Transcriptional activation by MafR, a global regulator of Enterococcus faecalis

Proteins that act as global transcriptional regulators play key roles in bacterial adaptation to new niches. These proteins recognize multiple DNA sites across the bacterial genome by different mechanisms. Enterococcus faecalis is able to survive in various niches of the human host, either as a commensal or as a leading cause of serious infections. Nonetheless, the regulatory pathways involved in its adaptive responses remain poorly understood. We reported previously that the MafR protein of E. faecalis causes genome-wide changes in the transcriptome. Here we demonstrate that MafR functions as a transcription activator. In vivo, MafR increased the activity of the P12294 and P11486 promoters and also the transcription levels of the two genes controlled by those promoters. These genes are predicted to encode a calcium-transporting P-type ATPase and a QueT transporter family protein, respectively. Thus, MafR could have a regulatory role in calcium homeostasis and queuosine synthesis. Furthermore, MafR recognized in vitro specific DNA sites that overlap the −35 element of each target promoter. The MafR binding sites exhibit a low sequence identity, suggesting that MafR uses a shape readout mechanism to achieve DNA-binding specificity.

Gene OG1RF_12294 encodes a putative p-type Atpase cation transporter. P-type ATPases constitute a large superfamily of cation and lipid pumps that use ATP hydrolysis for energy. They are integral, multispanning membrane proteins that are found in bacteria and in a number of eukaryotic plasma membranes and organelles 16 . The enterococcal OG1RF_12294 gene, which is adjacent to mafR (Fig. 1A), encodes a putative P-type ATPase cation transporter. Such a gene has been annotated as pmr1 (GeneID: 12289043) because it encodes a protein (850 amino acids) that has sequence similarity (∼52%) to eukaryotic PMR1 (plasma membrane ATPase related) P-type ATPases (Supplementary Table S1). Some PMR1-type pumps are able to transport calcium, as well as manganese, into the Golgi apparatus [17][18][19] .

MafR influences positively the transcription of OG1RF_12294.
To analyse whether MafR regulates the expression of the OG1RF_12294 gene, we determined its relative expression in OG1RF (wild-type) and OG1RFΔmafR (deletion mutant) by qRT-PCR. The log 2 FC in OG1RF_12294 expression due to the presence of MafR was ∼3, indicating that MafR has a positive effect on the transcription of such a gene. This conclusion was further confirmed by increasing the intracellular level of MafR. Specifically, we determined the relative expression of OG1RF_12294 in two strains: OG1RFΔmafR harbouring pDLF (absence of MafR) and OG1RFΔmafR harbouring pDLFmafR (plasmid-encoded MafR). In addition, we determined the relative expression of the OG1RF_10600 and OG1RF_11602 genes, which encode putative calcium-transporting ATPases (Supplementary  Table S1). In the presence of plasmid-encoded MafR, only transcription of OG1RF_12294 was increased (log 2 FC ∼4). Thus, MafR influences positively and specifically the transcription of the OG1RF_12294 gene.

MafR activates the P12294 promoter in vivo.
In the OG1RF genome 14 , the ATG codon at coordinate 2425611 is likely the translation start site of the OG1RF_12294 gene (Fig. 1A). It is preceded by a putative ribosome binding site sequence (AGGAGG). Upstream of such a sequence there is a putative promoter (here named P12294) that has a canonical −10 element (TATAAT) but lacks a potential −35 element (consensus TTGACA) at the optimal length of 17 nucleotides. Nevertheless, there is a near-consensus −35 element (TCGACC) at the suboptimal spacer length of 22 nucleotides. These features suggested that promoter P12294 could be recognized by a σ factor similar to the Escherichia coli σ 70 and that its activity could be enhanced by regulatory proteins. Sequence analysis of the region located between the TAA stop codon of the OG1RF_12295 gene (coordinate 2425761) and the P12294 promoter revealed the existence of an inverted-repeat (IR) that may function as a Rho-independent transcriptional terminator (Fig. 1A).
To characterize the P12294 promoter, a 255-bp DNA fragment (coordinates 2425885 to 2425631) ( Fig. 2) was inserted into the pASTT promoter-probe vector, which is based on the gfp reporter gene. The recombinant plasmid (pASTT-P12294) was first introduced into OG1RF and OG1RFΔmafR. In these strains, the expression of gfp (0.32 ± 0.02 and 0.26 ± 0.04 units, respectively) was similar to the basal level (OG1RF harbouring pASTT; 0.38 ± 0.02 units). Different results were obtained when pASTT-P12294 was introduced into OG1RFΔmafR harbouring either pDLF or pDLFmafR (plasmid-encoded MafR) (Fig. 2). The expression of gfp was ∼2.5-fold higher in the presence of plasmid-encoded MafR. This result indicated that the 255-bp DNA fragment contains a MafR-dependent promoter activity. Removal of the −10 element of the P12294 promoter resulted in loss of such  www.nature.com/scientificreports www.nature.com/scientificreports/ an activity (plasmid pASTT-P12294Δ-10). A further deletion analysis allowed us to conclude that the 186-bp region between coordinates 2425816 and 2425631 contains both the P12294 promoter and the site required for its activation by MafR (plasmids pASTT-P12294Δ69 and pASTT-P12294Δ208) (Fig. 2).

