Elucidation of roles for vitamin B₁₂ in regulation of folate, ubiquinone, and methionine metabolism

Significance

Using a chemical probe mimic of vitamin B₁₂, we reveal a light- and B₁₂-dependent DNA regulator, and make the unexpected discovery of B₁₂ having regulatory involvement in microbial folate, ubiquinone, and methionine processes. These findings suggest a pivotal role for B₁₂ in the control of cell growth, which may lead to coordination of cell behavior in complex multicellular systems. As key research questions emerge from host-associated and environmental microbiomes, we anticipate that B₁₂ regulatory control of metabolism will be found to be generalizable, will be critical for coordination of individual microbe and community metabolism, and that organismal interdependencies for B₁₂ may be pertinent to microbiome organization, stability, and overall function.

Abstract

Only a small fraction of vitamin B₁₂-requiring organisms are able to synthesize B₁₂ de novo, making it a common commodity in microbial communities. Initially recognized as an enzyme cofactor of a few enzymes, recent studies have revealed additional B₁₂-binding enzymes and regulatory roles for B₁₂. Here we report the development and use of a B₁₂-based chemical probe to identify B₁₂-binding proteins in a nonphototrophic B₁₂-producing bacterium. Two unexpected discoveries resulted from this study. First, we identified a light-sensing B₁₂-binding transcriptional regulator and demonstrated that it controls folate and ubiquinone biosynthesis. Second, our probe captured proteins involved in folate, methionine, and ubiquinone metabolism, suggesting that it may play a role as an allosteric effector of these processes. These metabolic processes produce precursors for synthesis of DNA, RNA, and protein. Thereby, B₁₂ likely modulates growth, and by limiting its availability to auxotrophs, B₁₂-producing organisms may facilitate coordination of community metabolism.

Vitamin B₁₂ encompasses a group of closely related corrinoid compounds best known for their role as cofactors of enzymes that mediate methyl transfer reactions, isomerase rearrangements, and dehalogenation (1). Although a common B₁₂-binding motif can be used to mine sequences for the presence of novel B₁₂-dependent enzymes (2, 3), new enzymes that lack the canonical motif continue to be discovered (4). Vitamin B₁₂ also regulates protein expression by binding to riboswitches (5), which are regulatory RNA. These riboswitches primarily attenuate synthesis of proteins associated with B₁₂ biosynthesis and salvage, but recent studies in the human microbiome suggest that nearly half of the 4,000 riboswitch-regulated proteins discovered are not associated with these processes, suggestive of considerable discovery opportunities (2). Recent studies have also revealed that vitamin B₁₂ is a cofactor of transcriptional regulators and antirepressors (2, 3, 6). The upper portion of vitamin B₁₂ is photolabile; thus, when bound to these proteins, it provides light-dependent regulation of transcription, controlling processes, such as biosynthesis of carotenoids, tetrapyrroles, and photosystems (6, 7).

The de novo biosynthesis of B₁₂ occurs only in bacteria and archaea, yet is used by all domains of life (8). B₁₂ is energetically costly, requiring nearly 30 different enzymes for its synthesis (1, 9), which likely explains why only a small fraction of prokaryotes have the genetic capacity to produce it. In natural microbial communities, from eukaryote-associated to those occurring within terrestrial, aquatic, or other environments, the limited sources of B₁₂ makes it a precious commodity (10). As such the availability of, and access to, B₁₂ has been suggested to impart a fundamental contribution to their spatial and functional organization (3).

To examine the role of metabolic interactions in community structure and function, we developed two nearly identical model photoautotrophic microbial consortia derived from a naturally occurring aquatic mat (11). Each consortium consists of one distinctive photosynthetic cyanobacterium and up to 18 heterotrophic bacteria. Through genome annotation we have identified extensive vitamin B₁₂ auxotrophy among the members, with only the cyanobacterium and three heterotrophs from each community capable of B₁₂ biosynthesis (12). All three heterotrophs have been isolated and can also salvage B₁₂, and are therefore receptive to exogenously supplied B₁₂ making probing of live cells with B₁₂ mimics feasible. However, only one, Halomonas sp. HL-48 (hereafter, Halomonas), can be propagated in defined media in which levels of exogenously supplied B₁₂ can be controlled.

