tag equilibration buffer containing 2.5 mM desthiobiotin and pooled together. Size-exclusion chromatografy was performed using the ÄKTA fast protein liquid chromatography (FPLC) system, equipped with a Superdex 200 h 16/60 column. After gel filtration, the protein was stored at − 80 °C in a buffer containing 50 mM Tris-HCl, 0.5 M NaCl, 1 mM MgCl 2 , 1 mM dithiotreitol and 5% glycerol, pH 7.5.
Enzymatic in vitro assay to determine diguanylate cyclase activity. Enzymatic in vitro activity assay was performed by adding 5 µM Strep-tagged SSFG_02181 460 protein into a reaction mixture (total volume 100 µl) containing 200 µM GTP, 50 mM Tris-HCl [pH 8.0], 50 mM NaCl and 10 mM MgCl 2 . The mixture was incubated at 37 °C for 2 h. The reaction was stopped by adding 10 mM CaCl 2 , followed by incubation at 70 °C for 5 min. LC-MS was used to analyze the reaction products, as described previously 12 . C-di-GMP and GTP standards were purchased from InvivoGen and Sigma-Aldrich, respectively.
Heme reconstitution and UV-vis spectroscopy. In order to determine whether Strep-tagged SSFG_02181 460 protein could bind heme, hemin (stock solution 1 mM in DMSO) was added to the purified protein (10 µM) at equimolar concentrations. Following a 30 min incubation at room temperature, the UV-vis absorption spectra were measured with 100 µl protein-hemin complex using a Jasco V-650 spectrophotometer. A solution of 10 µM hemin in a protein-storage buffer was prepared as a control. All experiments were performed under anaerobic conditions in a glove-box at room temperature.
Mutagenesis of BldD-binding sites on ssfg_02181 promoter. In order to identify the major BldD binding site in ssfg_02181 promoter region, three pairs of 53 bp complementary oligonuleotide strands were designed where BldD-boxes were replaced by a poly(A) non-sense sequence. Particularly, PmutI_for and PmutI_rev primers were used for mutating BldD-box I, PmutII_for and PmutII_rev for mutating BldD-box II and PmutIII_for and PmutIII_rev for mutating both boxes simultaneously. As a control, a pair of complementary oligonuclotide strands (Pnat_for and Pnat_rev) carrying both native sequences was used. Complementary oligonucleotides were mixed together at a 1:1 molar ratio and diluted in a buffer (10 mM Tris [pH 7.5], 100 mM NaCl and 1 mM EDTA) to a final concentration 1 pmol/µl. To anneal, the tube with oligonucleotides was incubated in a boiling water for 2 min and let slowly cool down to room temperature. Hybridized oligonucleotides were then 5′-end labelled with [γ-33 P]-ATP using T4 polynucleotide kinase and applied for electrophoretic mobility shift assay experiments.
β-glucuronidase activity measurements. To investigate the transcriptional activity of ssfg_02181 promoter, the DNA fragment containing the promoter region of ssfg_02181 was amplified by PCR (primers 02181_ script_for and 02181_script_rev) and digested with XbaI and KpnI. The obtained fragment was cloned into XbaI, KpnI-linearized pGUS vector, yielding pGUS02181. The latter was transferred into S. ghanaensis strains by conjugation. β-glucuronidase activity was measured as described previously 41 . The experiments were carried out at least three times. The values refer back to equal amount of dry biomass (10 mg) and correspond to the mean of independent experiments. Error bars indicate the standard deviations.

Semiquantitative (sq)RT-PCR.
Semiquantitative RT-PCR was used to determine differences in the transcription levels of ssfg_02181 between S. ghanaensis strains. sqRT-PCR experiments were performed as described previously 12 . Briefly, total RNA samples were isolated in triplicate from streptomycetes cultures grown in TSB for 48 h. One mcg of total RNA per reaction was used to synthesize cDNA using Photoscript II Reverse Transcriptase (NEB). Two hundred ng of cDNA were used for PCR (primers 02181_check_for and 02181_RT_rev) and the expression levels of the gene were estimated by visual examination. As a control, primers specific to the sequence of hrdB encoding the RNA polymerase principal sigma factor were used 12 .