MafR binds to the P12294 promoter region in vitro.
To investigate whether MafR activates directly the expression of the OG1RF_12294 gene, we performed DNase I footprinting experiments. We used a His-tagged MafR protein (MafR-His) and a 270-bp DNA fragment (coordinates 2425817 to 2425548). This fragment contains the P12294 promoter and the site required for its activation by MafR in vivo (Fig. 2). The presence of a His-tag at the C-terminal end of MafR does not affect its DNA-binding properties 4 . The 270-bp DNA fragment was radioactively labelled either at the 5′-end of the coding strand or at the 5′-end of the non-coding strand (Fig. 3). On the coding strand and at 100 nM of MafR-His, protections against DNase I digestion were observed within the region spanning coordinates 2425708 and 2425658. On the non-coding strand and at 125 nM of MafR-His, diminished cleavages were observed between coordinates 2425712 and 2425686. Thus, MafR-His recognizes a site overlapping the −35 element of the P12294 promoter (Fig. 3). This result allowed us to conclude that MafR activates directly the transcription of the OG1RF_12294 gene. Figure 1B shows the bendability/curvature propensity plot of the 270-bp DNA fragment according to the bend.it program 26 . The profile contains an intrinsic curvature of high magnitude (~13 degrees per helical turn), which is adjacent to the MafR binding site. In addition, the site recognized by MafR contains a region of potential bendability (~5.2 units). www.nature.com/scientificreports www.nature.com/scientificreports/ Gene OG1RF_11486 encodes a putative Quet transporter family protein. Energy-coupling factor (ECF) transporters are a family of ATP-binding cassette (ABC) transporters that are responsible for the uptake of essential micronutrients in prokaryotes. They consist of a membrane-embedded S-component that provides substrate specificity and a three-subunit ECF module that couples ATP hydrolysis to transport. In the so-called group II ECF transporters, different S-components share the same ECF module. Furthermore, the S-component genes are not located in the same operon as the genes for the ECF module [27][28][29] .
The enterococcal OG1RF_11486 gene encodes a putative QueT transporter family protein (GenBank AEA94173.1). Proteins identical to OG1RF_11486 (173 residues) are encoded by Mycobacterium abscessus (CPW17925.1), Listeria monocytogenes (CWW42654.1; 172 up to 173 residues are identical) and S. agalactiae (KLL29182.1). In the two former bacteria, the corresponding protein has been annotated as queuosine precursor ECF transporter S-component QueT. Therefore, protein OG1RF_11486 could be involved in the uptake of a queuosine biosynthetic intermediate. Using the BLASTP program 20 , we found that the OG1RF genome encodes an additional QueT transporter family protein (OG1RF_12031; 168 residues; AEA94718.1). It has 55% of similarity to the OG1RF_11486 protein.
MafR activates the P11486 promoter in vivo. By qRT-PCR assays, we found that MafR has a positive effect on the transcription of OG1RF_11486. Compared to strain OG1RFΔmafR, the relative expression of OG1RF_11486 was slightly higher in strain OG1RF (log 2 FC ∼0.9). Moreover, the relative expression of OG1RF_11486 was higher in strain OG1RFΔmafR harbouring pDLFmafR (plasmid-encoded MafR) than in strain OG1RFΔmafR harbouring pDLF (log 2 FC ∼2.4).
The BPROM program (Softberry, Inc.) predicts a promoter sequence (named P11486 herein) upstream of the OG1RF_11486 gene. The −35 (TTTACA) and −10 (TAACAT) elements of this promoter are separated by 17 nucleotides (Fig. 4A). By primer extension using total RNA from OG1RF cells, we demonstrated that the P11486 promoter is functional in vivo (Fig. 5). Oligonucleotide R11486-D was used as primer (Table 1). A cDNA product of 130 nucleotides was detected, indicating that transcription of OG1RF_11486 starts at coordinate 1543115 (Fig. 4A).
