DdlR, an essential transcriptional regulator of peptidoglycan biosynthesis in Clostridioides difficile

SUMMARY

D-Ala-D-Ala ligase, encoded by ddl genes, is responsible for the synthesis of a dipeptide, D-Ala-D-Ala, an essential precursor of bacterial peptidoglycan. In Clostridioides difficile, the single ddl gene is located upstream of the ddlR gene, which encodes a putative transcriptional regulator. Using mutational and transcriptional analysis and DNA-binding assays, DdlR was found to be a direct activator of the ddl ddlR operon. DdlR is a member of the MocR/GabR-type proteins that have aminotransferase-like, pyridoxal 5’-phosphate-binding domains. A DdlR mutation that prevented covalent binding of pyridoxal 5’-phosphate abolished the ability of DdlR to activate transcription. Addition of D-Ala-D-Ala to the medium inactivated DdlR, reducing dipeptide biosynthesis. In contrast, D-Ala-D-Ala limitation caused a dramatic increase in expression from the ddl promoter. Though uncommon for transcription regulators, C. difficile DdlR is essential, as the ddlR null mutant cells could not grow even in complex laboratory media in the absence of D-Ala-D-Ala. A dyad-symmetry sequence, which is located immediately upstream of the −35 region of the ddl promoter, serves as an important element of the DdlR-binding site. This sequence is conserved upstream of putative DdlR targets in other bacteria of classes Clostridia and Bacilli, indicating a similar mode of regulation of these genes.

Abbreviated Summary

The Clostridioides difficile D-Ala-D-Ala ligase is an essential metabolic enzyme that is encoded by the ddl gene. Expression of the ddl gene is directly and strongly activated by DdlR, a transcriptional regulator that responds to the availability of the dipeptide, D-Ala-D-Ala, or a derivative and requires pyridoxal 5’-phosphate, the active form of vitamin B₆, for its activity. DdlR is required for growth even in complex laboratory media, if D-Ala-D-Ala is not provided.

INTRODUCTION

Peptidoglycan is an essential component of the bacterial cell wall and is composed of glycan chains of alternating residues of N-acetylglucosamine and N-acetylmuramic acid to which crosslinked pentapeptide stems are attached (Scheffers and Pinho, 2005; Barreteau et al., 2008; Vollmer et al., 2008). The sequence of the pentapeptide unit, e.g., L-Ala - D-Glu - meso-diaminopimelic acid (DAP) - D-Ala - D-Ala, varies in different bacterial species, but the two terminal D-Ala residues are invariant. The D-Ala amino acids are incorporated as a single dipeptide unit, which is synthesized by D-Ala-D-Ala ligase, the product of the ddl gene(s). Some bacteria, like the model Gram-positive bacterium Bacillus subtilis, have a single ddl gene, which is essential for growth (Buxton and Ward, 1980; Kobayashi et al., 2003). Other bacteria, like the model Gram-negative bacterium Escherichia coli, have two ddl genes, ddlA and ddlB; inactivation of only one of these genes does not prevent growth, but a double mutant requires D-Ala-D-Ala for growth (Zawadzke et al., 1991; Flardh et al., 1998; McCoy and Maurelli, 2005). Importantly, D-Ala-D-Ala is not present even in complex laboratory media and is likely to be a very scarce metabolite in nature. Moreover, some bacteria, like B. subtilis, apparently cannot take up D-Ala-D-Ala from the environment (Buxton and Ward, 1980; Takenaka et al., 2015). Therefore, D-Ala-D-Ala ligase is generally considered to be an essential enzyme and a promising drug target (Tytgat et al., 2009). In fact, inhibition of D-Ala-D-Ala ligase activity serves as the basis for the antibacterial activity of D-cycloserine against Mycobacterium tuberculosis (Prosser and de Carvalho, 2013; Halouska et al., 2014).

Little is known about the regulation of ddl genes. Recently, it was reported that the Brevibacillus brevis ddl (ddlB) gene, one of the two ddl genes in this bacterium, is positively regulated by the product of the upstream gene, ddlR (Takenaka et al., 2015). B. brevis DdlR belongs to the undercharacterized but widespread MocR/GabR subfamily of the GntR family of transcriptional regulators. The members of this subfamily are composed of a short N-terminal helix-turn-helix-containing domain (a DNA-binding region) and a long C-terminal aminotransferase-like domain (Belitsky and Sonenshein, 2002; Rigali et al., 2002; Edayathumangalam et al., 2013; Milano et al., 2015; Tramonti et al., 2018).

Aminotransferases are ubiquitous enzymes of nitrogen metabolism that require pyridoxal 5’-phosphate (PLP), the biologically active form of vitamin B₆, as an essential covalently bound cofactor for catalysis (Schneider et al., 2000; Eliot and Kirsch, 2004). In MocR/GabR-like proteins, the aminotransferase-like domain apparently serves as an effector-binding domain that may have two distinct roles. In some proteins that are involved in the transcriptional regulation of genes of vitamin B₆ synthesis and usually named PdxR, the aminotransferase-like domain serves for interaction with PLP as a simple effector (Magarvey et al., 2001; Jochmann et al., 2011; El Qaidi et al., 2013; Belitsky, 2014; Liao et al., 2015; Tramonti et al., 2015; Tramonti et al., 2017). In other MocR/GabR-like proteins, the aminotransferase-like domain plays a more complex role and promotes a covalent interaction between PLP and an amino group-containing compound that is required for transcriptional regulation. PLP may be considered as a co-factor of the protein similar to its role in PLP-dependent enzymes, and an amino group-containing compound may serve as an effector. This mechanism has been studied in detail for B. subtilis GabR, a positive transcriptional regulator of γ-aminobutyrate (GABA) utilization (Belitsky and Sonenshein, 2002; Belitsky, 2004; Okuda, Ito et al., 2015; Okuda, Kato et al., 2015; Al-Zyoud et al., 2016). Determination of GabR’s crystal structure (Edayathumangalam et al., 2013; Okuda, Ito et al., 2015) revealed that the structural fold of the C-terminal domain of GabR is indeed very similar to the fold type I of aminotransferases and other PLP-binding proteins (Grishin et al., 1995) and that PLP appears to be a covalently-linked component of the protein. The interaction between GABA’s amino group and PLP’s aldehyde group results in the formation of an external aldimine and apparent conformational change of the protein that leads to its ability to activate transcription (Belitsky, 2004; Okuda, Kato et al., 2015; Wu et al., 2017).

The MocR/GabR subfamily includes at present thousands of members from both Gram-negative and Gram-positive bacteria (Bramucci et al., 2011; Suvorova and Rodionov, 2016), but only a handful of family members, other than GabR and PdxRs, have been characterized functionally and shown to be DNA-binding proteins that activate or repress transcription of their target genes (Wiethaus et al., 2008; Truong-Bolduc et al., 2011; Mostafavi et al., 2014; Takenaka et al., 2015; Schulz, Hermann et al., 2017; Schulz, Stoveken et al., 2017; Tramonti et al., 2017; Yu et al., 2017; Tramonti et al., 2018).

To characterize further the mechanism by which MocR/GabR-like proteins are involved in the regulation of peptidoglycan synthesis, we studied a putative DdlR protein of Clostridioides (formerly Clostridium) difficile that is similar in sequence to B. brevis DdlR (Takenaka et al., 2015). C. difficile is a Gram-positive, anaerobic bacterium and an important nosocomial pathogen that can cause severe disease in humans (Loo et al., 2005; Ricciardi et al., 2007; Ong et al., 2017). In this work, we showed that C. difficile DdlR is an efficient transcriptional activator of the ddl gene. D-Ala-D-Ala serves directly or via its derivative as an anti-activator of DdlR, decreasing or increasing its own synthesis when present in excess or in limiting amounts, respectively. Most regulatory proteins are not essential, but DdlR is absolutely required for growth in the absence of D-Ala-D-Ala even in complex laboratory media.

RESULTS

Phenotype of the ddlR null mutant.