Here we describe the development of an affinity-based vitamin B₁₂ probe (B₁₂-ABP), and its in situ application to Halomonas. Our probe selectively captured 41 proteins from Halomonas, including enzymes known to use it as a cofactor, a transcriptional regulator, three enzymes in the one carbon pool by folate, and two enzymes in ubiquinone biosynthesis. We demonstrate that the captured regulator binds—in a light-dependent manner—a conserved DNA motif upstream to several genes, including four that are associated with either folate or ubiquinone biosynthesis. The unexpected discovery of B₁₂ involvement in these processes suggests a pivotal role in the control of cell growth in response to photostress, potentially leading to coordination of cell behavior in complex multicellular systems.

Results

Development of an Affinity-Based B₁₂ Probe That Mimics Natural Vitamin B₁₂.

With the goal of conducting live-cell probe-labeling studies, we sought to develop a vitamin B₁₂ affinity-based probe that would be recognized, transported, and bind proteins akin to native vitamin B₁₂. Prior research with fluorescent analogs of B₁₂ found that modification of any of the carboxamide nitrogens emanating off the tetrapyrrole core resulted in moderate to severe impacts on protein binding affinity (13–15). However, modification of the ribose-5′-hydroxy moiety of the ribose sugar resulted in biologically well-tolerated B₁₂ analogs (16–21). To enable covalent bond-forming reactions between B₁₂-ABP and target proteins, we developed a “linker” moiety to attach at the 5′-hydroxy position of B₁₂ (Fig. 1A) (22). The linker, 3, contains a diazirine for irreversible photo–cross-linking of the probe to B₁₂-binding proteins, and an alkyne group to enable click chemistry that is used in postlive cell labeling to append fluorophores for fluorescence measurements or biotin for enrichment of probe targets, and subsequent LC-MS analysis. B₁₂-ABP synthesis required four steps to develop the linker and append it to the ribose-5′-hydroxy position (Fig. 1A).

Fig. 1.

B₁₂-ABP synthesis. (A) Cyanocobalamin (CNB₁₂) was converted to a photoreactive probe by carbamate bond formation to a linker moiety containing a diazirine and click chemistry-compatible alkyne. (B) B₁₂-ABP was added directly to live Halomonas cells, incubated for 60 min, then the samples were UV-irradiated to induce covalent bond formation between the probe and B₁₂-binding proteins. Cells were lysed, azido-biotin was added to probe-labeled proteins by click chemistry, and labeled proteins were enriched on a streptavidin agarose resin. Enriched proteins were digested on-resin, followed by quantitative LC-MS proteomic analysis of probe-labeled proteins by the accurate mass and time tag method.

To confirm that the probe binds proteins akin to native B₁₂, we demonstrated that transcobalamin, a B₁₂-binding protein, was probe-labeled in a concentration-dependent manner and that the probe was outcompeted by addition of excess native cyanocobalamin (CNB₁₂) (Fig. S1A). Live cell uptake and protein labeling by B₁₂-ABP in Halomonas was experimentally validated by fluorescence gel analysis (Fig. S1B). We also expressed and purified the B₁₂-dependent enzyme, MetH, from Halomonas sp. HL-48 and demonstrated labeling by B₁₂-ABP and ablation of the binding upon addition of excess CNB₁₂ (Fig. S2).

Fig. S1.

Assays of B₁₂-ABP specificity and live cell uptake. (A) Competition assays of B₁₂-ABP binding to transcobalamin, and (B) SDS/PAGE analysis of analysis of Halomonas HL-48 lysates after labeling with B₁₂-ABP. Fluorescence (Left) and Coomassie-stained (Right) images of the same gel show proteins bound by 0 (NP, no probe control), 10, 20, or 50 μM B₁₂-ABP probe.

Fig. S2.

Proteins (A) MetE, (B) the B₁₂-binding domain of MetH, (C) PhrR, and (D) FolD were recombinantly expressed and purified, then the proteins (2 μM) were labeled with B₁₂-ABP (2 μM). Protein labeling was also competed by concomitant addition of excess CNB₁₂ (10×, 25×, and 50× concentrations versus the B₁₂-ABP). Following labeling, azido-rhodamine was attached to B₁₂-ABP–labeled proteins via click chemistry, and the labeled protein solutions were separated by SDS/PAGE. The fluorescence gel images in the Left panels of A–D show probe labeling of the pure proteins and subsequent inhibition by increasing concentrations of CNB₁₂; panels on the Right side show Coomassie staining of protein abundance, displaying equivalent loading of gel lanes. Intensity of gel lanes was quantified using ImageJ, and shown in E as percent intensity compared with B₁₂-ABP labeling without competitive inhibition; n = 3 for the labeling and quantification experiments.

