Thermotoga lettingae Can Salvage Cobinamide To Synthesize Vitamin B₁₂

Abstract

We recently reported that the Thermotogales acquired the ability to synthesize vitamin B₁₂ by acquisition of genes from two distantly related lineages, Archaea and Firmicutes (K. S. Swithers et al., Genome Biol. Evol. 4:730–739, 2012). Ancestral state reconstruction suggested that the cobinamide salvage gene cluster was present in the Thermotogales' most recent common ancestor. We also predicted that Thermotoga lettingae could not synthesize B₁₂ de novo but could use the cobinamide salvage pathway to synthesize B₁₂. In this study, these hypotheses were tested, and we found that Tt. lettingae did not synthesize B₁₂ de novo but salvaged cobinamide. The growth rate of Tt. lettingae increased with the addition of B₁₂ or cobinamide to its medium. It synthesized B₁₂ when the medium was supplemented with cobinamide, and no B₁₂ was detected in cells grown on cobinamide-deficient medium. Upstream of the cobinamide salvage genes is a putative B₁₂ riboswitch. In other organisms, B₁₂ riboswitches allow for higher transcriptional activity in the absence of B₁₂. When Tt. lettingae was grown with no B₁₂, the salvage genes were upregulated compared to cells grown with B₁₂ or cobinamide. Another gene cluster with a putative B₁₂ riboswitch upstream is the btuFCD ABC transporter, and it showed a transcription pattern similar to that of the cobinamide salvage genes. The BtuF proteins from species that can and cannot salvage cobinamides were shown in vitro to bind both B₁₂ and cobinamide. These results suggest that Thermotogales species can use the BtuFCD transporter to import both B₁₂ and cobinamide, even if they cannot salvage cobinamide.

INTRODUCTION

The Thermotogales order is one of the deepest bacterial lineages in the “ribosomal tree of life” (1–3). Thermotogales genomes are subjected to frequent gene transfers (4), with the largest number of transfers from archaea and firmicutes (5). Our recent study provided strong evidence that this order has acquired two different gene clusters, corrinoid synthesis from the firmicutes and cobinamide salvage gene cluster from various archaeal and bacterial organisms. These gene clusters allow for de novo synthesis of vitamin B₁₂, also termed cobalamin (6), and the synthesis of B₁₂ from partial B₁₂ molecules, called cobinamides. The B₁₂ cofactor is required by all domains of life, and de novo synthesis requires over 30 enzymes to produce an active form of B₁₂ (7). B₁₂ is required as a growth factor for many bacteria and archaea that do not have genes for its de novo synthesis. Only 50% of sequenced bacterial genomes that indicate a need for B₁₂ appear to encode the ability to synthesize B₁₂ (8). Some of these bacteria salvage incomplete corrinoid rings called cobinamides and use these as precursors to synthesize an active form of B₁₂ (9). Other microorganisms import B₁₂ from the environment using a B₁₂/cobinamide BtuFCD ABC transporter (10).

We recently explored the evolutionary origins of B₁₂-related genes in the Thermotogales and showed that some members of the order, like the Thermosipho africanus strains, can synthesize B₁₂ de novo, while another species, like Thermotoga maritima, cannot (6). The cobinamide salvage gene cluster has a patchy distribution among Thermotogales species, and ancestral state reconstruction suggested that this pathway was present in the Thermotogales' most recent common ancestor (6). We made several predictions about the origin of the cobinamide salvage pathway and predicted that Thermotogales species are able to utilize different cobinamides. Here, we test some of those hypotheses by studying in detail one member of the Thermotogales, Thermotoga lettingae.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