The 299-amino acid product of the C. difficile ddl gene (CD630_14080; CD630DERM_RS07920) (Sebaihia et al., 2006; Monot et al., 2011; van Eijk et al., 2015) is similar (30 to 36% identity, respectively) to the products of the ddlA and ddlB genes of E. coli (Zawadzke et al., 1991); it is less similar (27% identity) to the ddl gene of B. subtilis (Barbe et al., 2009) and more similar (46% identity) to B. brevis DdlB (Takenaka et al., 2015). All these genes encode D-Ala-D-Ala ligase, an enzyme responsible for the synthesis of D-Ala-D-Ala, a dipeptide that is required for peptidoglycan biosynthesis (Neuhaus, 1962). A putative transcriptional regulator of the MocR/GabR subfamily (Belitsky and Sonenshein, 2002; Rigali et al., 2002; Tramonti et al., 2018) is encoded by the CD630_14070 (CD630DERM_RS07915) gene that is located downstream of the C. difficile ddl gene and named here as ddlR (Fig. 1). The open reading frames of the ddl and ddlR genes are separated by only 26 bp (Sebaihia et al., 2006; Monot et al., 2011), and they likely form a single transcriptional unit (see below). The ddlR gene encodes a 482-amino acid protein, which contains, as do other members of the MocR/GabR subfamily, an N-terminal putative DNA-binding domain (about 80 aa) and a C-terminal aminotransferase-like domain (about 360 aa) connected by a linker. Using TargeTron technology (Heap et al., 2007; Heap et al., 2010; Bouillaut et al., 2011), we inactivated the ddlR gene of C. difficile 630∆erm by the insertion of an ermB gene-containing group II intron as described in Experimental procedures. The resulting strain, LB-CD28, required D-Ala-D-Ala for growth in TY and BHIS complex media and CDMM defined medium (Fig. 2A and data not shown). Addition of D-Ala-D-Ala at a concentration of ≥2.5 µg/ml fully restored growth of the ddlR null mutant strain (Fig. 2A and 2C) showing that C. difficile cells are able to take up D-Ala-D-Ala similarly to Escherichia coli cells (McCoy and Maurelli, 2005). Some lysis of mutant cells was observed after lower concentrations of D-Ala-D-Ala were exhausted in the medium (Fig. 2A). The D-Ala-D-Ala requirement of the ddlR null mutant was abolished after introduction of an intact copy of the ddlR gene expressed from the ddl promoter on a multicopy plasmid, pBB1864 (Fig. 2D, strain BCD17). Thus, the product of the ddlR gene is required for D-Ala-D-Ala synthesis under the conditions tested, suggesting that it acts as a positive regulator of ddl transcription. C. difficile DdlR is 48% identical to B. brevis DdlR, which was shown to be an activator of the ddlR-associated ddlB gene (Takenaka et al., 2015).

Fig. 1.

Organization of the ddl ddlR operon and the sequence of the ddl and ddlR regulatory regions. The likely initiation codons, the ddl termination codon, the −10 and −35 promoter regions, and the +1 gene positions are in boldface. The directions of transcription and translation are indicated by the arrows. The sequence protected by DdlR on the template strand in DNase I footprinting experiments is underlined. The 33-bp dyad-symmetry sequence is in boldface and italicized. The coordinates of the 5’ and 3’ ends of the sequence with respect to the +1 position of the ddl gene and the locations of the p1 and p2 mutations are indicated.

Fig. 2.

Effect of D-Ala-D-Ala on growth of the C. difficile parent strain and ddlR and ddl mutants. Cells were grown overnight in TY complex medium with or without D-Ala-D-Ala (DADA) (20 µg/ml) and diluted 100-fold in the same medium without D-Ala-D-Ala or with indicated concentrations of D-Ala-D-Ala. 630 - strain 630∆erm; ddlR - strain LB-CD28, ddl - strain BCD27.

The C. difficile genome contains only one gene (ddl) encoding D-Ala-D-Ala ligase. A global transposon mutagenesis analysis of C. difficile strain R20291 revealed, as expected, that the ddl gene is essential, reflecting the absence of D-Ala-D-Ala in laboratory media (Dembek et al., 2015)(see below). In the same study and in accord with our results, the CDR20291_1254 (ddlR) gene was also found to be essential in strain R20291 (Dembek et al., 2015).

Transcriptional regulation of ddl and ddlR genes.

As determined by real-time RT-PCR experiments, ddl expression in the parent strain 630∆erm was moderately (5-fold) decreased in the complex TY medium in the presence of high concentrations of D-Ala-D-Ala (40 to 200 µg/ml), but was not significantly affected by D-Ala-D-Ala at a low concentration (2.5 µg/ml) (Table 1 and data not shown). In a ddlR null mutant strain, which cannot grow without supplementation of D-Ala-D-Ala in the growth medium, expression of the ddl gene was low in the presence of both low and high concentrations of D-Ala-D-Ala (Table 1, strain LB-CD28). The 8-fold different levels of expression of the ddl gene in the parent and ddlR null mutant strains in the presence of a low concentration of D-Ala-D-Ala confirmed that DdlR is an activator of ddl expression. We also conclude that D-Ala-D-Ala or a related compound, e.g., D-Ala-D-Ala derivative, reduces activity of DdlR. Also, our data indicate that even in the absence of DdlR we observe some expression from the ddl promoter.

Table 1.

Expression of the ddl and ddlR genes in C. difficile as determined by real-time RT-PCR

Strain	Relevant genotype	Additions to the medium (μg/ml)	ddl expression	ddlR expression
630∆erm	wild type	none	22.9±30%†	47.8±29%
		D-Ala-D-Ala, 2.5	31.1±20%†	28.9±17%
			39.3±40%‡	28.9±17%
		D-Ala-D-Ala, 40	4.70±47%†	26.6±63%
LB-CD28	ddlR::ermB	none	NG§	NG
		D-Ala-D-Ala, 2.5	2.82±26%†	NA¶
		D-Ala-D-Ala, 40	2.79±32%†	NA
BCD16	pBL58 Pddl-ddlR	none	29.2±28%†	190.8±66%
BCD17	ddlR::ermB	none	23.8±50%†	220.7±19%
	pBL58 Pddl-ddlR
BCD27	∆ddl	none	NG	NG
		D-Ala-D-Ala, 2.5	1,232±11%‡	529.2±8%
		D-Ala-D-Ala, 40	4.71±11%‡	18.4±39%
BCD36	ddlR1	none	NG	NG
		D-Ala-D-Ala, 2.5	4.12±2%†	25.0±34%

Cells were grown in TY medium with or without D-Ala-D-Ala, as indicated. Thiamphenicol was added for growth of plasmid-containing strains. The data are presented as the number of copies of the ddl or ddlR transcript per copy of the rpoC transcript as determined by real-time RT-PCR. All values are averages of at least three experiments. Relative standard deviations are shown as percentage of each value.

^†PCR primers oWN101 and oWN102 were used to detect ddl expression.

^‡A different pair of ddl PCR primers, oWN127 and oWN128, was used to allow detection of the remaining part of the ddl transcript in strain BCD27 ∆ddl and comparison with the parent strain.

^§NG – no growth.

^¶NA - the transcript abundance could not be determined due to the presence of an intron insertion in the corresponding gene.

In the parent strain, the level of ddlR expression was similar to that of ddl. However, the addition of D-Ala-D-Ala at a high concentration (40 µg/ml) caused only a 1.5-fold reduction in expression of ddlR (Table 1). The differential response (5-fold vs 1.5-fold) of ddl and ddlR transcription to the presence of D-Ala-D-Ala in the parent strain suggests that ddlR is expressed at least partly independently of the upstream ddl gene (see below). As the ddlR gene was altered by the intron insertion, we did not test the autoregulation of ddlR in the ddlR null mutant strain by real-time RT-PCR (see below).

The presence of the ddlR gene on a multicopy plasmid under the control of the ddl promoter did not affect ddl expression in a ddlR⁺ strain but restored ddl expression in the ddlR null mutant strain. Expression of the ddlR gene in these strains was elevated, as expected, due to the presence of multiple copies of the ddlR gene (Table 1, strains BCD16 and BCD17). ddl expression was not affected by the presence in cells of the empty vector plasmid or the plasmid containing just the ddl promoter but not the ddlR gene (data not shown).

Transcriptional regulation of a ddl-gusA fusion.