Affinity-Based Protein Profiling of Halomonas sp. HL-48.

Our analyses (12) of the Halomonas sp. HL-48 genome revealed that it encodes machinery for synthesis and salvage of B₁₂ and three B₁₂-dependent enzymes; thus, we anticipated that our probe would capture proteins involved in these processes. B₁₂-ABP was added directly to live Halomonas sp. HL-48 cells after they had reached exponential growth in B₁₂-deplete defined media (Fig. 1B). After 60 min, cells were UV-irradiated to stimulate probe–protein covalent bond formation via the diazirine moiety on B₁₂-ABP. Probed proteins were enriched by biotin affinity purification and quantitatively analyzed by LC-MS–based proteomics. B₁₂-ABP labeling of proteins was also competed with addition of excess CNB₁₂, and the competition experiment was quantitatively analyzed by proteomics. Accounting for only those proteins in which excess native CNB₁₂ results in statistically significant diminished B₁₂-ABP protein labeling, we identified a total of 41 proteins (summarized in Table 1; full details in Dataset S1). Five of these proteins are components of three enzymes known to use B₁₂ as a cofactor, namely the methionine synthase (MetH), ribonucleotide reductase (NrdJa/NrdJb), and ethanolamine ammonia-lyase (EutB/EutC). The probe also detected cob(I)alamin adenosyltransferase (BtuR), an enzyme that reacts with B₁₂ and its biosynthetic precursors as a substrate, catalyzing their adenylosylation during de novo biosynthesis or salvage. CobW, also bound by the probe, is known to be involved in B₁₂ biosynthesis, but its exact function is still not clear. It has been proposed to be involved in cobalt insertion into a corrinoid precursor (23). It should be noted that we did not detect the vitamin B₁₂ transporter, which we speculate is because of the low levels of the probe still retained on the transporter at the time UV was applied, or it may have been excluded from proteomic analysis because of it being a membrane component.

Table 1.

Categorization of 41 proteins detected by B₁₂-ABP

Functional category	No. of proteins	Protein designation
Nucleotide metabolism	2	NrdJa*, NrdJb*
Ethanolamine metabolism	2	EutB*, EutC*
Methionine metabolism	6	MetH*, MetE, LuxS, MsrB, MetK, MetZ
Porphyrin synthesis	3	HemE, CobW*, BtuR*
SAM/SAH-dependent enzymes	8	QueA, UbiE, UbiG, Rlml, RlmL, RlmN, RsmB, COG4106
Folate cycle	2	FolD, MetF
Regulation	2	PhrR, Mcp
Amino acid metabolism	2	DapC, AstB
Other systems	11	Hcp, Ssb, MmsA1, MmsA2, AptA, GatA, GapA, AtoB, RplN, TadA, CY41DRAFT_2254
Unknown function	3	CY41DRAFT_0508, CY41DRAFT_0804, CY41DRAFT_0947

^*Proteins that are components of enzymatic complexes known to use B₁₂ as a cofactor or substrate.

To confirm that probe-labeled proteins are not artifacts of the labeling process, proteomic controls were performed. First, samples treated with DMSO only (no probe), but for which all click chemistry steps were performed akin to the probed samples, were analyzed. Second, competition experiments with CNB₁₂ were performed in which the B₁₂-ABP (2 μΜ) and CNB₁₂ (100 μM) were added concomitantly to HL-48 proteins. The LC-MS proteomics data were analyzed, and the control samples were used to statistically validate probe labeling (Dataset S1). For additional validation, we expressed and purified FolD, MetE (Table 1), and MetH (positive control). These proteins were labeled with B₁₂-ABP, and labeling was competed by addition of excess CNB₁₂ during the labeling experiment, which resulted in significantly inhibited or near complete ablation of labeling (Fig. S2). Additionally, the carbene generated by UV irradiation of the probe diazirine is rapidly protonated by water (24–26), so it is highly unlikely that the probe with the carbene moiety transiently diffused away from one protein through aqueous cytosol and incidentally labeled other proteins. As demonstrated by the labeling and competitive inhibition of purified MetH, FolD, and MetE, it is probable that the proteins identified in our proteomic measurements of B₁₂-ABP labeling do indeed rely on B₁₂ binding, or they are in very tight multienzyme complexes. Finally, we also performed a global proteomic analysis of Halomonas HL-48 cells. When comparing the order of quantitative values from the global analysis to the B₁₂-ABP chemoproteomic analysis, meaning the highest to lowest values for the 41 B₁₂-binding proteins, they are correlative. This finding indicates that the probe labeling results in specific binding events, and does not follow the order of protein abundance; in fact, three proteins were not detected in the global proteome analysis that were identified by B₁₂-ABP chemoproteome analysis. In summary, our results confirm that the probe binds to and labels expected enzymes that require B₁₂ as a cofactor or use it as a substrate, and identify 34 candidate B₁₂-binding proteins.

