The conserved actinobacterial transcriptional regulator FtsR controls expression of ftsZ and further target genes and influences growth and cell division in Corynebacterium glutamicum

Background Key mechanisms of cell division and its regulation are well understood in model bacteria such as Escherichia coli and Bacillus subtilis. In contrast, current knowledge on the regulation of cell division in Actinobacteria is rather limited. FtsZ is one of the key players in this process, but nothing is known about its transcriptional regulation in Corynebacterium glutamicum, a model organism of the Corynebacteriales. Results In this study, we used DNA affinity chromatography to search for transcriptional regulators of ftsZ in C. glutamicum and identified the Cg1631 protein as candidate, which was named FtsR. Both deletion and overexpression of ftsR caused growth defects and an altered cell morphology. Plasmid-based expression of native ftsR or of homologs of the pathogenic relatives Corynebacterium diphtheriae and Mycobacterium tuberculosis in the ΔftsR mutant could at least partially reverse the mutant phenotype. Absence of ftsR caused decreased expression of ftsZ, in line with an activator function of FtsR. In vivo crosslinking followed by affinity purification of FtsR and next generation sequencing of the enriched DNA fragments confirmed the ftsZ promoter as in vivo binding site of FtsR and revealed additional potential target genes and a DNA-binding motif. Analysis of strains expressing ftsZ under control of the gluconate-inducible gntK promoter revealed that the phenotype of the ΔftsR mutant is not solely caused by reduced ftsZ expression, but involves further targets. Conclusions In this study, we identified and characterized FtsR as the first transcriptional regulator of FtsZ described for C. glutamicum. Both the absence and the overproduction of FtsR had severe effects on growth and cell morphology, underlining the importance of this regulatory protein. FtsR and its DNA-binding site in the promoter region of ftsZ are highly conserved in Actinobacteria, which suggests that this regulatory mechanism is also relevant for the control of cell division in related Actinobacteria. Electronic supplementary material The online version of this article (10.1186/s12866-019-1553-0) contains supplementary material, which is available to authorized users.


Supplementary methods
Coulter counter measurements Bacterial size distribution was determined via the Coulter Principle using a MultiSizer 3 (Beckman Coulter, Krefeld, Germany) particle counter equipped with a 30 µm capillary. Briefly, for each measurement the bacterial cells at an OD600 of ~0.1 were 20-fold diluted in CASYton assay buffer (Shärfe Systems, Reutlingen/Germany). Each sample (100 µl) was analyzed twice in the volumetric measurement mode. The data were visualized and extracted using the Beckman Coulter Multisizer 3 software package.

Quantitative PCR
To elucidate how many copies of the genes cg0834 and cg0840 were present in the genome of different C. glutamicum strains, quantitative PCR (qPCR) was performed. For this purpose, genomic DNA was isolated as following. Selected strains were incubated overnight in 5 mL BHI medium and harvested by centrifugation. The cells were washed in TE buffer (10 mM Tris-HCl, 1 mM Na2EDTA, pH 7.6), centrifuged again and suspended in 1 mL TE buffer with 15 mg lysozyme. After three hours incubation at 37°C, 3 mL lysis buffer (10 mM Tris-HCl, 400 mM NaCl, 2 mM Na2EDTA, pH 8.2), 220 µL 10% (w/v) SDS and 150 µL Proteinase K (20 µg/mL) were added to the suspension and mixed carefully, followed by an additional incubation for two hours at 60°C. 2 mL saturated saline solution was added and the samples were shaken vigorously until appearance of white precipitate, which was spun down by centrifugation for 30 minutes and 16.000 g at room temperature. The cleared supernatant was transferred into a fresh reaction tube and the DNA was precipitated by careful mixing with 2.5 volumes of ice cold ethanol absolute. The DNA was removed from the tube with a Pasteur pipette whose tip was bent before using a Bunsen burner, dipped into 70 % Ethanol in a fresh Eppendorf tube for washing, and air-dried for a few seconds. Subsequently, the DNA was solubilized in 200 µL TE buffer and incubated at 4°C overnight. For quality control, the isolated genomic DNA was analyzed by agarose gel electrophoresis. The DNA concentration was determined using a Colibri Microvolume Spectrophotometer (Berthold Detection Systems GmbH, Pforzheim, Germany) and adjusted to 50 ng/µL. For the qPCR experiment, PCR fragments for the generation of standard curves were amplified with chromosomal C. glutamicum DNA as template and the oligonucleotides given in table S5. The gene recF was used as reference gene, as it has proven to be well suitable in previous projects. The primers were designed with the Primer3plus online tool (1), using the standard settings for qPCR and an annealing temperature of 60°C. The amplified fragments were checked for purity by agarose gel electrophoresis. The standards were used in concentrations of 10 pg/µL to 100 ag/µL in a gradient cycle protocol to determine the optimal annealing temperature for the qPCR experiment. The setup of the cycling protocol was as following. 