To further characterize the P11486 promoter, we constructed several transcriptional fusions (Fig. 6). A 284-bp DNA fragment (coordinates 1542902 to 1543185) was inserted into pASTT. The recombinant plasmid (pASTT-P11486) was first introduced into OG1RF and OG1RFΔmafR. In both strains, gfp expression (1.48 ± 0.10 and 1.51 ± 0.16 units, respectively) was ∼4-fold higher than the basal level (OG1RF harbouring pASTT). This result indicated that the 284-bp DNA fragment has promoter activity, however, the chromosomal copy of mafR is not sufficient to activate such a promoter located on pASTT (multicopy plasmid). Next, we introduced pASTT-P11486 into OG1RFΔmafR harbouring pDLFmafR (plasmid-encoded MafR). In this strain, gfp expression was ∼3-fold higher than in the control strain (OG1RFΔmafR harbouring pDLF) (Fig. 6). Similar results were obtained with plasmids pASTT-P11486Δ66 and pASTT-P11486Δ145, which allowed us to conclude that the 139-bp region between coordinates 1543047 and 1543185 contains both the P11486 promoter and the site required for its activation by MafR. A further deletion analysis showed that sequences between coordinates 1543047 and 1543071 (plasmid pASTT-P11486Δ169) are needed for MafR-mediated activation of the P11486 promoter but not for promoter activity. Moreover, deletion of the region that spans coordinates 1543071 and 1543090 (plasmid pASTT-P11486Δ188) removes the −35 element of the P11486 promoter and, consequently, reduces the expression of gfp to basal levels (Fig. 6).
MafR binds to the P11486 promoter region in vitro. By DNase I footprinting assays, we analysed whether MafR-His binds to the P11486 promoter region (Fig. 7). We used a 275-bp DNA fragment (coordinates 1542969 to 1543243), which contains both the P11486 promoter and the site required for its activation by MafR in vivo (Fig. 6). On the coding strand and at 350 nM of MafR-His, changes in DNase I sensitivity (diminished cleavages) were observed within the region spanning coordinates 1543047 and 1543110. On the non-coding strand and at 300 nM of MafR-His, diminished cleavages were observed between coordinates 1543043 and 1543110. On both strands and at 400 nM of MafR-His, regions protected against DNase I digestion were observed along the DNA fragment, which is consistent with the ability of MafR-His to generate multimeric complexes 4 . Therefore, MafR-His recognizes preferentially a DNA site overlapping the P11486 core promoter. Such a DNA site includes sequences needed for MafR-mediated activation of the P11486 promoter in vivo (Fig. 6). According to the bendability/curvature propensity plot of the 275-bp DNA fragment, the MafR binding site contains regions of potential bendability (Fig. 4B).

Discussion
Gene regulation plays a key role during bacterial adaptation to environmental fluctuations. The ability of enterococci to metabolize numerous carbohydrates enables them to colonize diverse environments 1 . Our previous work showed that MafR activates, directly or indirectly, the transcription of numerous genes on a genome-wide scale. Many of such genes encode proteins involved in transport or metabolism of carbon sources 5 . Now, by qRT-PCR, transcriptional fusions and DNase I footprinting experiments, we have demonstrated that MafR functions as a transcription activator. It activates directly the transcription of the OG1RF_12294 and OG1RF_11486 genes. Gene OG1RF_12294 encodes a protein that has sequence similarity to several eukaryotic and prokaryotic proteins characterized as calcium P-type ATPases (Supplementary Table S1). This finding suggests that MafR could have a regulatory role in maintaining cellular calcium homeostasis. Calcium ions are known to affect different physiological processes in prokaryotic organisms, such as division, secretion, transport, and stress response 30 . Gene OG1RF_11486 encodes a putative ECF transporter S-component, likely involved in the uptake of a queuosine precursor. Thus, MafR could have an additional regulatory role in the biosynthesis of queuosine, a modified nucleoside found at the wobble position of particular transfer RNAs 31 . There is evidence that queuosine contributes www.nature.com/scientificreports www.nature.com/scientificreports/ to the efficiency of protein synthesis. In Shigella flexneri, the intracellular concentration of the virulence-related transcriptional regulator VirF is reduced in the absence of queuosine 32 . Moreover, it has been reported that the lack of queuosine affects the growth of some bacteria under stress conditions 33,34 .