To confirm the results obtained by real-time RT-PCR experiments, a ddl-gusA transcriptional fusion containing the entire ddl regulatory region was constructed and introduced into C. difficile cells on a multicopy plasmid, pRPF185 (Fagan and Fairweather, 2011). Expression of the ddl-gusA fusion was readily detected in the parent strain in TY medium and was decreased about 7-fold in the presence of high concentrations (40 to 200 µg/ml) of D-Ala-D-Ala. The addition of D-Ala-D-Ala at a low concentration (2.5 µg/ml) had only a small (1.8-fold) effect on ddl-gusA expression (Table 2, strain BCD4). In the ddlR null mutant strain, very weak expression of the ddl-gusA fusion (16-fold less than in the parent strain) was found even in the presence of a low concentration of D-Ala-D-Ala; this expression level was not further reduced by increasing the D-Ala-D-Ala concentration (Table 2, strain BCD5). The results are consistent with the data obtained by real-time RT-PCR analysis of RNA isolated from C. difficile cells and the notion that DdlR is a positive regulator of ddl transcription and that D-Ala-D-Ala or a related compound acts as an anti-activator of DdlR.

Table 2.

Expression of the ddl-gusA fusion in C. difficile cells

Strain	Fusion type	Genotype (ddlR allele)	Additions to the medium (µg/ml)	β-glucuronidase activity (ddl-gusA)
BCD6	none†	wild type	none	0.08±104%
BCD4	ddlp+	wild type	none	266.0±20%
			D-Ala-D-Ala, 2.5	151.4±13%
			D-Ala-D-Ala, 40	39.9±34%
			D-Ala-D-Ala, 200	35.3±17%
			D-Ala, 1,000	296.5±18%
			Gly-Gly, 200	282.3±9%
			Pyridoxal, 100 µM	253.1±1%
BCD5	ddlp+	ddlR::ermB	none	NG
			D-Ala-D-Ala, 2.5	9.55±26%
			D-Ala-D-Ala, 40	11.5±5%
BCD77	ddlp+	∆ddl	none	NG
			D-Ala-D-Ala, 2.5	38,345±14%
			D-Ala-D-Ala, 40	86.7±11%
BCD78	ddlp+	∆ddl	none	224.6±38%
		pyrE::ddl	D-Ala-D-Ala, 40	50.9±27%
BCD43	ddlp+	ddlR1	none	NG
			D-Ala-D-Ala, 2.5	19.7±4%
BCD7	ddlp1	wild type	none	11.5±6%
BCD8	ddlp2	wild type	none	5.17±18%
BCD9	ddlp1/p2	wild type	none	5.60±19%
BCD15	ddlp1/p2	ddlR::ermB	D-Ala-D-Ala, 2.5	5.83±6%

Cells were grown in TY medium plus thiamphenicol with or without D-Ala-D-Ala or other additions, as indicated. β-Glucuronidase specific activity was assayed and expressed in Miller units. All values are averages of at least two experiments. Relative standard deviations are shown as percentage of each value.

^†Strain BCD6 contains a derivative of pRPF185 with a promoterless gusA gene. The other strains have the indicated promoter regions upstream of gusA.

No effects of D-Ala, a precursor of D-Ala-D-Ala, or another dipeptide, Gly-Gly, or pyridoxal, a PLP precursor, on ddl expression was detected (Table 2).

Transcription start points of the ddl and ddlR genes.

Using a RACE assay (Frohman, 1994), a purine nucleotide G(+1), located 53 bp upstream of the ddl initiation codon, was identified as the apparent 5’ end of ddl mRNA (Fig. 1 and S1A). The sequences TTTCAT and TATAAT with 4 and 0 mismatches, respectively, to the consensus −35 and −10 regions of σ^A-dependent promoters of most bacteria and a 17-bp spacer region can be identified upstream of the apparent ddl transcription start point (Fig. 1). Due to the nature of the RACE experiment, the possibility that T(-1) serves as an actual or alternative transcription start point for ddl cannot be excluded (Fig. S1A).

Considering that our expression data suggested the existence of an additional promoter for ddlR, we used the same RACE approach to identify a purine nucleotide G(+1), located 161 bp upstream of the ddlR start codon, as the apparent 5’ end of the ddlR mRNA (Fig. 1 and S1B). The sequences TTGCAA and TATGAT with 2 and 1 mismatches to the consensus −35 and −10 regions of σ^A-dependent promoters, respectively, and a 18-bp spacer region can be identified upstream of the apparent ddlR transcription start point (Fig. 1). In an unusual manner, the ddlR transcription start point is located within the ddl gene, 135 bp upstream of the ddl termination codon. Due to the nature of the RACE experiment, the possibility that T(-1) serves as an actual or alternative transcription start point for ddlR cannot be excluded (Fig. S1B).

Finally, we used real-time RT-PCR and primers located upstream and downstream of the ddlR promoter (oBB754 and oBB734) to reveal a transcript that spans the entire region between the primers and includes parts of both ddl and ddlR open reading frames (data not shown). This confirms that ddlR is co-expressed with ddl from the ddl promoter in addition to being expressed from its own promoter.

Effect of a ddl mutation on the activity of the ddl promoter.

An in-frame deletion covering 67% of the 897-bp ddl gene was introduced by a gene-replacing technique (Ng et al., 2013) into the chromosome of strain 630∆erm; the deletion left intact the ddlR promoter at the 3’ end of the ddl gene. As expected, the Δddl mutant strain, BCD27, was not able to grow without D-Ala-D-Ala and was lysing after a limiting concentration of D-Ala-D-Ala was exhausted in the medium; growth was fully restored at higher concentrations of D-Ala-D-Ala (Fig. 2B and 2C). Growing a ddl null mutant strain in the presence of a low concentration of D-Ala-D-Ala (2.5 µg/ml) led to more than 250-fold elevated expression of the ddl-gusA fusion compared to a parent strain (Table 2, strain BCD77). This higher expression from the ddl promoter is apparently due to the inability of the ddl mutant to synthesize D-Ala-D-Ala and the resulting decrease in the D-Ala-D-Ala pool that cannot be compensated by a low concentration of D-Ala-D-Ala in the growth medium. The low D-Ala-D-Ala pool presumably results in higher activity of DdlR, which activates the ddl promoter. Increasing D-Ala-D-Ala concentration to 40 or 200 µg/ml reduced expression of the ddl-gusA fusion more than 440-fold (Table 2). Complementation of the ddl deletion by placing the ddl locus at an ectopic locus of the chromosome restored cell growth in the absence of D-Ala-D-Ala (Fig. 2D, strain BCD28) and reduced expression from the ddl promoter to the level comparable to that of the parent strain (Table 2, strain BCD78).

A more than 1,000-fold difference between the low level of expression from the ddl promoter in the parent strain in the presence of D-Ala-D-Ala excess and the high level of promoter expression in a ddl null mutant grown with limiting concentrations of D-Ala-D-Ala demonstrates the dramatic physiological dynamic range of DdlR-mediated regulation. The pattern of ddl expression in parent and ddl mutant cells growing in the complex TY medium indicates that most of regulation (144-fold) occurs when cells experience D-Ala-D-Ala limitation; only 7-fold regulation is achieved by D-Ala-D-Ala excess. The full dynamic range of ddl regulation increases to 4,000-fold when a ddlR null mutant strain is included in the analysis.

Using real-time RT-PCR assays and primers specific to the remaining part of the partially deleted ddl gene, we confirmed that ddl expression, and therefore DdlR activity, were increased in ddl mutant cells grown in the presence of a low concentration of D-Ala-D-Ala (Table 1, strain BCD27). The magnitude of this regulation (40-fold) was high but smaller than the level detected using a transcriptional fusion (250-fold). This may be due to differences between the two approaches or the multicopy nature of the fusion. A possibility of altered stability of mRNA encoded by the partially deleted ddl gene also contributes to some uncertainty about the exact level of DdlR-mediated regulation. Expression of the ddlR gene was also increased (18-fold) in a ddl null mutant strain grown with 2.5 µg/ml of D-Ala-D-Ala compared to the ddlR expression level in the parent strain, indicating efficient positive autoregulation of the ddlR gene due to much stronger expression from a fully active ddl promoter than from the ddlR promoter (Table 1, strain BCD27). A higher concentration of D-Ala-D-Ala resulted in the decrease of expression of the ddlR gene and the remaining part of the ddl gene in a ddl null mutant strain to basal levels (Table 1, strain BCD27).

