Chemicals used in this study were of the highest commercially available purity and purchased from Kanto Chemical (Tokyo, Japan), Wako Pure Chemical (Tokyo, Japan), and Tokyo Kasei Kogyo (Tokyo, Japan). The stock solution of D-lactic acid was neutralized to pH 7.4 with sodium hydroxide before use as a growth substrate for S. oneidensis strains.

Bacterial strains and plasmids used in the present study are listed in Table 1. Escherichia coli strains were cultivated in Luria-Bertani (LB) or 2 × yeast extract-tryptone (2 × YT) medium at 37°C. The E. coli mating strain (WM6026) required 100 μg/ml 2,6-diaminopimelic acid (DAP) for growth. S. oneidensis strains were cultured at 30°C in LB or minimal medium (MM) (Nakagawa et al., 2015) containing a racemic mixture of DL-lactate, D-lactate, L-lactate, or pyruvate as the carbon and energy source. S. oneidensis strains were grown under aerobic or anaerobic TMAO-reducing conditions, since a CRP-deletion mutant (Δcrp) cannot utilize other electron acceptors (Saffarini et al., 2003). For aerobic cultivation, MM supplemented with each substrate (10 mM) in a test tube was inoculated with an S. oneidensis strain and shaken at 180 rpm. For anaerobic cultivation, MM supplemented with each substrate and TMAO (10 mM or 30 mM) in a test tube was inoculated with an S. oneidensis strain and incubated without shaking. The test tubes containing the anaerobic cultures were capped with butyl rubber septa and polycarbonate screw caps, and purged with pure nitrogen gas. Optical density at 600 nm (OD₆₀₀) was measured using a mini photo 518R photometer (Taitec, Tokyo, Japan). When necessary, 50 μg/ml kanamycin (Km) or 15 μg/ml gentamicin (Gm) was added to a culture medium. Agar plates contained 1.6% Bacto Agar (Difco).

To construct plasmid pBBRcrp, the crp gene was amplified using Phusion High-Fidelity DNA polymerase (New England Biolabs, Beverly, MA, United States) and primers crp-F-EcoRI and crp-R-BamHI (Supplementary Table S1). The PCR product was digested using EcoRI and BamHI, and cloned between the corresponding sites of pBBR1MCS-5 (Kovach et al., 1995). To construct plasmid pBBRdld, the dld gene was amplified using primers dld-F-BamHI and dld-R-XbaI (Supplementary Table S1). The PCR product was digested using BamHI and XbaI, and cloned between the corresponding sites of pBBR1MCS-2 (Kovach et al., 1995). The resulting plasmids, pBBRcrp and pBBRdld, were introduced into Δcrp cells by filter mating with E. coli WM6026 (Kouzuma et al., 2010).

Concentrations of D- and L-lactate in culture supernatants were determined according to a method described previously (Nakagawa et al., 2015). S. oneidensis strains were cultivated in liquid MM until a stationary growth phase, and cells were removed by filtration through a membrane filter unit (0.20 μm pore size, DISMIC-13JP; Advantec, Tokyo, Japan). Concentrations of D- and L-lactate in the filtrate were measured using F-kit reagents (J. K. International, Tokyo, Japan) according to the manufacturer’s instructions.

Shewanella oneidensis cells were cultivated in MM containing 10 mM racemic DL-lactate, D-lactate, or L-lactate under aerobic or 10 mM TMAO-reducing conditions and harvested at the logarithmic growth phase (OD₆₀₀ of 0.2–0.3 or 0.04–0.06, respectively). RNA was extracted from the cells using a Trizol reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions. The extracted RNA was purified using an RNeasy Mini Kit and RNase-Free DNase Set (Qiagen, Valencia, CA, United States). The quality of purified RNA was evaluated using an Agilent 2100 Bioanalyzer with RNA 6000 Pico reagents and RNA Pico Chips (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer’s instructions.