All cultures were grown in glassware that had been soaked in 65 to 70% sulfuric acid (vol/vol) for at least 24 h to remove traces of B₁₂ and then rinsed with distilled water (11). Tt. lettingae, Tt. maritima, and Th. africanus were grown on B₁₂-deficient medium, TL-1, DB-B, and DB-B with 0.5 g/liter vitamin-free casein (MP Biomedicals), respectively (see Table S1 in the supplemental material). TL-1 was modified from previously used DB-B medium (6). TL-1 medium is different from DB-B medium in that it contains 0.75 g/liter vitamin-free casein (MP Biomedicals) and its pH is 7.2. The casein was autoclaved in the TL-1 medium. Media were made anaerobic as previously described (12). Tt. lettingae was passed at least 5 times in TL-1 medium containing 92.5 nM cyanocobalamin, before growth assays. Tt. lettingae was washed 6× anaerobically by spinning down the cultures and resuspending the pellets in 1.5 ml of TL-1. The final pellet was resuspended in 1/8 the volume of the original culture, and 2% of washed cells were used to inoculate 10 ml of TL-1 containing 5 mg/ml maltose with or without cyanocobalamin or dicyanocobinamide and incubated at 65°C. Growth rate and doubling times were calculated using an online calculator (V. Roth, Doubling Time [http://www.doubling-time.com/compute.php]).

Lactobacillus bioassay.

Tt. lettingae B₁₂ production was measured using a Lactobacillus delbrueckii subsp. lactis ATCC 7830 bioassay as described in the Official Methods of Analysis of the AOAC International (13–15) using B₁₂ assay medium from BD Diagnostic Systems (Sparks, MD). All reagents used to grow cultures were tested for B₁₂ contamination using the Lactobacillus bioassay. The detection limit of the bioassay for B₁₂ was empirically determined to be 25 pg B₁₂, consistent with the USP cyanocobalamin reference standard. All reagents, excluding water, were tested at 10-fold or greater concentrations than the amount of these reagents that could be present in the Tt. lettingae (grown with cobinamide) extract used in the bioassay. No reagents had detectable B₁₂. Tt. lettingae, Tt. maritima, and Th. africanus cultures were grown to mid-log phase in 100- to 200-ml anaerobic cultures. Boiled cell extracts were prepared as described previously (6). Briefly, cells were centrifuged at 3,441 × g for 10 min. (Beckman GPR centrifuge using a GH-10 fixed-angle rotor), washed 3 times (centrifuged at 16,000 × g for 3 min in an Eppendorf benchtop centrifuge), resuspended in B₁₂-free water, incubated at 100°C for 30 min, and centrifuged for 20 min. The resulting supernatant was filtered through a 0.2-μm-pore-size filter. The extract was stored at −20°C until use. Before cell extracts were prepared, the weights of the cell pastes were measured. From the volumes of cell-free extracts harvested from those pastes, the concentrations of B₁₂ (μg/g cell paste) were calculated after the Lactobacillus assays. The concentrations of B₁₂, expressed as μg/liter, were calculated in a similar manner based on the volume of culture harvested.

Transcriptional analysis.

Quantitative real-time PCR (qRT-PCR) was used to measure the transcription of genes in the presence or absence of cyanocobalamin or dicyanocobinamide as previously described (6). RNA was isolated from cultures in mid-log growth phase using ZR fungal/bacterial RNA MiniPrep protocol (Zymo Research, Irvine, CA) and treated with RQ1 RNase-free DNase (Promega, Madison, WI). RNA concentration was determined using the Qubit RNA assay kit (Invitrogen, Carlsbad, CA), and cDNA was generated using the cDNA GoScript reverse transcription system (Promega, Madison, WI). Transcriptional levels were determined using SsoFast EvaGreen supermix and the CFX96 real-time PCR detection system (Bio-Rad, Hercules, CA). CFX Manager software (Bio-Rad) was used for transcriptional analyses.

Transcription was measured by qRT-PCR in cells grown with or without 92.5 nM cyanocobalamin (B₁₂) or dicyanocobinamide (cobinamide). Values were normalized relative to two endogenous controls, rspB (Tlet_1394) and gap (Tlet_0998). Both endogenous controls have previously been used in Tt. maritima qRT-PCR experiments (16). Averages and standard deviations for transcriptional analyses were from three determinations and at least three biological replicas using both endogenous controls.

BtuF purification.