DdlR is a DNA-binding protein.

A C-terminally His₆-tagged form of C. difficile DdlR was purified to near homogeneity (see Experimental Procedures). This version of DdlR is able to complement a ddlR null mutation (Table 1, strain BCD17). In a gel shift experiment (Fig. 3), DdlR was able to bind a DNA fragment containing the entire ddl regulatory region with moderately high affinity [an apparent equilibrium dissociation constant (K_D) was 25–50 nM; “K_D” is defined here as the DdlR concentration needed to shift 50% of DNA fragments under conditions of vast DdlR excess over DNA]. Binding of DdlR appeared not to be affected by the presence of its possible effector, D-Ala-D-Ala, and was only weakly negatively affected by the presence of PLP alone or a combination of PLP and D-Ala-D-Ala (K_D≈50–100 nM) (Fig. 3).

Fig. 3.

Binding of DdlR to the ddl regulatory region as detected by a gel-shift assay. A radioactively labeled ddlp⁺ DNA fragment was incubated with increasing amounts of purified DdlR without or with 2 mM D-Ala-D-Ala and 100 µM PLP. DdlR monomer concentrations used (nM) are indicated below each lane, and the concentrations needed to shift ~50% of DNA fragments, are underlined. The arrows indicate the bands corresponding to unbound ddl DNA and the complex of DdlR with ddl DNA. Each gel-shift assay was repeated at least two times.

No binding to the regulatory regions of the C. difficile CD630_23440 gene (Fig.4A) or leuA gene (data not shown) was detected at ≤800 nM DdlR indicating specificity of DdlR interaction with ddl. Also, no binding of DdlR to the ddlR promoter was detected, indicating that DdlR does not regulate this promoter (Fig. 4B).

Fig. 4.

Binding of DdlR or DdlR1 to various regulatory regions as detected by a gel-shift assay. Radioactively labeled CD630_23440 (A), ddlR (B), ddlp1 (C), ddlp2 (D), and ddlp⁺ (E, F) DNA fragments were incubated with increasing amounts of purified DdlR (A-D) or DdlR1 (E, F). 100 µM PLP was added for the experiment shown in panel F. DdlR or DdlR1 monomer concentrations used (nM) are indicated below each lane, and the concentrations, which are needed to shift ~50% of DNA fragments, are underlined.

DNase I footprinting experiments showed that DdlR protects a 37-bp region of the template DNA strand from positions −74 to −38 with respect to the ddl transcription start point (Fig. 1 and 5). The protected region is located 1 nt upstream of the ddl −35 promoter region and includes a 33-bp sequence ACCACACATAATATAAACTTCGTATATAGTGGT, containing inverted 5-bp repeats at its ends (underlined) separated by 23 nt (Fig. 1 and 5, positions −71 to −39). Two hypersensitive bands, separated by one helical turn of DNA, were detected at positions −61 and −49 with respect to the ddl transcription start point, indicating DNA distortion (e.g., bending) upon DdlR binding (Fig. 5). Similarly to our gel shift experiment, addition of PLP or a combination of D-Ala-D-Ala and PLP did not alter significantly the pattern of DdlR-mediated protection or the strength of DdlR-DNA interaction (Fig. 5).

Fig. 5.

DNase I footprinting analysis of DdlR binding to the ddl regulatory region. ddlp⁺ (A) or ddlp2 (B) DNA fragments, radioactively labeled on the template strand, were incubated with increasing concentrations of purified DdlR in the absence or presence of 100 µM PLP or 100 µM PLP and 2 mM D-Ala-D-Ala. DdlR monomer concentrations used (nM) are indicated below each lane. The corresponding A + G sequencing ladders is shown in the left lane. The protected area is shown by a vertical line. The positions of hypersensitive bands are indicated by arrows.

Mutational analysis of the DdlR-binding site.

Double mutations were introduced into each of the 5-bp inverted repeats within the 33-bp dyad symmetry sequence associated with the DdlR-binding site. The p1 mutation, affecting positions −70 and −69 of the distal (with respect to ddl) repeat (Fig. 1), significantly reduced the ability of DdlR to activate ddl expression (Table 2, strain BCD7). A gel-shift experiment revealed that the ddlp1 mutation abolished the ability of DdlR to bind to the regulatory region (Fig. 4C).

The p2 mutation, at positions −41 and −40 in the proximal repeat of the DdlR-binding site, and the p1/p2 double mutation (Fig. 1) caused expression phenotypes even stronger than that of the p1 mutation (Table 2, strains BCD8 and BCD9), but we cannot exclude the possibility that the p2 mutation affects promoter activity per se because of its proximity to the −35 region. More importantly, no binding of DdlR to the p2- or p1/p2-containing DNA fragments was observed (Fig. 4D and data not shown). The strong negative effect of the p1 and p2 mutations on DdlR binding was confirmed by footprinting experiments (Fig. 5B and data not shown). We conclude that the integrity of both arms of the dyad-symmetry sequence is essential for efficient binding of DdlR and therefore for the activation of the ddl promoter.

The combination of the p1 and p2 mutations restored a perfect inverted repeat within the DdlR-binding region (Fig. 1). Thus, it is the exact sequence of the element and not the dyad symmetry per se that is important for interaction with DdlR.

Mutations in the aminotransferase domain of DdlR.

DdlR appears to have two functional domains, a short N-terminal DNA-binding domain and a longer C-terminal aminotransferase-like domain. Most MocR/GabR-type proteins are able to form a covalent linkage (Schiff base) with PLP (Edayathumangalam et al., 2013; Okuda, Kato et al., 2015); a similar covalent bond between PLP and a conserved lysine residue is formed by aminotransferases and other PLP-dependent enzymes (Mehta et al., 1993). Purified DdlR protein had yellow coloration (data not shown) reflecting the expected presence of PLP as a tightly bound component of DdlR. The presence of covalently bound PLP was proved by spectrophotometric analysis of the purified protein demonstrating a peak at A₄₃₀ characteristic of PLP in its Schiff-base (i.e., covalently bound to a lysine residue) form (Fig. 6A). Similar assays indicated PLP presence in B. brevis DdlR and other MocR/GabR-type proteins (Okuda, Kato et al., 2015; Takenaka et al., 2015). The lysine residue that interacts with PLP is conserved in all PLP-containing proteins and is located at position 328 in C. difficile DdlR. The conserved lysine of DdlR was modified to glutamine by site-directed mutagenesis [glutamine is found at this position in some MocR/GabR proteins but cannot form a covalent bond with PLP (Bramucci et al., 2011; Jochmann et al., 2011; El Qaidi et al., 2013)]. Purified DdlR1(K328Q) had the same yield as wild-type DdlR in E. coli cells but no yellow coloration, consistent with DdlR1 being a stable protein defective in interaction with PLP. As expected, the spectroscopic peak characteristic of PLP presence was not detected in purified DdlR1 (Fig. 6B). The mutant protein retained the ability to bind the ddl regulatory region in vitro as efficiently as wild-type DdlR, implying that the DdlR1 protein is still folded and that PLP that is bound to wild-type DdlR does not contribute significantly to its ability to bind DNA (Fig. 3, 4E and 4F).

Fig. 6.

UV-visible spectra of DdlR, DdlR1, and PLP. 40 µM DdlR (A) or DdlR1 (B) were in 50 mM HEPES - 500 mM NaCl - 92.5 mM imidazole (pH 7.5) buffer. The inset in panel A compares the spectra of DdlR and of free 40 µM PLP in 50 mM HEPES (pH 7.5) buffer. The intensity of the absorption peak at A₄₃₀ manifested by wild-type DdlR indicates that more than half of DdlR molecules are bound to PLP.