3 min preincubation at 95°C (step 1), 5 sec denaturation at 95°C (step 2), 25 sec elongation at a temperaturegradient from 55.1°C to 66.9°C (step 3), 40 times repetition of steps two and three, followed by a melting curve analysis (step 4) from 60°C to 95°C, with T = 1°C for every 6 seconds. Again, agarose gel electrophoresis of the PCR products was performed by agarose gel electrophoresis. Based on this experiment, an annealing temperature of 59°C was chosen for the qPCR experiment. The qPCR reaction was performed using the innuMIX qPCR MasterMix SyGreen (Analytik Jena, Jena, Germany) and the qTOWER 2.2 (Analytik Jena, Jena, Germany), and the protocol stated above. The reaction mix contained 10 µL master mix (2x), 1 µL primer 1 (10 µM), 1 µL primer 2 (10 µM), 2 µL template (50 ng/µL) and 6 µL H2O. The data was analyzed using the program qPCRsoft 3.1 (Analytik Jena, Jena, Germany) and the Ct method.

Genome resequencing
For genome re-sequencing, C. glutamicum ATCC13032 and three independent clones of C. glutamicum ATCC13032ΔftsR were cultivated overnight in 20 mL BHI medium and the DNA prepared (2). Genomic DNA was purified using the NucleoSpin ® Microbial DNA Kit (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany). 4 µg were used for library preparation and indexing with the TruSeq ® DNA PCR-Free Sample Preparation Kit (illumina Inc., San Diego, CA, USA). Quantifications of the resulting libraries were conducted using KAPA Library Quantification Kits (PEQLAB Biotechnologie GmbH, Erlangen, Germany) and were normalized for pooling. A MiSeq TM sequencing device (illumina Inc., San Diego, CA, USA) was used for paired-end sequencing with a read-length of two times 150 bases. Data analysis and base calling were accomplished with the illumina ® instrument software and stored as fastq output files. The sequencing data obtained were imported into CLC Genomics Workbench (Qiagen Aarhus A/S, Aarhus, Denmark) for trimming and base quality filtering. The output was mapped to accession BX927147 as the C. glutamicum ATCC13032 reference genome. The resulting mappings were used for the quality-based single nucleotide polymorphisms (SNPs) variant detection with CLC Genomics Workbench. The detected SNPs were manually inspected for relevance.

Supplementary results
Genome-resequencing of ATCC13032ftsR uncovers amplification of the DNA region encompassing cg0828 -cg0840 The transcriptome comparison of ATCC13032ftsR with ATCC13032 revealed 2.6-to 4.6-fold increased mRNA levels of the genomic region encompassing cg0830 -cg0838 in the ftsR mutant (Table S3). Interestingly, two of the nine genes annotated in this region are oriented in the opposite direction than the other seven, but nevertheless also showed an increased mRNA level. Although this could be a consequence of a similar regulation of the nine genes, it might also result from an amplification event leading to an increased DNA copy number and therefore increased mRNA levels. We therefore sequenced the entire genome of three clones of strain ATCC13032ftsR. For all three clones, this analysis indeed revealed an amplification event. As shown in Table S3, the sequence coverage of the DNA region from cg0828 to cg0840 was 5-to 7-fold (mean value 6.36 ± 0.57) higher than that of the residual genome in all three clones. When looking at the DNA microarray data, it becomes evident that the mRNA levels of cg0828, cg0829, cg0839, and cg0840 were also 2.5-to 3.4fold increased in the ftsR mutant, but since the p-value was above 0.05, they were not included in Table  S2. The results of genome resequencing were confirmed by qPCR, which revealed a ≥5-fold increased DNA level for cg0834 and cg0840 compared to the reference gene recF (cg0005) in the ATCC13032ftsR strain (Fig. S7). The mechanism of the DNA amplification is unknown and the exact genomic structure of the amplification cannot be deduced from the short sequencing reads. In four independent clones of MB001ftsR, qPCR did not reveal an increased DNA level of cg0834 and cg0840 (Fig. S7), indicating that the amplification event only occurred in ATCC13032ftsR, but not in the prophage-free strain MB001ftsR. Since ftsR deletion mutants were constructed only once for each strain, firm conclusions on a functional correlation between the observed amplification event and the presence of the prophages cannot be drawn. However, the amplification could be responsible for the differences in the complementation studies between the wild type and the MB001 strain described in the main manuscript.