Bacteria use a variety of mechanisms to activate transcription from specific promoters. Genetic and biochemical studies have shown that some proteins stimulate transcription by binding to a specific DNA site either upstream of or overlapping the core promoter 35 . By DNase I footprinting experiments, we have found that MafR recognizes a site overlapping the P12294 core promoter, as well as a site overlapping the P11486 core promoter (this work). These results suggest that MafR might enhance the efficiency of both promoters by recruitment of RNA polymerase through direct interactions with the sigma factor. In addition, MafR might induce conformational changes in the target promoters, as it has been described for some transcription activators 35 . Transcriptional activation from specific promoters has also been reported for other members of the Mga/AtxA family. The pneumococcal MgaSpn regulator stimulates transcription of a four-gene operon (spr1623-spr1626) by binding to a specific DNA site upstream of the promoter (positions −60 to −99) 12 . Regarding the Mga regulator from S. pyogenes, the position of its DNA-binding site with respect to the start of transcription varies among the promoters tested. Nevertheless, the majority of the promoters contain an Mga binding site located around position −54, thereby overlapping the −35 element of the promoter 8 .
Simple protein-DNA recognition mechanisms do not exist 36 . Based on the structures of various protein-DNA complexes, Rohs et al. proposed that particular proteins use likely a combination of readout mechanisms: base readout and shape readout 6 . The DNA sites recognized by MafR on the P12294 and P11486 promoters have a low sequence identity: they share the GG( (Fig. 8A). Moreover, both MafR binding sites contain regions of potential bendability (Figs 1B and 4B). We have also shown that MafR recognizes a DNA site upstream of the Pma promoter (positions −69 to −104) 4 . The function of this interaction remains unknown. Such a MafR binding site is adjacent to the peak of a potential intrinsic curvature 4 and shares a short DNA sequence motif (TGATAT) with the two MafR binding sites identified in this work (Fig. 8B). www.nature.com/scientificreports www.nature.com/scientificreports/ Therefore, MafR does not seem to recognize a specific nucleotide sequence. Several findings suggest that recognition of particular DNA shapes could be a characteristic of the global regulators that constitute the Mga/AtxA family. MgaSpn from S. pneumoniae recognizes a DNA site upstream of the P1623B promoter (positions −60 to −99), as well as a DNA site overlapping the Pmga promoter (positions −23 to +21) 12 . The former interaction enhances the efficiency of the promoter 11 , whereas the function of the latter remains unknown. Such MgaSpn binding sites have a low sequence identity and, according to predictions, they contain an intrinsic curvature flanked by regions of bendability 12 . Furthermore, MgaSpn was shown to have a preference for AT-rich DNA regions 13 . Concerning Mga from S. pyogenes, several DNA-binding sites have been identified. These sites exhibit a low sequence identity (13.4%) 37 , although a consensus Mga binding sequence was initially proposed 38 . In the case of AtxA from B. anthracis, in vitro protein-DNA interaction studies have not been reported. Nevertheless, sequence similarities are not apparent in its target promoter regions, and some of them are intrinsically curved 39 .

C/A)C(A/C)(C/A)TGAAAT(T/A)A sequence element
In conclusion, our study shows for the first time that MafR is a transcription activator. It stimulates transcription from the P12294 and P11486 promoters in vivo. Moreover, MafR binds in vitro to a specific DNA site that overlaps the −35 element of each promoter. The two MafR binding sites have a low sequence identity but share a six-base pair motif. We propose that MafR would recognize intrinsic DNA structural features rather than particular DNA sequences on its target DNAs.
Growth and transformation of bacteria. E. faecalis was grown in BHI medium, which was supplemented with tetracycline (4 μg/ml) and/or with kanamycin (250 μg/ml) when strains carrying plasmids were used. Experiments were performed at 37 °C without aeration. The protocol used to transform E. faecalis by electroporation was described 41 . DNA and RNA isolation. Genomic DNA was prepared using the Bacterial Genomic Isolation Kit (Norgen Biotek Corporation). Plasmid DNA was prepared using the High Pure Plasmid Isolation Kit (Roche Applied Science) as described 5 . Total RNA was isolated using the RNeasy mini Kit (QIAGEN). In general, bacteria were grown to an optical density at 650 nm (OD 650 ) of 0.4 (logarithmic growth phase). For stationary phase, bacteria were grown to an OD 650 of 0.8 and then incubated for two hours at the same temperature. Then, cultures were processed as reported 5 . The integrity of rRNAs was analysed by agarose gel electrophoresis. RNA concentration was determined using a NanoDrop ND-2000 Spectrophotometer. polymerase chain reaction (pCR). The Phusion High-Fidelity DNA polymerase (Thermo Scientific) and the Phusion HF buffer were used. Reaction mixtures (50 μl) contained 5-30 ng of template DNA, 20 pmol of each primer, 200 μM each deoxynucleoside triphosphate (dNTP), and one unit of DNA polymerase. PCR conditions were reported 40 . To amplify the 270-bp DNA fragment (promoter P12294) used in footprinting experiments, the Phusion GC buffer was used. In this case, reaction mixtures were supplemented with 7% DMSO and the annealing step was performed at 59 °C. PCR products were purified with the QIAquick PCR purification kit (QIAGEN).