To study the functional properties of DdlR1 in vivo, we have replaced the wild-type ddlR gene by its ddlR1 allele within the chromosome of C. difficile cells (see Experimental procedures). The ddlR1 mutant was unable to grow in the absence of D-Ala-D-Ala, and the level of expression from the ddl promoter was low even if the D-Ala-D-Ala concentration was low indicating that proper interaction between DdlR and PLP is required for the activation of the ddl promoter (Table 1, strain BCD36 and Table 2, strain BCD43). Furthermore, DdlR1 did not regain activity in vivo even in the presence in the medium of an excess of pyridoxal (100 µM), a precursor of PLP (data not shown).

Distribution of DdlR proteins.

Genes, encoding proteins highly similar to C. difficile DdlR (more than 45% identity), are present only in low GC-containing Gram-positive bacteria (phylum Firmicutes, classes Bacilli and Clostridia); no more than one such gene is present in each genome. Some of these genes are located immediately upstream of a ddl gene, as reported for B. brevis (Takenaka et al., 2015), or immediately downstream of a ddl gene, as in C. difficile. Other putative ddlR genes are not linked to the ddl gene(s) (Fig. S2). Interestingly, in several genomes, e.g., Clostridium butyricum, a putative ddlR gene is located next to the alr gene, which encodes alanine racemase (Fig. S2). Alanine racemase catalyzes the conversion of L-Ala to D-Ala, providing a substrate for the ddl product, D-Ala-D-Ala ligase (Walsh, 1989). In some bacteria, alanine racemase is the major or only enzyme that performs this reaction and is essential if D-Ala is not provided (Ferrari et al., 1985).

The analysis of sequences located upstream of the ddl genes from the genomes containing putative ddlR genes revealed that one ddl gene (or a ddl-containing operon) in each tested genome (some genomes contained more than one ddl gene) was preceded by a conserved 35-bp dyad-symmetry motif, AACCAC-N₁₁-A-N₁₁-GTGGTT (Bailey and Elkan, 1994)(Fig. 7). In C. difficile, as shown above, a version of this motif constitutes a major part of the DdlR-binding site and its integrity is required for DdlR binding and regulation. Apparently, this sequence serves as a DdlR-binding site in other genomes, including B. brevis [though a different sequence was proposed for this role, based on the deletion and mutational analysis of the regulatory region of the B. brevis ddlR ddl operon (Takenaka et al., 2015)]. In different genomes, the 35-bp motif possesses a varying degree of an additional dyad symmetry and is easily identified by the presence of absolutely conserved C nucleotides at the positions 3, 4, and 6 and the corresponding G nucleotides at positions 30, 32 and 33 (Fig. 7). Sequences highly similar to the DdlR-binding motif are found only in genomes containing DdlR-like proteins and only upstream of putative DdlR targets (i.e., ddl, alr, or ddlR genes) (Grant et al., 2011). In genomes that contain two or more ddl genes, only one of them is regulated by DdlR, as indicated by the presence of the putative DdlR-binding sites. The alignment of DdlR-binding motifs from different genomes revealed a conserved TTT stretch 1 nt downstream of the motif (Fig. 7). In C. difficile, these T residues constitute a part of the −35 region of the ddl promoter (Fig. 1), a location perfectly consistent with DdlR being a transcription activator; a similar arrangement between the DdlR-binding motifs and the −35 promoter regions of target genes apparently is conserved in other bacteria.

Fig. 7.

Motif logo for the conserved sequence in the putative DdlR-binding sites of different bacteria. The logo was generated by the MEME function of the MEME suite (Bailey and Elkan, 1994). The dyad symmetry region is indicated by horizontal arrows and the conserved central nucleotide by a vertical arrow. The partial, putative −35 region is underlined. The nucleotide sequences of the regulatory regions of genes from 11 Clostridia and Bacilli strains were analyzed: ddl from C. difficile, Clostridium tetani, Clostridium butyricum, Clostridium acetobutylicum, Clostridium pasteurianum, and Clostridium ljungdahlii, ddlR from Bacillus anthracis, Lysinibacillus sphaericus, and Paenibacillus polymyxa, and alr from Clostridium botulinum and C. butyricum. The simplified consensus sequence and the corresponding actual sequences from three genomes are shown below the logo. The elements of dyad symmetry are underlined.

Other proteins that are less similar to DdlRs are encoded in many bacterial genomes, but are likely to be either functionally unrelated MocR/GabR-type proteins or aminotransferases and similar PLP-dependent enzymes (Milano et al., 2015) (the latter is the case for proteins, encoded by ddl-associated ddlR-like genes identified in Actinobacteria and suggested to encode DdlR (Takenaka et al., 2015); it is possible that these genes encode alanine racemases).

DISCUSSION

We have shown that C. difficile DdlR, a member of the MocR/GabR subfamily of the GntR family of transcriptional regulators, is a highly efficient activator of the ddl gene responsible for the de novo biosynthesis of D-Ala-D-Ala, an essential precursor of peptidoglycan. Physiologically, D-Ala-D-Ala serves as the negative effector (anti-activator) of C. difficile DdlR in accord with the reduced or increased need for expression of the D-Ala-D-Ala biosynthetic pathway in the presence of excess D-Ala-D-Ala or under conditions of D-Ala-D-Ala limitation, respectively. It is unclear whether D-Ala-D-Ala itself or a derivative interacts directly with C. difficile DdlR. However, direct interaction between B. brevis DdlR and D-Ala-D-Ala was demonstrated previously (Takenaka et al., 2015). A function of B. brevis DdlR in the regulation of one of the two ddl genes in this organism similar to that of C. difficile DdlR was suggested, although the role of the regulator was shown in a heterologous host and the functional role of D-Ala-D-Ala in interaction with DdlR was not elucidated (Takenaka et al., 2015). DdlR, according to our knowledge, is the only described dedicated transcriptional regulator of peptidoglycan synthesis.

A C. difficile ddlR null mutant strain requires D-Ala-D-Ala for growth; therefore, the ddlR gene is conditionally essential. However, even complex growth media do not contain sufficient amounts of D-Ala-D-Ala to support the growth of the ddlR null mutant strain. Therefore, under laboratory conditions, ddlR behaves as an essential gene. It is likely that ddlR is also essential under natural growth conditions, including growth in animals. Thus, C. difficile DdlR belongs to a rather rare category of essential transcriptional regulators. In the model Gram-positive bacterium B. subtilis, only one such regulator, WalR, a two-component system response regulator, is known. WalR, as DdlR, is involved in the control of cell wall metabolism (Dubrac et al., 2008).

An extremely high dynamic range of regulation of ddl, a gene essential for bacterial growth, is very unusual. It is not entirely clear what can lead to variations in the D-Ala-D-Ala pool that would result in altered DdlR activity. It is unlikely that bacterial cells can ever encounter high D-Ala-D-Ala concentrations in the environment. The rather unusual mode of peptidoglycan crosslinking in C. difficile (3–3 crosslink connecting two residues of meso-DAP instead of a more common 3–4 crosslink between meso-DAP and D-Ala residues) (Peltier et al., 2011) may allow the release of D-Ala-D-Ala from the pentapeptide stem by L,D-endopeptidase at the transpeptidation (crosslinking) step of peptidoglycan maturation reducing the demand for D-Ala-D-Ala synthesis due to the reutilization of the dipeptide. An enzyme with such specificity, i.e., one that acts on a pentapeptide stem and releases D-Ala-D-Ala, was described in Enterococcus faecalis (Magnet et al., 2007). A homologous protein (CD630_29630; 23% identity) is present in C. difficile, though the specificity of this enzyme has not been described.

It is more likely, however, that the main role of DdlR is to ensure sufficient supply of D-Ala-D-Ala under conditions that lead to the drop in the intracellular concentration of the dipeptide. Under such conditions, DdlR activity would increase and this would lead to an increase in ddl expression. Perturbations in the intracellular concentration of the dipeptide may be caused by changes in the rate of D-Ala-D-Ala synthesis or consumption, e.g., due to changes in the substrate supply (D-Ala and ATP) or variations in the demand for D-Ala-D-Ala during peptidoglycan synthesis at different stages of growth. Interestingly, all bacteria that contain DdlR-like proteins form endospores and require massive peptidoglycan synthesis during spore formation and germination. In fact, expression of ddl during sporulation is induced (but only 4-fold) in a σ^E-specific manner (Fimlaid et al., 2013) and increased 7-fold during germination and outgrowth of C. difficile spores (Dembek et al., 2013).