A Potential Allosteric Control Role for Vitamin B₁₂.

Three probe-labeled proteins were identified that are involved at different points of the tetrapyrrole biosynthetic pathway that yields heme and B₁₂ in Halomonas. The probe labeled uroporphyrinogen decarboxylase (HemE), which catalyzes the first reaction in heme biosynthesis from uroporphyrinogen III. This metabolite is also the precursor to vitamin B₁₂ biosynthesis and, therefore, these anabolic processes compete for the same precursor. Allosteric control of HemE would provide Halomonas a means by which to control flux through these pathways based on B₁₂ availability, and suggests a fundamentally new role for B₁₂ in cellular metabolism. Prior reports on control of the tetrapyrrole biosynthetic pathway in other microbes have identified regulatory feedback controls by B₁₂-dependent riboswitches (27) and redox signaling cascades (28). Taking these data together, we find that vitamin B₁₂ regulation of these steps could result in redirection of metabolism between biosynthesis of heme versus B₁₂ biosynthesis.

B₁₂ Interdependencies in Folate and Methionine Metabolism.

Examination of additional enzymes bound by the B₁₂-ABP revealed a remarkable connection to processes linked by methionine synthase. Two variants of methionine synthase, MetH and MetE, are encoded by Halomonas and responsible for conversion of homocysteine to methionine (Fig. 2). Both enzymes were captured by the B₁₂-ABP, yet only MetH is known to depend on B₁₂ for function. In many bacteria, MetE translation is repressed by an upstream cobalamin-binding riboswitch (29). In Halomonas the transcription of metE (CY41DRAFT_1840) is activated by MetR (CY41DRAFT_1841), but no upstream riboswitch for repression of MetE was detected (12). A new mechanism of control involving an allosteric interaction between MetE and B₁₂ is suggested by our results. To confirm that MetE binds B₁₂, we expressed and purified the enzyme and labeled it with B₁₂-ABP, and also demonstrated that addition of excess CNB₁₂ during the labeling experiment results in significantly inhibited probe labeling (Fig. S2). Additionally, given the number of replicate analyses that were performed, if probe labeling of the methionine cycle and 5-methyl tetrahydrofolate (5mTHF) recycling pathways was purely ancillary, the proteomic results would likely be highly variable, but they are not (Dataset S1).

Fig. 2.