To synthesize cDNA, 1.0 μg of total RNA extracted from MR-1 cells grown in MM containing 10 mM racemic DL-lactate as described above was reverse-transcribed using Superscript III Reverse Transcriptase (Invitrogen) and Random Primers (Invitrogen) according to the manufacturer’s instructions. The synthesized cDNA was amplified using Ex Taq DNA polymerase (Takara, Tokyo, Japan) and the primer sets listed in Supplementary Table S1. Amplification steps consisted of an initial denaturation step at 96°C for 30 s, followed by 30 amplification cycles of 96°C for 30 s, 52°C for 30 s, and 72°C for 2 min, and a final extension step at 72°C for 7 min. The PCR products were electrophoresed on 2% agarose gels. Negative control reactions without reverse transcriptase were also performed.

Quantitative RT-PCR (qRT-PCR) was performed using a LightCycler 1.5 instrument (Roche, Indianapolis, IN, United States) according to a method described previously (Kouzuma et al., 2014). A PCR reaction mixture contained 15 ng total RNA, 1.3 μL of 50 mM Mn(OAc)₂ solution, 7.5 μL of LightCycler RNA Master SYBR Green I (Roche), and 0.15 μM primers listed in Supplementary Table S1. To generate standard curves, DNA fragments of target genes (dld, lldP, lldF, and 16S rRNA genes) were amplified by PCR using Ex Taq DNA polymerase (Takara) and the primer sets listed in Supplementary Table S1, and purified by gel electrophoresis using a QIAEX II Gel Extraction Kit (Qiagen) according to the manufacturer’s instructions. Standard curves were generated by amplifying a dilution series of the purified DNA fragments of each gene. The specificity of quantitative PCR was verified by a dissociation-curve analysis. Expression levels of target genes were normalized based on expression levels of the 16S rRNA gene.

5′-Rapid amplification of cDNA ends (5′-RACE) PCR reactions were carried out using 1.0 μg total RNA extracted from MR-1 cells grown in 10 mM racemic DL-lactate up to the logarithmic growth phase and a SMATer RACE cDNA Amplification Kit (Takara) in accordance with the manufacturer’s instructions. The first single-strand cDNA was synthesized using the gene-specific primer lldP_race_out (Supplementary Table S1) and subsequently amplified using Universal Primer A Mix (Takara) and the primer lldP_race_in (Supplementary Table S1). The amplified DNA fragments were purified using a QIAquick PCR Purification Kit (Qiagen). The Purified DNA fragments were cloned into T-Vector pMD20 (Takara) and sequenced to determine the 5′-end points.

To construct a series of reporter plasmids containing promoter regions upstream of lldP, these DNA regions were amplified using Phusion High-Fidelity DNA polymerase (New England Biolabs) and the primer sets listed in Supplementary Table S1. After purification using a QIAquick PCR Purification Kit (Qiagen), the DNA fragments were digested with EcoRI and SalI, and cloned between the corresponding sites of pMElacZ (Endoh et al., 2003). The constructed reporter plasmids were introduced into MR-1 and Δcrp by electroporation according to a method reported previously (Choi et al., 2006). The resulting reporter strains were aerobically cultivated in MM containing 10 mM racemic DL-lactate up to an OD₆₀₀ of 0.3–0.5. β-Galactosidase activity was measured in triplicate according to the method of Miller (1972).

The electrophoretic mobility shift assay (EMSA) using purified N-terminal histidine-tagged CRP (N-his-CRP) was performed according to a method described previously (Kasai et al., 2015). Briefly, N-his-CRP was produced by E. coli BL21(DE3) harboring pET-crp (Kasai et al., 2015), and purified using a QuickPick IMAC Metal Affinity Kit for Proteins (Bio-Nobile, Turku, Finland). The purity of the protein samples was assessed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and Coomassie blue staining. Cy3-labeled DNA probes PBlldP1 and PBlldP2 were generated by PCR with the 5′-Cy3 labeled primer set listed in Supplementary Table S1. PBlldP3 and PBlldP3m were generated by annealing of the complementary single-strand oligonucleotides listed in Supplementary Table S1. DNA-binding reactions were performed in 20 μl of a previously described reaction mixture (Kasai et al., 2015) containing 2 nM Cy3-labeled DNA probe, 50 ng N-his-CRP, and 50 μM cAMP.