The efficiency of DdlR-mediated activation of the ddl gene is amplified due to co-transcription of the ddl and ddlR genes and, therefore, positive autoregulation of DdlR. On the other hand, in the presence of excess D-Ala-D-Ala, cells need neither expression of the D-Ala-D-Ala-biosynthetic pathway nor high expression of the activator of this pathway. However, the existence of the dedicated ddlR promoter, which is not regulated by DdlR, suggests that the cells maintain a certain level of DdlR even in the presence of D-Ala-D-Ala. Also, even when DdlR is completely inactive, the cells maintain a residual level of ddl expression, though this level is not sufficient for growth.

As shown here, a 33-bp sequence, ACCACACATAATATAAACTTCGTATATAGTGGT, with two 5-nt dyad-symmetry elements, separated by 23 nt, lies within the DdlR-binding site almost immediately upstream of the −35 promoter region of ddl. From the mutational analysis, we conclude that both the distal and proximal arms of the dyad-symmetry element found within the DdlR-binding site are essential for efficient binding of DdlR; this interaction appears to be mostly independent of the presence of PLP or D-Ala-D-Ala in vitro. Based on their similarity to aminotransferases, the MocR/GabR subfamily members were hypothesized to form head-to-tail dimers and, therefore, bind to direct DNA repeats (Rigali et al., 2002). The head-to-tail dimer configuration of such proteins was confirmed for GabR (Edayathumangalam et al., 2013) and may be the case for DdlR. Nevertheless, as shown here, DdlR is apparently able to interact with inverted repeats, possibly due to the flexibility of the linker that connects its DNA-binding and effector-binding domains. In fact, a combination of direct and inverted repeats was suggested as a binding site for several other MocR/GabR-like proteins (Al-Zyoud et al., 2016; Suvorova and Rodionov, 2016).

Sequences similar to the DdlR-binding motif, with a consensus AACCAC-N₁₁-A-N₁₁-GTGGTT, were detected upstream of putative targets of DdlR-proteins in other genomes (Fig. 7) but not detected in the regulatory regions of genes controlled by other proteins of the MocR/GabR subfamily (Suvorova and Rodionov, 2016). However, similar to the case of DdlR, doublets of C and G nucleotides were shown to be important for DNA binding of GabR and PdxR and for regulation by these proteins. In all cases, these nucleotides were parts of direct or inverted repeats of (A)CCA or TGG(T) (Belitsky, 2004, 2014; Al-Zyoud et al., 2016; Suvorova and Rodionov, 2016). The spacing and the relative orientation of these tri- or tetra-nucleotide basic units differ for binding sites of these three proteins. Similar repeated short sequences were detected in putative binding sites of other members of the MocR/GabR subfamily or other proteins of the GntR family (Rigali et al., 2002; Wiethaus et al., 2008; Suvorova and Rodionov, 2016).

Only a handful of proteins of the MocR/GabR subfamily, have been characterized in mechanistic details. The founding member of this group, GabR, requires both PLP and another low-molecular weight compound, GABA, for modulation of its activity (Belitsky, 2004; Okuda, Kato et al., 2015; Wu et al., 2017). GabR forms a covalent Schiff-base linkage with PLP, in which the ε-amino group of Lys312 forms an imine bond with the aldehyde group of PLP (Edayathumangalam et al., 2013). PLP appears to act as a tightly bound intrinsic component of GabR, akin to the cofactor role that PLP plays in aminotransferases and other PLP-dependent enzymes (Schneider et al., 2000; Eliot and Kirsch, 2004). Furthermore, to achieve its active state, GabR performs a transimination reaction between PLP and GABA, in which the imine linkage between Lys312 and PLP is broken and a new imine bond is formed between PLP, which remains bound to GabR non-covalently, and GABA (Belitsky, 2004; Okuda, Kato et al., 2015; Wu et al., 2017).

A similar cofactor-like role of PLP and the need for an amino group-containing compound as an effector are likely for many other members of the MocR/GabR-like proteins. For instance, N-acetyl-2,4-diaminobutyric acid apparently interacts covalently with PLP to inactivate Ruegeria pomeroyi EnuR as a repressor of genes of ectoine catabolism (Schulz, Hermann et al., 2017). For these proteins, we hypothesize that the likely conformational change that affects their activity is associated not with PLP binding per se but with binding of another low-molecular compound, such as GABA or N-acetyl-2,4-diaminobutyric acid, and with the subsequent modification of both this compound and PLP.

In contrast to this GabR-like group of proteins, PdxR proteins, which regulate genes involved in PLP biosynthesis (Magarvey et al., 2001; Jochmann et al., 2011; El Qaidi et al., 2013; Belitsky, 2014; Liao et al., 2015; Tramonti et al., 2015; Tramonti et al., 2017), apparently form a special group within the MocR/GabR subfamily. It is likely that all members of the PdxR group, like Listeria monocytogenes PdxR, respond only to PLP and do not require additional low-molecular compounds for modulation of their activity (Belitsky, 2014).

The signaling role of D-Ala-D-Ala availability with respect to DdlR activity is similar to the mode of regulation of the activity of GabR by GABA. However, in contrast to GABA, which acts as a co-activator of GabR, D-Ala-D-Ala or a related compound acts as an anti-activator of DdlR. DdlR is the first member of the MocR/GabR subfamily with this mode of interaction with an effector. It is very likely that DdlR-like proteins in other bacteria respond to the availability of D-Ala-D-Ala in the same manner. The summary of functional properties of different types of MocR/GabR proteins and their low-molecular-weight partners is shown in Table 3.

Table 3.

Comparison of MocR/GabR proteins

Regulator	Role	Effector’s identity and role	Role of PLP	Activity of the Lys mutant†	Reference
GabR	Activator	GABA, co-activator	“Co-activator”‡	No	(Belitsky and Sonenshein, 2002; Belitsky, 2004)
PdxR	Activator	PLP, anti-activator	Anti-activator, direct effector	Yes§¶	(Belitsky, 2014)
EnuR	Repressor	N-acetyl-2,4-di-aminobutyric acid, anti-repressor	“Co-repressor”‡	Yes§	(Schulz, Hermann et al., 2017)
DdlR	Activator	D-Ala-D-Ala or derivative, anti-activator	“Co-activator”‡	No	this work

^†The conserved Lys residue involved in the interaction with PLP was replaced by Ala or Gln.

^‡The quotation marks indicate the unusual role of PLP as a covalently bound component of the protein.

^§The protein lost ability to respond to the effector.

^¶B. R. Belitsky, unpublished.

EXPERIMENTAL PROCEDURES

Bacterial strains and culture media.

The C. difficile strains constructed and used in this study were all derivatives of strain 630∆erm, an erythromycin-sensitive derivative of strain 630 (Hussain et al., 2005), and are described in Table 4 or in the text. Escherichia coli strain JM107 (Yanisch-Perron et al., 1985) was used for isolation of plasmids; strain HB101 (pRK24 bla tet) (Trieu-Cuot et al., 1991) was used as a plasmid donor for conjugation with C. difficile. Plasmids used in this work are listed in Table S1. Cells were grown in TY or Brain Heart Infusion - 0.1% cysteine (BHIS) complex medium (Dupuy and Sonenshein, 1998; Bouillaut et al., 2011), for C. difficile; and in L broth (Miller, 1972), for E. coli. The same media with addition of agar were used for growth of bacteria on plates. Defined CDMM medium was used for selection of fluoroorotic acid resistant colonies and pyrE⁺ strains (Cartman and Minton, 2010).

Table 4.