B₁₂-ABP captures 17 proteins in methionine, folate, and ubiquinone metabolism. Metabolites are shown in open boxes: 5,10-CH = THF, 5,10-methenyltetrahydrofolate; 5,10-CH2-THF, 5,10-methylene-THF; 5mTHF, 5-methyl-THF; 10f-THF, 10-formyl THF; DHF, dihydrofolate; H₂-MPt, dihydromonapterin; H₄-MPt, tetrahydromonapterin; H₂-NPt-P3, dihydroneopterin triphosphate; H₂-NPt, dihydroneopterin; HCY, homocysteine; Met, methionine; pABA, p-aminobenzoate; SAH, S-adenosylhomocysteine; SRH, S-ribosylhomocysteine; THF, tetrahydrofolate;. Enzyme abbreviations: CheR, chemotaxis signal relay system methyltransferase; FolD, bifunctional methylenetetrahydrofolate dehydrogenase (NADP+)/methenyltetrahydrofolate cyclohydrolase; FolE, GTP cyclohydrolase; FolK, 2-amino-4-hydroxy-6-hydroxymethyldihydropteridine diphosphokinase; FolM, alternative dihydrofolate reductase 1; GlyA, glycine hydroxymethyltransferase; LuxS, S-ribosylhomocysteine lyase; MsrB, methionine-R-sulfoxide reductase; PurN, phosphoribosylglycinamide formyltransferase 1; MetE, B₁₂-independent methionine synthase; MetF, 5,10-methylenetetrahydrofolate reductase; MetH, B₁₂-dependent methionine synthase; MetK, S-adenosylmethionine synthetase; Me-MCP, methyl accepting chemotaxis protein; MtnN, adenosylhomocysteine nucleosidase; MetX, homoserine O-acetyltransferase; MetZ, O-succinylhomoserine sulfhydrylase; PanB, 3-methyl-2-oxobutanoate hydroxymethyltransferase; PurU, formyltetrahydrofolate deformylase; ThyA, thymidylate synthase; UbiB, ubiquinone biosynthesis monooxygenase; UbiG, bifunctional 2-polyprenyl-6-hydroxyphenyl methylase/3-demethylubiquinone-9 3-methyltransferase; UbiE, 2-methoxy-6-polyprenyl-1,4-benzoquinol methylase; UbiI, 2-octaprenylphenol hydroxylase. ROS, reactive oxygen species.

The B₁₂-ABP also captured all three enzymes needed to synthesize 5mTHF, the methyl donor used in the MetH reaction, and five enzymes associated with methionine metabolism and repair (Fig. 2). These proteins are not known to be B₁₂-dependent; however, they are used in pathways that are linked by MetH and thereby may be probe-labeled because of close proximity effects. In correlation to the role B₁₂ plays in methionine cycling, nine S-adynosyl methionine (SAM)-dependent enzymes were probe -labeled (Table 1). Most of these enzymes are methyltransferases involved in the modification of rRNA and tRNA, or synthesis of ubiquinone. In total, the B₁₂-ABP–labeling results point to significant and previously unknown roles in control of methionine and 5mTHF recycling, and the processes in which intermediates are used.

Identification of a B₁₂-Dependent DNA Transcription Factor PhrR.

Probe labeling of Halomonas resulted in the identification of a B₁₂-dependent transcription factor from the MerR family, which was named PhrR (Table 1). Comparative genomics analysis of PhrR orthologs in Proteobacteria suggests that they belong to the previously uncharacterized group of light-controlled regulators of genes coding for DNA photolyases and other light dependent processes (see below). N-terminal DNA-binding domains in PhrR proteins are 41–49% similar to DNA-binding domains of the previously characterized B₁₂-dependent light-inducible regulators LitR from Bacillus megaterium and Thermus thermophilus and CarH from Myxococcus xanthus (30–32). However, PhrR proteins lack a canonical C-terminal B₁₂-binding domain, and would not be characterized as B₁₂-binding proteins by BLAST and domain searches using the trusted cut-off. B₁₂ is known to act as a photosensitive regulator of transcription factors, where photolysis of B₁₂ leads to altered DNA binding (7, 32). In Halomonas sp. and other Proteobacteria, orthologs of phrR are often colocalized on the chromosome with genes involved in photo-damage repair (phr) and cyclopropane fatty acid biosynthesis (cfa) (Fig. 3 and Dataset S2). Both of these activities are beneficial under light stress, further supporting the idea that PhrR is a B₁₂-dependent light-sensitive transcriptional regulator. Subsequently, we set out to more fully characterize the B₁₂-dependency, light regulation, and regulatory role of PhrR in Halomonas.

Fig. 3.

Comparative genomics reconstruction of PhrR regulons in Gammaproteobactreria. (A) Predicted genes and candidate operons under regulation by PhrR in Halomonas HL-48. Genes, candidate PhrR-binding sites, and putative promoters are shown as rectangles, yellow circles, and small arrows, respectively. Sequence logo for PhrR-binding motif in the Halomonadaceae is shown in a box. Names and locus tags for PhrR-regulated genes are shown on top and bottom lines, respectively. The phrR (regulator) and phr (DNA photolyase) genes are in black and yellow, respectively. Genes in green and orange are involved in folate biosynthesis (fol) and cyclopropane fatty acid biosynthesis (cfa), respectively. Conserved members of the PhrR regulons in Gammaproteobacteria encoding functionally uncharacterized proteins (designated by Pfam/COG family numbers) are shown by gray rectangles. Experimentally tested PhrR sites are marked with a “T” in yellow circles. (B) Conserved core of PhrR regulons in 20 genomes of the Halomonadaceae. The table shows gene orthologs that are predicted to be regulated (light green squares) or not regulated (pink squares) by PhrR in each analyzed genome. The absence of a gene ortholog is shown by a blank space. (C) Genomic organization of PhrR-controlled loci in other Gammaproteobacteria.