The genes encoding the B₁₂ periplasmic binding protein from Thermotogales species were identified by running BLASTp (version 2.2.28+) using the amino acid sequence from E. coli BtuF against the GenBank nr database (accessed April 2012) (17). The signal peptide was identified for each BtuF using LipoP (18) and SignalP (19), and primers were designed to amplify the genes without their signal peptide-coding regions. The btuF genes of Tt. lettingae, Fervidobacterium nodosum, and Tt. maritima were amplified by PCR using the Failsafe enzyme mix (Epicentre) and incorporating restriction sites (see Table S2 in the supplemental material). The recombinant genes were cloned into the pBAD TOPO TA vector (Invitrogen, Carlsbad, CA). One Shot TOP10 chemically competent E. coli cells (Invitrogen) were transformed with these vectors. The cells containing the plasmid were selected for by growth on Bertani's lysogeny broth (LB)-ampicillin (100 μg/ml). Clones that contained the genes were identified by PCR or restriction digests; these genes were confirmed by sequencing at the University of Connecticut DNA Biotechnology Facility.

BtuF genes were expressed in pBAD TOPO vector in E. coli TOP10 cells, and proteins were purified as described by the manufacturer (Invitrogen) with a few modifications. The cultures were grown in LB-ampicillin (100 μg/ml) and induced with arabinose overnight at 18°C. The cultures were pelleted at 4,955 × g for 30 min and washed with 500 mM NaCl, 0.5 mM dithiothreitol (DTT), and 20 mM imidazole-HCl, pH 7.8. Bacterial protein extraction reagent (B-PER; Thermo Scientific) containing lysozyme and DNase I was used to resuspend the pellets as described by the manufacturer with a few modifications. The reaction mixture contained Halt protease inhibitor cocktail EDTA-free (10 μl/ml B-PER; Thermo Scientific) and RNase A (Qiagen; 1 ng/ml of B-PER reagent). The reaction mixture was incubated for at least 15 min at room temperature and heat treated (60°C for Tlet_1275 and Fnod_0257 and 70°C for TM0080) for 10 min to denature nonthermophilic proteins. The cell extracts were pelleted, and the supernatants were filtered through a 0.2-μm-pore-size filter (Thermo Scientific). Immobilized metal ion affinity chromatography was performed with a His SpinTrap (GE Healthcare) or Ni Sepharose high-performance resin (GE Healthcare) using the protocol of the manufacturer with the following modifications: 500 mM NaCl, 0.5 mM DTT, and 20 mM imidazole-HCl, pH 7.8, were used as the wash buffer and 500 mM NaCl, 0.5 mM DTT, and 500 mM imidazole-HCl, pH 7.8, were used as the elution buffer. The proteins were concentrated with an Ultra-4 10K centrifugal filter (Amicon), washed 3 to 5 times, resuspended in 500 mM NaCl, 0.5 mM DTT, and 50 mM imidazole-HCl, pH 7.8, and stored at 4°C. Protein concentrations were determined using the Bradford reagent and bovine serum albumin as the standard by following the manufacturer's instructions (Thermo Scientific).

DSF.

Differential scanning fluorimetry (DSF) was done as previously described with modifications (20). The proteins were assayed in a final volume of 20 μl with 150 mM NaCl, 0.16% (vol/vol) 5,000× Sypro orange (Invitrogen), and 20 mM citric acid-sodium phosphate buffer at pH from 3.5 to 7.0. A range of pH was used to determine the pH that allows for unfolding of the proteins in the absence and presence of potential ligands. The fluorescence intensities were measured using a CFX96 real-time PCR detection system (Bio-Rad) with excitation at 490 nm and emission at 530 nm. The samples were heated from 30 to 99°C, with a heating rate of 0.5°C per min. All the assays were done in triplicate. The melting temperature (T_m) is defined as the temperature at half the maximal fluorescence and was determined using gnuplot, with curve fitting to the Boltzmann equation with in-house scripts (21). The ΔT_m of the protein for a specific ligand was calculated as the difference between the T_m values in the presence and absence of a ligand.

Spectral analysis of BtuF binding to B₁₂ or cobinamide.