C. difficile strains used

Strain	Genotype	Source or reference
630∆erm	EmS derivative of strain 630	(Hussain et al., 2005)
630∆erm pyrE	pyrE	(Heap et al., 2012)
LB-CD28	ddlR::ermB	630∆erm × pBL141
BCD4	pRPF185 (Pddlp+-gusA catP)	630∆erm × pBB1849
BCD5	ddlR::ermB pRPF185 (Pddlp+-gusA catP)	BLM28 × pBB1849
BCD6	pRPF185 (‘gusA catP)	630∆erm × pBB1853
BCD7	pRPF185 (Pddlp1-gusA catP)	630∆erm × pBB1854
BCD8	pRPF185 (Pddlp2-gusA catP)	630∆erm × pBB1855
BCD9	pRPF185 (Pddlp1/p2-gusA catP)	630∆erm × pBB1856
BCD12	ddlR::ermB pBL58 (bla catP)	BLM28 × pBL58
BCD15	ddlR::ermB pRPF185 (Pddlp1/p2-gusA catP)	BLM28 × pBB1856
BCD16	pBL58 (Pddl p+-ddlR-His6 bla catP)	630∆erm × pBB1864
BCD17	ddlR::ermB pBL58 (Pddl p+-ddlR-His6 bla catP)	LB-CD28 × pBB1864
BCD25	∆ddl pyrE	630∆erm pyrE × pBB1906
BCD27	∆ddl	BCD25 × pMTL-YN1C
BCD28	∆ddl pyrE::ddl	BCD25 × pBB1911
BCD35	ddlR1 pyrE	630∆erm pyrE × pBB1920
BCD36	ddlR1	BCD35 × pMTL-YN1C
BCD43	ddlR1 pRPF185 (Pddlp+-gusA catP)	BCD36 × pBB1849
BCD77	∆ddl pRPF185 (Pddlp+-gusA catP)	BCD27 × pBB1849
BCD78	∆ddl pyrE::ddl pRPF185 (Pddlp+-gusA catP)	BCD28 × pBB1849

C. difficile strains were maintained at 37°C in an anaerobic chamber (Coy Laboratory Products) with an atmosphere of 10% H₂, 5% CO₂ and 85% N₂. E. coli strains were grown at 37°C under aerobic conditions. The following antibiotics were used when appropriate: erythromycin, 10 µg/ml or thiamphenicol, 10 µg/ml for C. difficile strains and ampicillin, 50–100 µg/ml or chloramphenicol, 20–25 µg/ml for E. coli strains. BHIS plates containing thiamphenicol, 10 µg/ml, kanamycin, 50 µg/ml, and D-cycloserine, 250 µg/ml were used for conjugation. When present, D-Ala-D-Ala concentration was 20 µg/ml or as specified.

General molecular genetic methods.

Methods for common DNA manipulations, E. coli electroporation, and sequence analysis were as previously described (Belitsky and Sonenshein, 1998). Chromosomal DNA of C. difficile was isolated as described previously (Bouillaut et al., 2011) after the cells were disrupted using 0.1 mm silica beads and a Mini-BeadBeater (Biospec Products) for two 30-sec cycles at the maximal setting. All oligonucleotides used in this work are described in Table S2. Chromosomal DNA of C. difficile or plasmids constructed in this work were used as template for PCR. All cloned PCR-generated fragments were verified by sequencing.

Construction of the C. difficile ddlR null mutant.

To create a ddlR insertional mutation by TargeTron mutagenesis (Heap et al., 2007; Heap et al., 2010; Bouillaut et al., 2011), we first generated by PCR a ddlR-targeted intron fragment using primers oLB380, oLB381, oLB382 and EBS universal and a 1:1 mixture of pBL64 and pBL65 as template as previously described (Bouillaut et al., 2013). The PCR fragment was cloned between the HindIII and BsrGI restriction sites of pBL100 (Bouillaut et al., 2013). The resulting plasmid, pBL141, was introduced into E. coli conjugation donor strain HB101 (pRK24), and the resulting strain was mated with C. difficile strain 630∆erm as a plasmid recipient as described (Bouillaut et al., 2011). The transconjugants were selected for erythromycin resistance, conferred by the intron, in the presence of D-Ala-D-Ala and screened for the loss of thiamphenicol resistance, indicating the loss of the plasmid. The insertional disruption of ddlR in the resulting strain LB-CD28 was verified by comparing the sizes of PCR fragments from the wild-type and mutant chromosomal loci using flanking primers; the loss of the plasmid was established in separate PCR reactions using vector-specific primers. Whole-genome sequencing of strain LB-CD28 confirmed integration of the intron at the single position of the chromosome, 99 nt downstream of the ddlR translation start point.

Complementation of the ddlR mutation.

The ddlR gene was introduced in C. difficile strains as a part of an E. coli-C. difficile shuttle vector pBL58 (bla cat) (McBride and Sonenshein, 2011) under the control of the ddl promoter. In the first cloning step, a 0.3-kb PCR product containing the entire ddl regulatory region was synthesized by PCR using oRG464 and oRG465 as primers, cut with EcoRI and XbaI and cloned in pBL58 to create pBB1861. In the second step, a 1.5-kb PCR product containing the entire ddlR gene with six histidine codons at the 3’ end was synthesized by PCR using oRG466 and oRG461 primers, cut with XbaI and BamHI, cloned in pCR2.1-TOPO (Invitrogen) to obtain pBB1858, and then recloned, using XbaI and HindIII sites, in pBB1861 to create pBB1864 (Pddl-ddlR). pBL58, pBB1861, and pBB1864 were introduced into C. difficile cells by conjugation (Bouillaut et al., 2011).

Construction of ddl-gusA transcriptional fusions.

Plasmid pBB1849 (ddlp⁺-gusA) was created by replacing the tetR fragment of an E. coli-C. difficile shuttle plasmid pRPF185 (cat) (Fagan and Fairweather, 2011) with the 0.29-kb KpnI- and SacI-digested PCR product, containing the entire ddl regulatory region and synthesized with oBB732 and oBB733 as primers. Plasmids pBB1854 (ddlp1-gusA) and pBB1855 (ddlp2-gusA), containing 2-bp substitution mutations in the DdlR-binding site, were constructed as described above using fragments generated by two-step overlapping PCR. In the first step, products containing the 5’ part of the corresponding regulatory regions were synthesized by using oligonucleotide oBB732 as the forward primer and mutagenic oligonucleotides oBB737 (ddlp1) or oBB739 (ddlp2) as the reverse primer. Products, containing the 3’ part of the regulatory regions, were synthesized by using mutagenic oligonucleotides oBB738 (ddlp1) or oBB740 (ddlp2) as the forward primer and gusA-specific oligonucleotide oNM20 as the reverse primer. The PCR products were used in a second, splicing step of PCR mutagenesis as overlapping templates to generate a modified fragment containing the entire ddl regulatory region; oligonucleotides oBB732 and oNM20 served as PCR primers. pBB1856 (ddlp1/p2-gusA) was constructed as described above for pBB1854 but using DNA of pBB1855 as template.

A control plasmid, pBB1853, lacking a promoter upstream of the gusA gene was constructed by deleting the tetR gene and the tet promoter after digesting pRPF185 with KpnI and SacI, blunt ending the resulting vector fragment with the DNA polymerase I Klenow fragment, and self-ligating. The plasmids were introduced by transformation into C. difficile cells by conjugation (Bouillaut et al., 2011).

Isolation of RNA.

Samples of 1 ml of C. difficile cells growing exponentially in TY medium with or without D-Ala-D-Ala or other additions were collected at the OD₆₀₀≈0.5 by mixing with an equal volume of an 1:1 (vol/vol) solution of ethanol/acetone (−20°C) and kept at −80°C until further use. The cells were pelleted, washed with 1 ml of 10 mM tris-HCl (pH 8.0) - 1 mM EDTA buffer, and resuspended in 0.8 ml of the TRIzol Reagent. The cells were disrupted using a beadbeater as described above, and RNA was purified using the Direct-zol RNA MiniPrep Plus kit (Zymo Research). For real-time RT-PCR experiments, purified RNA (~10 µg) was further treated with Turbo DNA-free DNase I (Ambion) according to the manufacturer’s instructions. RNA was quantified using the NanoDrop ND-1000 spectrophotometer (Thermo Scientific).