Influence of FtsR on ftsZ promoter activity in strains with FtsR-independent ftsZ expression.
The ftsR deletion in C. glutamicum MB001 led to a significant reduction of ftsZ promoter activity (Fig.  4). In order to confirm that the observed activation of the ftsZ promoter by FtsR is independent of the actual ftsZ expression, strains MB001::PgntK-ftsZ and MB001ftsR::PgntK-ftsZ were transformed with the reporter plasmid pJC1-PftsZ-venus and either pEC-ftsR or pEC-XC99E as empty plasmid control (Fig. S13). With ftsZ being under control of PgntK and addition of the same gluconate concentration to the medium, expression of the chromosomal ftsZ should be identical for all strains. In plasmid pEC-ftsR the ftsR gene is expressed under control of the IPTG-inducible but leaky trc promoter (Fig. S13). The growth conditions were chosen according to the previous experiment where the mixture of 0.01% (w/v) gluconate and 1.99% (w/v) glucose led to comparable growth of strains MB001::PgntK-ftsZ and MB001ftsR::PgntK-ftsZ. The latter strain carrying pJC1-PftsZ-venus and the vector pEC-XC99E showed a significant growth defect in comparison to the other strains tested (Fig. S13A, red curve). This defect must be due to altered expression of genes besides ftsZ that are regulated directly or indirectly by FtsR. Accordingly, the growth defect was reversed in the presence of the ftsR expression plasmid (Fig. S13A, blue curve). The activity of the native ftsZ promoter on plasmid pJC1-PftsZ-venus was much lower in the strain lacking FtsR (Fig. S13B, red curve) and was strongly increased when ftsR was expressed via pEC-ftsR (Fig. S13B, blue curve). Growth and specific fluorescence was comparable for the two strains with chromosomal ftsR expression (Fig. S13, black and green curves). These results confirm transcriptional activation of ftsZ expression by FtsR and that the phenotype of ftsR deletion mutants is not solely caused by reduced ftsZ expression.  a Three biological replicates of the experiment were performed and genes are listed whose average mRNA ratio was increased or decreased at least 2-fold and whose p-value was ≤0.05.   15 Figure S1. Alignment of C. glutamicum FtsR and homologous proteins of other Actinobacteria. The alignment was prepared using Clustal Omega (12) and edited using ESPript 3.0 (13). Residues shown in yellow are at least 70% identical and residues indicated in red are fully conserved.  Figure S2. Morphology of C. glutamicum ATCC13032 (A) ATCC13032ΔftsR (B) and ATCC13032 pAN6-ftsR (C). The cells were first cultivated in BHI medium followed by two consecutive cultivations in CGXII minimal medium with 2% (w/v) glucose as carbon source. For the plasmid-based overexpression (C), kanamycin (25 µg/mL) and IPTG (100 µM) were added. Fluorescence microscopy of stationary phase cells was performed. DNA was stained with Hoechst 33342 (cyan) and membranes with Nile red (red) as described in the Methods section. The scale bar is 5 μm. Figure S3. Morphology of C. glutamicum ATCC13032, ATCC13032ΔftsR, ATCC13032 pAN6, and ATCC13032 pAN6-ftsR in different growth phases. The cells were first cultivated in BHI medium followed by two consecutive cultivations in CGXII minimal medium with 2% (w/v) glucose as carbon source. For the strains carrying plasmids, kanamycin (25 µg/mL) and IPTG (100 µM) were added. Fluorescence microscopy was performed with staining of DNA with Hoechst 33342 (cyan) and membranes with Nile red (red) as described in the Methods section.  The cells were first cultivated in BHI medium followed by two consecutive cultivations in CGXII minimal medium with 2% (w/v) glucose as carbon source. For the plasmid-based (over)expression, kanamycin (25 µg/mL) and IPTG was added as indicated. Averages and standard deviations of three biological replicates of the second CGXII culture are presented. Figure S6. Complementation of the growth defect (left panel) and the morphological phenotype (right panel) of the MB001ftsR mutant by plasmid-encoded FtsR and homologous proteins of related pathogenic species. The cells were first cultivated in BHI medium with kanamycin (25 µg/mL) followed by two consecutive cultivations in CGXII minimal medium with