When B₁₂ or cobinamide bind to intrinsic factor and haptocorrin (a glycoprotein), the corrinoid visible spectrum changes (22, 23). Spectra of cyanocobalamin (15 μM) and dicyanocobinamide (15 μM) from 320 to 580 nm were recorded in a quartz cuvette in 1 mM imidazole-HCl buffer (pH 7.8) using a bandwidth of 1.0 nm, scan speed of 240 nm/min, and data intervals of 1.0 nm. The purified BtuF proteins (15 μM) were then added, and spectra were measured again from 320 to 580 nm.

RESULTS

Tt. lettingae salvages cobinamide to produce vitamin B₁₂.

Tt. lettingae was grown with and without B₁₂ or cobinamide to determine if these are required for growth. Tt. lettingae was able to grow without the addition of either B₁₂ or cobinamide, but the addition of either clearly decreased its doubling time. Tt. lettingae doubling times with no addition, with B₁₂ (92.5 nM), or with cobinamide (92.5 nM) were 57 ± 2.1 h, 40 ± 2.2 h, and 40 ± 2.0 h, respectively. Tt. lettingae growth rates with no addition, with B₁₂, or with cobinamide were 0.012 ± 0.001 h⁻¹, 0.017 ± 0.001 h⁻¹, and 0.017 ± 0.001 h⁻¹, respectively. The addition of cobalt (92.5 nM), the central atom of both B₁₂ and cobinamide, had no apparent effect on Tt. lettingae growth (data not shown). A Lactobacillus delbrueckii bioassay was performed to test for B₁₂ production. L. delbrueckii grows only in the presence of B₁₂ and is unable to grow with cobinamide. This was confirmed by incubating L. delbrueckii with 963 pg to 9.6 μg (0.09 to 925 nM) cobinamide for 48 h. Tt. lettingae is unable to produce B₁₂ de novo in B₁₂- and cobinamide-deficient media. Based on the empirically determined detection limit for the bioassay, there is less than 0.003 μg B₁₂/g cell paste when Tt. lettingae is grown without cobinamide or B₁₂. However, the addition of cobinamide (92.5 nM) allows Tt. lettingae to produce B₁₂ at 2.0 ± 0.3 μg/g cell paste (3.4 ± 0.3 μg/liter), over 660 times greater than the detection limit of the bioassay. As a control, cell extracts of Tt. lettingae cells were prepared from cells grown without B₁₂ and cobinamide, and no B₁₂ was detected in cells from a culture equivalent to 58 times the weight of cell paste used to assay Tt. lettingae cells grown with cobinamide. Cell extracts of Tt. maritima (does not synthesize B₁₂ [6]) were prepared from cells grown without B₁₂ and cobinamide, and no B₁₂ was detected when assayed using 5,800 times the weight of cell paste that had been used to assay Tt. lettingae cell extract from cells grown with cobinamide.

Tt. lettingae utilizes the cobinamide salvage pathway to produce vitamin B₁₂.

In our previous study of the evolution of the B₁₂ synthesis and transport genes, we used a computational approach (24) to predict that the Tt. lettingae salvage pathway gene cluster is transcriptionally controlled by a B₁₂ riboswitch (6). To test if Tt. lettingae utilizes the salvage pathway to produce B₁₂ from cobinamide (Fig. 1), in this study we used quantitative real-time PCR (qRT-PCR) to measure the transcription of pathway genes in the presence and absence of B₁₂ or cobinamide. We found that all seven genes encoding enzymes of the salvage pathway are repressed in cells grown in the presence of either B₁₂ or cobinamide compared to cells grown with B₁₂ (Table 1). Although it was expected that salvage genes would be repressed by growth with B₁₂, it was not obvious why these genes would be repressed in the presence of the precursor cobinamide. However, we found that the transcription levels of four of the seven salvage pathway genes were statistically higher when Tt. lettingae was grown with cobinamide than when cells were grown with B₁₂ (Table 1). This is consistent with the use of the salvage pathway that uses cobinamides for synthesis of B₁₂. In both Escherichia coli and Salmonella enterica serovar Typhimurium, genes encoding the B₁₂ ABC transporter and B₁₂ de novo synthesis pathway have higher transcription in the absence of B₁₂ (24–26). The B₁₂ ABC transporter and B₁₂ synthesis genes have been shown to be regulated by B₁₂ riboswitches in several organisms, including E. coli and S. enterica (27–31). The pattern of transcription we observed in Tt. lettingae is consistent with similar patterns found in other organisms that utilize B₁₂ riboswitches.