Real-time RT-PCR.

cDNA was synthesized starting from 1 or 2 µg of RNA using random hexamer primers and SuperScript II or SuperScript III reverse transcriptase (Invitrogen) per the manufacturer’s instructions. The reactions were performed using a LightCycler 480 instrument (Roche Diagnostics Corporation) and SYBR Green I Master (Roche) or HOT FIREPol EvaGreen qPCR mix Plus (no ROX) (Solis BioDyne) according to the manufacturer’s instructions. The reactions were done in a total volume of 20 µl and contained 4 µl of 5-fold diluted cDNA or control RNA samples. Primer pairs oWN101-oWN102 or oWN127-oWN128 and oWN105-oWN106 were used for detection of ddl or ddlR transcripts, respectively. The rpoC transcript, detected with primers oWN99 and oWN100, was used for normalization. All primers were designed using the PrimerQuest tool (Integrated DNA Technologies). Serial dilutions of C. difficile chromosomal DNA (from 3.2 to 10,000 pg per reaction) were used to create calibration curves for each transcript and to compare amounts of different transcripts.

Determination of ddl transcription start point using rapid amplification of cDNA ends (RACE).

cDNA samples were synthesized from 1 µg of RNA in 20-µl reactions using 2 pmol of ddl- or ddlR-specific primer, oBB734 or oBB742, respectively, and SuperScript II reverse transcriptase (Invitrogen) per the manufacturer’s instructions. 3’ poly(A) tails were added using terminal transferase (New England Biolabs), and cDNA was purified using the PCR Purification kit (Qiagen) and eluted in a 50-µl volume. First-round PCR products were generated in 50-µl reactions using 2 µl of purified cDNA as template, universal anchor primer oKZ69, and ddl- or ddlR-specific primer, oBB745 or oBB747, respectively. The ddl-specifc product was sequenced directly using primer oBB745. The ddlR-specific product (1 µl) was PCR-amplified again using the universal amplification primer oKZ70 and the ddlR-specific primer oBB734 and then sequenced using oBB734. The nucleotide(s) at the junction between the gene-specific sequence and the stretch of A nucleotides, generated from sequencing the poly(A) tail, was assumed to be an apparent 5’ end of mRNA (Frohman, 1994).

Overexpression and purification of DdlR.

A 1.48-kb PCR product containing the entire ddlR gene with six histidine codons at the 3’ end was synthesized using oBB720 and oBB721 as primers, cloned in pCR2.1-TOPO between the NdeI and BamHI sites to create pBB1852, and then recloned, using the same sites, downstream of the T7 promoter of the expression vector pET-26b (Invitrogen). The resulting plasmid, pBB1859, was introduced into E. coli strain BL21/DE3 (Studier and Moffatt, 1986), and expression of DdlR was induced in the resulting strain in L broth (at OD₆₀₀≈1.0) by adding IPTG to 0.1–0.2 mM and incubation of the cells for 1 hour (addition of IPTG caused an abrupt arrest of growth indicating DdlR toxicity to E. coli cells). DdlR-His₆ was purified to near homogeneity as described previously for B. subtilis GabR-His₆ (Belitsky, 2004). Elution from the Ni²⁺-affinity column (His·Bind resin; Novagen) was with a buffer containing 185 or 385 mM imidazole. DdlR1(K328Q)-His₆ was purified in a similar way using pBB1875. The latter plasmid is similar to pBB1859 but the ddlR1 insert was created using overlapping PCR as described above with pairs of primers M13R-oBB742 and oBB743-M13F and pBB1852 as template.

Construction of ddl and ddlR1 mutants.

The ddl in-frame deletion and the ddlR1 point mutation were introduced into the chromosome by allelic exchange using pyrE mutant strain as described (Ng et al., 2013). The 1.25-kb fragment containing a 597-bp deletion within the ddl gene was created using overlapping PCR as described above with pairs of primers oBB795-oBB796 and oBB797-oBB798, digested with SacI and XhoI, and cloned into pMTL-YN3 (Ng et al., 2013) to create pBB1906. The deletion does not affect the internal ddlR promoter. The 1.48-kb ddlR1-containing PCR fragment described above was digested with SacI and BamHI and cloned into pMTL-YN3 to create pBB1920. Plasmids pBB1906 and pBB1920 were introduced into strain 630∆erm pyrE (Heap et al., 2012) by conjugation. The integration of the plasmid into the chromosome and subsequent excision of the plasmid and gene replacement events by selecting colonies resistant to fluoroorotic acid (in the presence of D-Ala-D-Ala) were performed as described (Ng et al., 2013). The resulting colonies were screened for the D-Ala-D-Ala auxotrophy. The replacement of the ddl⁺ or ddlR⁺ allele by the Δddl or ddlR1 allele in the chromosome of strain BCD25 (Δddl pyrE) or BCD35 (ddlR1 pyrE), respectively, was confirmed by PCR and sequencing. The pyrE mutation was replaced by a wild-type allele after conjugation with pMTL-YN1C (Ng et al., 2013) to create strains BCD27 (Δddl) and BCD36 (ddlR1).

To complement the Δddl mutation, a 1.55-bp PCR fragment containing the entire ddl gene with its promoter was generated using primers oBB732 and oBB798, digested with KpnI and XhoI, and cloned in pMTL-YN1C. The resulting plasmid, pBB1911, was introduced into strain BCD25 (Δddl pyrE) by conjugation, and the replacement of the pyrE mutation by the intact allele and the accompanying integration of the ddl gene at the pyrE locus was selected in the resulting strain BCD28 as described (Ng et al., 2013).

Labeling of DNA fragments.

The 302-bp PCR products containing the entire wild-type or mutants ddl intergenic regions were synthesized using vector-specific oligonucleotides oBB732 and oBB733 as primers and pBB1849 or its derivatives as template. The 256-bp PCR product containing the ddlR promoter region was synthesized using primers oBB757 and oBB758 and chromosomal DNA as template. Fragments containing the regulatory regions of C. difficile genes CD630_23440 (0.49 kb) or leuA (0.25 kb) were synthesized using primers M13R and M13F and plasmids pSD30 (Dineen et al., 2010) or pBL144 (a PCR fragment, generated with primers oJG1 and oJG2, was cloned in pCR2.1-TOPO), respectively, as templates. One of the primers for each PCR reaction was labeled using T4 polynucleotide kinase and [γ-32P]-ATP. The labeled PCR products were purified on an 8% non-denaturing polyacrylamide gel or used without purification.

Gel shift assays and DNase I protection experiments.

Incubation of DdlR with the 32P-labeled promoter fragments was performed in a binding buffer containing 20 mM Tris-Cl (pH 8.0) - 50 mM KCl - 2 mM MgCl2 – 5% glycerol - 0.5 mM EDTA - 1 mM DTT - 0.05% Nonidet P-40 – 25 µg/ml sonicated salmon sperm DNA. Samples (10 µl) containing varying amounts of DdlR and less than 1 fmole of DNA were incubated for 16 min at room temperature and separated on 8% non-denaturing 50 mM Tris - 384 mM glycine - 1 mM EDTA polyacrylamide gels in 35 mM Hepes - 43 mM imidazole buffer. In some experiments, 2 mM D-Ala-D-Ala or 100 µM PLP was present in the incubation mixture, as indicated.

For DNase I protection experiments, samples containing 20–40 fmoles of labeled DNA were incubated with DdlR as described above. One µl of the binding buffer containing 0.1–0.2 U RQ1 DNase I (Promega), 10 mM MgCl2 and 20 mM CaCl2 was then added, followed by addition, after 1 min, of 4 µl of 20 mM EDTA-95% formamide dye solution and subsequent heating of the samples at 80°C for 5 min. The samples were loaded without further purification on 7 M urea - 6% polyacrylamide DNA sequencing gels. The G+A sequencing ladder, generated according to a published procedure by boiling the appropriate samples of labeled DNA for 20 min (Liu and Hong, 1998), served to locate precisely the protected region.

The gels were dried, and the radioactive bands were detected and quantified using storage screens, a Storm PhosphorImager, and ImageQuant software (GE Healthcare).

Enzyme assays.

β-Glucuronidase specific activity was determined in Miller units as described previously; cell permeabilization was achieved by addition of 1% (vol/vol) toluene (Belitsky et al., 1995).

Spectrophotometry.

The UV-visible spectra of DdlR and PLP were determined using a Beckman DU640 spectrophotometer in a volume of 60 µl at a concentration of 40 µM.