To investigate the involvement of CRP in the catabolism of D-lactate and other organic acids in S. oneidensis MR-1, we compared growth characteristics of wild-type MR-1 (WT) and Δcrp (Kasai et al., 2015) on a 10 mM racemic mixture of D- and L-lactate (DL-lactate; 5 mM each isomer), D-lactate, L-lactate, and pyruvate (Figure 1). When these strains were grown on DL-lactate, the final cell density of Δcrp was approximately half that of WT (Figure 1A). Measurements of residual D- and L-lactate in the culture media (Figure 2) revealed that a large portion of D-lactate remained in the Δcrp culture. These results suggest that Δcrp has a decreased capacity for metabolizing D-lactate. This notion is supported by the growth deficiency of this mutant in the medium containing D-lactate as the sole substrate (Figure 1B). Complementary expression of crp in Δcrp restored the ability to grow on D-lactate (Supplementary Figure S1), demonstrating the involvement of CRP in D-lactate utilization. However, Δcrp was capable of growing on L-lactate and pyruvate, albeit that growth rates and final cell densities were slightly lower than those of WT (Figures 1C,D). These results suggest that CRP is required for the expression of gene(s) involved in D-lactate utilization, such as the dld gene encoding D-LDH in MR-1, although this regulator is not essential for the expression of genes involved in L-lactate and pyruvate utilization.

In the MR-1 genome, the dld gene is located within a gene cluster consisting of lldP (a putative lactate permease gene), dld, and lldEFG (SO_1522 to SO_1518; Figure 3A) (Pinchuk et al., 2009). Thus, to examine the involvement of CRP in the transcription of genes in this cluster, we first determined transcriptional units by RT-PCR and subsequently compared amounts of transcripts from these genes in WT and Δcrp by qRT-PCR. RT-PCR analysis using total RNA extracted from WT cells grown on DL-lactate detected a transcript containing the intergenic region between lldP and dld (Figure 3B, lane 2), indicating the polycistronic transcription of these two genes. The analysis also amplified the lldE–lldF and lldF–lldG intergenic regions (Figure 3B, lanes 5 and 6). However, no transcripts were detected when the dld–lldE intergenic region was analyzed (Figure 3B, lane 4), demonstrating that the transcription of dld is terminated within this region. These results indicate that the five genes in this gene cluster are transcribed as two operons, the lldP-dld and lldEFG operons.

We performed qRT-PCR analyses to determine expression levels of lldP, dld, and lldF in WT and Δcrp (Figure 4). The results revealed that expression levels of lldP and dld in Δcrp were decreased to approximately 20% of those in WT, although the expression of lldF did not significantly differ between these two strains. This result indicates that CRP is required for the transcriptional activation of the lldP-dld operon, although this regulator is not involved in the transcription of the lldEFG operon. Decreased expression of lldE and dldP in Δcrp was also observed when cells were grown under TMAO-reducing conditions (Supplementary Figure S2), indicating that CRP activates the transcription of these genes under both aerobic and anaerobic conditions.

The above-mentioned results suggest that the decreased expression of the lldP-dld operon in Δcrp is related to the deficient growth of this mutant on D-lactate. However, it remained possible that transcriptional activation of lldP and/or other CRP-regulated genes might also be required for D-lactate utilization. To test this possibility, we transformed Δcrp with a plasmid constitutively expressing the dld gene, pBBRdld (Table 1), and examined the growth of the resultant strain, Δcrp(pBBRdld), on D-lactate. When cells were grown under aerobic conditions (Figure 5), the growth of Δcrp(pBBRdld) on D-lactate was comparable to that of the control strain, MR-1(pBBR1MCS-2). The restored growth of Δcrp(pBBRdld) on D-lactate was also observed when cells were grown under TMAO-reducing conditions (Supplementary Figure S3), although its growth was slower than that of MR-1(pBBR1MCS-2). We consider that this growth retardation is due to the decreased ability of Δcrp to grow under anaerobic conditions because similar growth retardation was observed when Δcrp(pBBR1MCS-2) was cultivated using L-lactate and TMAO (Supplementary Figure S4). We therefore concluded that the deficient growth of Δcrp on D-lactate is attributable to the decreased expression of the dld gene.