Fig 1.

Proposed cobinamide salvage pathways in the Thermotogales. The color box indicates that members of the Thermotogales group contain a protein homolog, while an empty box indicates no protein homolog. Thermotogales groups: Thermosipho group (Th. melanesiensis, Th. africanus TCF52B, and H1760334; black), Thermotoga/P. mobilis/M. prima (Tt. maritima, Thermotoga sp. strain RQ2, Thermotoga sp. strain cell2, Tt. naphthophila, Tt. neapolitana, Tt. petrophila, Petrotoga mobilis, Mesotoga prima; light blue), Tt. lettingae (dark blue), Tt. thermarum (light green), Fervidobacterium nodosum (red), Kosmotoga olearia (yellow). B₁₂ de novo synthesis can be broken into two separate steps, corrinoid synthesis and cobinamide salvage (6). CM, cytoplasmic membrane; Cbi, cobinamide; AdoCbl-P, adenosylcobalamin phosphate; AdoCbi, adenosylcobinamide; AdoCbi-P, adenosylcobinamide phosphate; AdoCbi-GDP, adenosylcobinamide GDP; Cbl, cobalamin; AdoCbl, adenosylcobalamin; Cby, cobyric acid; AdoCbr, adenosylcobyrinic acid; AdoCby, adenosylcobyric acid; Thr, threonine; Thr-P, threonine phosphate; AP-P, aminopropanol phosphate; a-R-P, α-ribazole-5-phosphate; DMB, 5,6-dimethylbenzimidazole; a-DAD, α-5,6-dimethylbenzimidazole adenine dinucleotide; Cbr, cobyrinic acid; Cbr(I), cob(I)yrinate a,c-diamide. Proteins that have been identified in other microbes but have no homologs in the Thermotogales genomes are indicated by question marks. Diagram was modified from reference 6 and used here with permission from Oxford University Press, and reaction pathways were also predicted based on evidence from other microbes (9, 38, 39).

Table 1.

Differential transcription of a putative B₁₂ riboswitch-regulated cobinamide salvage gene cluster, btuF, and B₁₂-dependent and independent enzyme-encoding genes in Tt. lettingae^a

Function	Gene	Relative fold expression in cells grown with:
		B12	Cbi	No B12 or Cbi
Cobinamide salvage	cobT′	1.0 ± 0.1	1.6 ± 0.2	11.6 ± 3.7
	cobU	1.0 ± 0.1	1.5 ± 0.2	8.7 ± 4.3
	Tlet_1238b	1.0 ± 0.1	1.5 ± 0.2	13.0 ± 6.9
	cobS	1.0 ± 0.2	0.9 ± 0.1	2.9 ± 1.1
	cbiB	1.0 ± 0.1	1.3 ± 0.2	7.0 ± 2.7
	cobD	1.0 ± 0.2	1.6 ± 0.2	5.5 ± 2.6
	cobA	1.0 ± 0.1	1.2 ± 0.2	5.3 ± 2.2
B12-ABC transporter binding protein	btuF	1.0 ± 0.2	1.9 ± 0.2	43.4 ± 6.9
B12-dependent methionine synthase	Tlet_1365	1.0 ± 0.1	0.9 ± 0.1	0.2 ± 0.1
B12-independent ribonucleotide reductase	Tlet_0432	1.0 ± 0.1	1.0 ± 0.1	10.5 ± 4.3
B12-dependent ribonucleotide reductase	Tlet_0333	1.0 ± 0.2	0.8 ± 0.1	2.8 ± 1.7c

^aTranscription was measured by qRT-PCR in cells grown in the presence or absence of 92.5 nM B₁₂ and cobinamide (Cbi). Genes of the salvage cluster, cobT′, cobU, Tlet_1238, cobS, cbiB, cobD, and cobA, are listed in the order they appear in the genome, with cobT′ adjacent to the putative B₁₂ riboswitch. Values were normalized relative to two endogenous controls, rspB (Tlet_1394) and gap (Tlet_0998). For relative fold expression of genes, expression values of cells grown with B₁₂ were normalized to 1. Values are averages from three determinations from at least three biological replicates, and standard deviations are shown.