Quantitative Rt-pCR (qRt-pCR). For cDNA synthesis with random primers, the iScript Select cDNA Synthesis kit (Bio-Rad) was used as described 5 . Quantitative PCRs were performed using the iQ SYBR Green Supermix (Bio-Rad) and a iCycler Thermal Cycler (Bio-Rad) as reported 5 . Forward (Fgene-q) and reverse (Rgene-q) primers used in the quantitative PCRs are listed in Table 1. Relative quantification of gene expression was performed using the comparative C T method 15 as described 5 . Except for gene mafR, the internal control gene was recA (OG1RF_12439; recombination protein RecA). In the case of mafR, the internal control gene was zwf (OG1RF_10737; glucose-6-phosphate 1-dehydrogenase) because its expression level was similar at the logarithmic and stationary growth phases. primer extension. Oligonucleotide R11486-D was radioactively labelled at the 5′-end using [γ-32 P]-ATP (PerkinElmer) and T4 polynucleotide kinase (New England Biolabs) as reported 12 . Primer extension reactions (20 μl) contained 1.2 pmol of 32 P-labelled oligonucleotide and 5 μg of total RNA isolated from strain OG1RF. The ThermoScript Reverse Transcriptase enzyme (Invitrogen) was used. Reactions were incubated at 55 °C for 45 min. After heating at 85 °C for 5 min, samples were ethanol precipitated and dissolved in loading buffer (80% formamide, 1 mM EDTA, 10 mM NaOH, 0.1% bromophenol blue, 0.1% xylene cyanol). cDNA products were analysed by sequencing gel (8 M urea, 6% polyacrylamide) electrophoresis. Dideoxy-mediated chain termination sequencing reactions were run in the same gel. Labelled products were visualized using a Fujifilm Image Analyser FLA-3000.
Fluorescence assays. Plasmid-carrying cells were grown to an OD 650 of 0.4 (logarithmic phase).
Then, different volumes of culture (0.4 to 1 ml) were centrifuged, and cells were resuspended in 200 μl of  www.nature.com/scientificreports www.nature.com/scientificreports/ phosphate-buffered saline (PBS). In each case, three independent cultures were analysed. Fluorescence intensity was measured using a Thermo Scientific Varioskan Flash instrument (excitation at 488 nm and emission at 515 nm). The fluorescence corresponding to 200 μl of PBS buffer without cells was ~0.03 arbitrary units. purification of MafR-His. The procedure to overproduce and purify a His-tagged MafR OG1RF protein (herein MafR-His) was reported 4 . MafR-His carries the Leu-Glu-6xHis peptide (His-tag) fused to its C terminus. Protein concentration was determined using a NanoDrop ND-2000 Spectrophotometer (Thermo Scientific).
DNase I footprinting assays. Oligonucleotides were 32 P-labelled at the 5′-end as described 12 . 32 P-labelled oligonucleotides were used for PCR amplification to obtain double-stranded DNA fragments labelled at either the coding or the non-coding strand. Two regions of the OG1RF chromosome were amplified: a 270-bp region (coordinates 2425817-2425548) using the F12294-D and R12294-D oligonucleotides, and a 275-bp region (coordinates 1542969-1543243) using the F11486-D and R11486-D oligonucleotides. Binding reactions (8 μl) contained 30 mM Tris-HCl, pH 7.6, 1 mM DTT, 1 mg/ml BSA, 1.25% glycerol, 0.25 mM EDTA, 50 mM NaCl, 10 mM MgCl 2 , 1 mM CaCl 2 , 2-4 nM 32 P-labelled DNA and different concentrations of MafR-His (100 to 600 nM). Reaction mixtures were incubated at room temperature for 20 min. Then, 0.015 units of DNase I (Roche Applied Science) was added and the reaction proceeded for 5 min at the same temperature. DNase I digestion was stopped by adding 1 μl of 250 mM EDTA. Then, 4 μl of loading buffer (80% formamide, 1 mM EDTA, 10 mM NaOH, 0.1% bromophenol blue and 0.1% xylene cyanol) was added. Samples were heated at 95 °C for 5 min and loaded onto sequencing gels (6% polyacrylamide, 8 M urea). Dideoxy-mediated chain termination sequencing reactions were run in the same gel. Labelled products were visualized using a Fujifilm Image Analyser FLA-3000. The intensity of the bands was quantified using the Quantity One software (Bio-Rad).