kanamycin (25 µg/mL) and 2% (w/v) glucose as carbon source. MB001ΔftsR was transformed with pAN6 encoding ftsR, CDC7B_1201, or rv1828 under control of the leaky tac promoter. IPTG was only added for expression of rv1828, where 100 µM were required to obtain optimal complementation. Averages and standard deviations of three biological replicates of the second CGXII culture are presented. Microscopy was performed with stationary phase cells. The scale bar represents 5 μm. Figure S7. Normalized DNA levels of cg0834 and cg0840 in different ftsR deletion strains and the corresponding reference strains. Chromosomal DNA of ATCC13032, MB001, and several ΔftsR clones of the two strains was isolated and the relative copy numbers were determined. The DNA level of recF was used as reference and was set to 1. Illustrated are the normalized DNA levels of the genes cg0834 and cg0840 determined by qPCR and calculated using the Ct quantification with the program qPCRsoft 3.1.  Figure S9. Fluorescence microscopy of the C. glutamicum strains ATCC13032::ftsZ-venus and ATCC13032ΔftsR::ftsZ-venus carrying a second chromosomal copy of ftsZ fused in frame to the coding sequence of the fluorescent protein Venus. The cells were first cultivated in BHI + 2 % (w/v) glucose and afterwards transferred to CGXII + 2 % (w/v) glucose, both with 25µg/mL kanamycin. Samples for microscopy were taken after about 6 h cultivation of the main culture, which is approximately in the middle of the exponential growth phase. Two representative pictures for each culture are shown. The experiment was performed with two biological replicates each. Figure S10. DNA regions showing the highest coverage (red peaks) in the ChAP-Seq experiment with FtsR-Strep (strain C. glutamicum ΔftsR pAN6-ftsR-Strep). As negative control, the DNA enriched by StrepTactin affinity chromatography from strain C. glutamicum ΔftsR pAN6 was used (blue background). The red peak in the ftsZ promoter region (between cg2365 and cg2366), which had the 3 rd highest coverage, is depicted in Fig. 5. Due to the high peaks in the negative control of sample peaks 1 and 2, an independent ChAP-Seq experiment was performed, which confirmed binding of FtsR between cg2477 and cg2478 and between cop1 and cg3185 (see Table 1) without the high background peaks in the negative control.  Figure S13. Promoter exchange of ftsZ and growth of the resulting strains. (A) Strains with FtsRindependent ftsZ expression were constructed using a DNA fragment with a terminator sequence and the gluconate-inducible gntK promoter, which was inserted between the native ftsZ promoter and the ftsZ start codon in the chromosomes of MB001 and the MB001ftsR mutant. (B-D) The growth of the resulting promoter exchange strains was tested in comparison to the strains with the native ftsZ promoter using either sucrose (B), glucose (C), or glucose plus gluconate (C) as carbon source(s). (E) Cultivation with a gluconate/glucose ratio which enables a similar growth behavior of MB001::PgntK-ftsZ and MB001ftsR::PgntK-ftsZ. Cells were pre-cultivated either in BHI medium (B and C) or in BHI medium supplemented with 0.1% (w/v) gluconate for PgntK induction when gluconate was also used as carbon source in the main culture (D and E), followed by two consecutive cultivations in CGXII minimal medium supplemented with the indicated carbon sources. Averages and standard deviations from three biological replicates are presented. Figure S14. Influence of ftsR expression on (A) growth and (B) plasmid-based ftsZ promoter activity in the ftsZ promoter exchange strains MB001::PgntK-ftsZ and MB001ftsR::PgntK-ftsZ. The two strains were transformed with plasmid pJC1-PftsZ-venus for monitoring ftsZ promoter activity and either with the ftsR expression plasmid pEC-ftsR or the vector pEC-XC99E. Cells were cultivated first in BHI complex medium supplemented with 0.1% (w/v) gluconate followed by a second pre-culture and the main culture in CGXII minimal medium containing 0.01% (w/v) gluconate and 1.99% (w/v) glucose as carbon source. Kanamycin (25 µg/mL) and chloramphenicol (10 µg/mL) were added to all cultures. Averages and standard deviations from three biological replicates are shown.