^bTlet_1238 encodes a hypothetical protein.

^cNot statistically different based on a paired t test.

The Tt. lettingae genes for the BtuFCD transporter may also be under B₁₂ riboswitch control. Based on a computational examination (24), we predicted that Tt. lettingae and all sequenced Thermotogales btuFCD genes are regulated at the transcriptional level by a directly upstream putative B₁₂ riboswitch (see Fig. S1 in the supplemental material). Our qRT-PCR results indicate that btuF transcription is indeed regulated at the transcriptional level and is more highly expressed in the absence of B₁₂ and cobinamide (Table 1). As noted above, this pattern of transcription is also found in E. coli and S. Typhimurium. Apparently the constitutive level of transcription of the transporter is sufficient for cellular needs when B₁₂ or cobinamide are available, but a larger amount of transporter is needed to harvest those compounds when their concentrations are very low. The differential transcription of btuF when Tt. lettingae is grown in the presence or absence of cobinamide or B₁₂ suggests that the BtuFCD transporter functions to import both cobinamide and B₁₂ into the cell.

Tt. lettingae expresses B₁₂-dependent and -independent enzymes differently in the presence or absence of cobinamide and B₁₂.

Tt. lettingae contains a B₁₂-dependent methionine synthase (MS) and both B₁₂-dependent and -independent forms of ribonucleotide reductase (RNR) (6). B₁₂-dependent and -independent proteins perform the same function, but the former requires the cofactor B₁₂, while the latter does not. Though there is no evidence in the genome sequence for the regulation of B₁₂-dependent and -independent genes by B₁₂ riboswitches, we hypothesize that their transcription is linked to the availability of B₁₂. Tt. lettingae B₁₂-dependent MS was downregulated (4.1 ± 0.6-fold) in the absence of B₁₂ or cobinamide compared to when these cofactors were available (Table 1).

One would expect the B₁₂-independent RNR gene to be upregulated in the absence of B₁₂ compared to when B₁₂ is present. Our result supports this hypothesis (Table 1). In addition, one would expect the B₁₂-dependent RNR gene to be downregulated with no addition of B₁₂ compared to when B₁₂ is present. Four biological replicas with triplicate replicates were used to determine the transcription of the B₁₂-dependent RNR gene in the presence and absence of B₁₂, 2.8 ± 1.7 and 1.0 ± 0.2, respectively. A paired t test showed that there is no significant difference in B₁₂-dependent RNR transcription in the presence or absence of B₁₂.

There was no significant change in transcription of btuF when Tt. maritima, an organism unable to use cobinamide, was grown with cobinamide (0.9 ± 0.2) compared to growth in the absence of B₁₂ (1.0 ± 0.2). Transcription was measured by qRT-PCR, and values were normalized relative to two endogenous controls, deoD2 (TM1737) and eno (TM0877), as described in reference 6. This pattern of transcription is consistent with the inability of Tt. maritima to use cobinamide.

Overall, our transcription analysis results suggest that B₁₂ is utilized by Tt. lettingae and that cobinamide elicits a transcriptional response similar to that of B₁₂, because cobinamide can be processed into B₁₂ (Fig. 1 and Table 1).

BtuF proteins from Tt. lettingae, Tt. maritima, and F. nodosum bind B₁₂ and cobinamide in vitro.

The homology of the Thermotogales BtuFCD proteins to those of E. coli suggests that they function to import B₁₂. Tt. lettingae BtuF protein is 26% identical and 51% similar (BLASTp E value of 6e−14) to its E. coli strain K-12 substrain W3110 homolog. E. coli BtuF binds B₁₂ (17, 32) but has not been shown to bind cobinamide, and a Halobacterium sp. strain NRC-1 btuF mutant is unable to utilize cobinamide (10). The differential transcription of the btuF genes in the presence compared to the absence of B₁₂ from Tt. lettingae (see above), Tt. maritima, and Th. africanus strains (6) indicate that these genes encode a B₁₂ transporter.