A previous study reported that D-LDH activity in MR-1 was independent of growth substrates, whereas L-LDH activity was increased when it was grown on L-lactate (Brutinel and Gralnick, 2012). However, it was unclear how the presence of D- and L-lactate affects transcription of the lldP-dld and lldEFG operons. To address this, we cultivated MR-1 cells in the presence of D- or L-lactate, and investigated the expression levels of lldP and lldF by qRT-PCR analyses (Figure 6). We found that, when cells were grown on L-lactate, the expression of lldF was markedly increased, suggesting that the lldEFG operon is regulated by an L-lactate-dependent regulatory mechanism (i.e., the LlpR transcriptional regulator) (Brutinel and Gralnick, 2012). In contrast, the expression of lldP did not significantly differ between the two growth conditions. These results support the idea that the lldP-dld and lldEFG operons are regulated by different mechanisms, and that expression of the lldP-dld operon is not affected by the presence or absence of D-lactate.

To investigate the transcriptional mechanism of the lldP-dld operon in more detail, we determined a transcription start site (TSS) and upstream promoter region for the lldP-dld operon. 5′-RACE PCR analysis detected a TSS 192 bp upstream of the ATG start codon of lldP (TSS_lldP; Figure 7). A sequence similar to the consensus sequence of the E. coli σ⁷⁰-dependent promoter (–10 and –35 regions) (deHaseth et al., 1998) was found in the upstream region of TSS_lldP (Figure 7). Although a previous study reported that the RpoE-binding motif is detected upstream of many anaerobic respiratory genes in MR-1 (Barchinger et al., 2016), a sequence similar to that motif was not found in the upstream of lldP. However, a candidate CRP-binding sequence (5′-TTAAGTGACACCGATCACAGTT-3′) predicted by Gao et al. (2010) was located in position –70 to –49 relative to TSS_lldP. We therefore hypothesized that CRP binds to this sequence and activates the downstream lldP promoter (P_lldP).

To determine the DNA regions involved in the activation of P_lldP by CRP, we performed 5′-deletion analysis of the sequence upstream of TSS_lldP. For this purpose, WT and Δcrp strains were transformed with the lacZ reporter plasmids with promoter region deletions (pMElldP-541 to pMElldP+1; Table 1), and resultant transformants were grown on DL-lactate until the logarithmic growth phase. Measurement of the LacZ activities of these strains (Figure 8) revealed that, while Δcrp cells transformed with these reporter plasmids consistently exhibited very low levels of LacZ activities, WT cells transformed with pMElldP-541 and pMElld-182 exhibited much higher LacZ activities than these Δcrp cells. This result supports the idea that CRP plays a critical role in the activation of P_lldP. The analysis also revealed that LacZ activity in WT cells was decreased to the same level as that in Δcrp cells when they were transformed with pMElld-60. This result demonstrates that the activation of P_lldP by CRP requires the DNA region from –182 to –60 relative to TSS_lldP, which partially contains the putative CRP-binding site (Figure 7).

To investigate whether CRP binds directly to the upstream activation region of P_lldP, we performed an EMSA using purified CRP protein (Figure 9A) and DNA probes containing the upstream regions of P_lldP (Figure 9B). When a DNA probe containing the region from –182 to –61 relative to TSS_lldP (PBlldP1) was mixed with CRP, no shifted bands were observed (Figure 9C). However, when a probe containing the region from –182 to –1 relative to TSS_lldP (PBlldP2) was used, a shifted band was detected in a cAMP-dependent manner (Figure 9C), indicating that CRP directly binds to this region. Together with the results of the 5′-deletion analysis, these results suggested that CRP binds to a region around position –60, which is consistent with the position of the putative CRP-binding motif (Figure 7). To examine the binding of CRP to this motif, we performed an EMSA using a 50-bp probe containing the region from –84 to –35 (PBlldP3) and a mutated probe in which the core motif sequence was replaced (PBlldP3m). The results revealed that CRP specifically bound to PBlldP3, but not to PBlld3m (Figure 9C). We therefore conclude that this motif is involved in the binding of CRP and the transcriptional activation of the lldP-dld operon.