Three Thermotogales BtuF proteins were examined for their binding properties: (i) TM0080 from Tt. maritima, an organism that cannot salvage cobinamide; (ii) Tlet_1275 from Tt. lettingae, an organism that salvages cobinamide; and (iii) Fnod_0257 from F. nodosum, an organism predicted to salvage several forms of cobinamides (cobinamide, cobyric acid, cobyrinic acid) (6) (Fig. 1). All three BtuF proteins were extremely thermostable, with melting temperatures (T_m) over 99°C at pH 7.0 as determined using differential scanning fluorimetry (DSF). Acidic conditions have been used to determine the T_m and ΔT_m by DSF for proteins stable over 96°C, the thermal limit of the instrument (20). Under such conditions (TM0080, pH 4.0; Fnod_0257, pH 4.5; Tlet_1275, pH 4.5), all three BtuF proteins were shown to interact with B₁₂ and cobinamide (Fig. 2A and B). In the presence of the ligand, the T_m value for each protein was greater than in the absence of the ligand. A positive difference, ΔT_m, is evidence for an interaction between the protein and the ligand (20, 21, 33–35). Spectral analyses were used to confirm that all three BtuF proteins bind both B₁₂ and cobinamide. This assay has previously been used to show that B₁₂ or cobinamide binds to intrinsic factor and haptocorrin (a glycoprotein) (22, 23). Spectral analyses show that all three BtuF proteins bind both B₁₂ and cobinamide (Fig. 2C and D). Cobalt alone (15 μM for spectral analysis and 75 μM for DSF) did not bind to the three BtuF proteins (data not shown).

Fig 2.

Differential scanning fluorimetry (DSF) and visible spectral analysis of BtuF binding of B₁₂ and cobinamide. For both B₁₂ (A) and cobinamide (B), DSF analyses were done with 1.5 μg, 5.0 μg, and 4.5 μg of TM0080 (pH 4.0), Fnod_0257 (pH 4.5), and Tlet_1275 (pH 4.5) proteins, respectively. Black triangles, TM0080; gray squares, Fnod_0257; red circles, Tlet_1275. The ΔT_m values of BtuF proteins were determined from three replicates. For both B₁₂ binding (C) and cobinamide binding (D), visible light spectra were recorded from 320 to 580 nm at 24°C using 15 μM BtuF protein in 1 mM imidazole-HCl buffer (pH 7.8). The three BtuF proteins had no detectable absorbance between 320 to 580 nm at 15 μM. Black, TM0080 with ligand; gray, Fnod_0257 with ligand; red, Tlet_1275 with ligand; light blue, B₁₂ (in panel C) and cobinamide (in panel D).

DISCUSSION

We developed a model of B₁₂ usage and production among the different species of Thermotogales (6). Here, we add evidence to this model that Tt. lettingae utilizes both cobinamide and B₁₂ and can produce B₁₂ from cobinamide. We also show in vitro that three different Thermotogales BtuF proteins can bind both B₁₂ and cobinamide, indicating its importance in importation of these molecules into the cell. Our previous paper on the ability of Th. africanus to synthesize B₁₂ de novo and the inability of Tt. maritima to synthesize B₁₂ (6) and the salvage of cobinamide to synthesize B₁₂ by Tt. lettingae shown here demonstrate the diversity of B₁₂ usage and production in this order (Fig. 1).

The growth rate of Tt. lettingae increases upon the addition of B₁₂ or cobinamide to its growth medium. Tt. lettingae is able to synthesize B₁₂ when cobinamide is provided. All seven cobinamide salvage pathway genes had significantly higher transcription when cells were grown without B₁₂ or cobinamide than when these additives are present (Table 1). These data support the hypothesis that this gene cluster allows for B₁₂ synthesis from cobinamide.

Our working hypothesis is that high concentrations of B₁₂ activate the B₁₂ riboswitch that represses the transcription of the salvage pathway genes, resulting in extremely low transcription. Cobinamide appears to similarly activate the riboswitch but perhaps less than B₁₂ does, allowing modest transcriptional activity. It has been demonstrated that the corrin ring is important for E. coli btuB riboswitch binding for four different B₁₂ derivatives: B₁₂, coenzyme B₁₂, adenosyl factor A, and adenosyl-cobinamide. B₁₂ and adenosyl-cobinamide A do not have the upper axial ligand or the lower nucleotide ligand found in coenzyme B₁₂ (see Fig. 1 from reference 36). The Thermotogales riboswitch may be similar to that of E. coli btuB and binds to corrins with little specificity for the axial ligands. Consequently, it is unsurprising that both B₁₂ and cobinamide elicit similar gene expression responses. Apparently, the low, constitutive transcription of the salvage and transporter genes is sufficient to allow B₁₂ synthesis since addition of cobinamide resulted in a similarly increased growth compared to the addition of B₁₂.

The btuF gene of Tt. lettingae, which encodes a putative substrate-binding protein of the BtuFCD ABC transporter, is involved in transporting B₁₂ and cobinamide, as shown by transcription analysis of btuF and the protein-binding assay. Our qRT-PCR results suggested that btuF is regulated at the transcriptional level, possibly by a B₁₂ riboswitch (Table 1). BtuF proteins from Tt. lettingae, Tt. maritima (an organism that cannot salvage cobinamide), and F. nodosum (an organism predicted to salvage several forms of cobinamides) were shown to interact with low concentrations of B₁₂ and cobinamide (Fig. 2A and B) and bind both ligands in vitro (Fig. 2C and D). Additionally, there was no significant change in transcription of btuF when Tt. maritima was grown with cobinamide, consistent with its inability to use cobinamide. Tt. maritima likely utilizes BtuF to import B₁₂ for use in B₁₂-dependent pathways; however, why this protein also binds cobinamides is less apparent. The E. coli BtuF protein shows a similar ability to bind a variety of corrins (37), so this feature of the Tt. maritima BtuF may simply be a function of the lack of specificity of BtuF-type proteins. This low specificity likely was selected for in Thermotoga ancestors who had the cobinamide salvaging pathway (6). Alternatively, maybe there is a function for cobinamides that is not yet known for which further investigations are required to elucidate.

Tt. lettingae is able to grow without the addition of B₁₂ or cobinamide, suggesting that either B₁₂ is not essential for growth or the cell growth requires only an amount of B₁₂ that is below the detection limit of our assay. Supplementing B₁₂ or cobinamide significantly promoted its growth rate. This suggests that Tt. lettingae may use B₁₂ for enzymatic processes (including methionine synthesis). Indeed, Tt. lettingae contains six putative B₁₂-dependent proteins (see Table S2 from reference 6). One such protein is a B₁₂-dependent methionine synthase (MS); no B₁₂-independent MS or known alternative pathways for methionine synthesis are present in the Tt. lettingae genome. The medium used to grow Tt. lettingae, TL-1 (see Table S1 in the supplemental material), contained vitamin-free casein, which contains trace amounts of methionine, so it is possible that methionine is supplied by the medium. We were able to grow Tt. lettingae only with media that contained casein or yeast extract. Without B₁₂, Tt. lettingae may have only one method of acquiring the essential amino acid methionine, and the energy used in this acquisition may be substantial. Perhaps the increase in growth rate seen when cells were grown with B₁₂ is due to the activation of B₁₂-dependent MS to produce methionine. To test this, Tt. lettingae was grown on the TL-1 medium (B₁₂ and cobinamide deficient) with and without methionine (92.5 nM). The addition of methionine caused no significant increase in its growth rate (data not shown). Furthermore, the addition of all 20 amino acids resulted in no apparent change in its growth rate (data not shown). This suggests that Tt. lettingae may be using B₁₂ for other processes in addition to methionine synthesis, which is consistent with the fact that Tt. lettingae contains six putative B₁₂-dependent proteins (see Table S2 from reference 6).

Thus far, our results are consistent with our model (Fig. 1) and the evolutionary history we previously proposed (6). As more sequenced Thermotogales genomes become available and more studies are done on this lineage, perhaps using genetic analyses, the model can be more thoroughly tested and the evolutionary history further refined.