Reconstruction of PhrR Regulons in Halomonas and Other Gammaproteobacteria.

Orthologs of phrR were identified in all 20 Halomonas species with sequenced genomes. In most of these genomes, phrR is clustered on the chromosome with the photolyase gene phr, suggesting it is a primary target gene for PhrR-dependent transcriptional regulation. We applied the comparative genomics approach to reconstruct the PhrR regulons. A conserved 21-bp palindrome was identified as a candidate PhrR-binding motif (Fig. 3A). The reconstructed PhrR regulons in the Halomonas genomes include several genes involved in light-dependent processes, such as DNA photolyases (phr, phr2), a blue light- and temperature-regulated antirepressor (bluF), the photoactive yellow protein (pyp), three folate biosynthesis genes (folE, folK, folM), two methyl-accepting chemotaxis proteins (mcp1, mcp2), one ubiquinone biosynthetic gene (ubiB), and several hypothetical enzymes and uncharacterized proteins (Fig. 3B; for detailed list of binding sites detected and domains and their putative functional roles, see Datasets S2 and S3). The comparative analysis of upstream gene regions in multiple Halomonas genomes (Fig. S3) suggests that PhrR likely functions as a repressor of its target genes, similarly to LitR and CarH (30–32).

Fig. S3.

(A) Palindromic DNA motifs are shown as sequence logos that were constructed by WebLogo using sequences of all predicted PhrR-binding sites for each taxonomic group. Candidate PhrR-binding motifs in different lineages of Gammaproteobacteria are characterized by similar 7-bp half-sites and an internal linker of variable length. (B) Common consensus of all PhrR-binding DNA motifs identified in Gammaproteobacteria. (C) Sequence logo of known binding sites for homologous light-inducible B₁₂-dependent regulator LitR from Bacillus megaterium and related Bacillus spp. (D) Consensus of known binding sites for homologous light-inducible B₁₂-dependent regulator LitR from Thermus thermophiles.

Orthologs of phrR were also identified in several species that belong to other lineages of Gammaproteobacteria, where they are also colocated with phr (Fig. 3C). By applying a similar bioinformatics approach, we identified DNA binding site motifs for these PhrR orthologs (Fig. S4). In most of the genomes, the reconstructed PhrR regulons control from one to four candidate operons (Datasets S2 and S3). This finding is in contrast with Halomonas spp., which possess up to 14 candidate PhrR-controlled operons per genome.

Fig. S4.

Phylogenetic footprinting of upstream regions of predicted PhrR regulated operons in Halomonas spp. Each gene is named according to Dataset S3; gene locus tags in Halomonas sp. HL-48 are given in parentheses. Candidate PhrR-binding sites are highlighted in yellow. Consensus sequences of the PhrR motif are shown in the top line in red. Nucleotides in the PhrR binding sites that correspond to the consensus motif are in red. PhrR binding site scores are given to the right of the first line of sequence for each entry. Strong binding sites have a score above 4.5; weak sites have scores between 4.0 and 4.5. Putative promoter elements (−35 and −10 boxes) are underlined. Coding regions of genes that are immediately downstream to PhrR binding sites are in blue.

Structural Analysis of PhrR Proteins.

The PhrR proteins from Halomonas species are distantly related to the B₁₂-dependent repressors CarH from M. xanthus and LitR from T. thermophilus and B. megaterium. LitR and CarH regulators are characterized by three Pfam domains: PF13411 (MerR HTH), PF02607 (B₁₂-binding-2), and PF02310 (B₁₂-binding). The PhrR proteins have highly conserved DNA binding domains from the MerR family, and their C-terminal domains are very weakly similar to those in LitR/CarH. The structure of the B₁₂-binding domains in the T. thermophilus LitR regulator was recently solved in complex with B₁₂ (PDB ID code 3WHP) (7). We used the structure-based multiple alignment of the PhrR and LitR/CarH proteins to check the conservation of B₁₂-binding residues in their C-terminal domains (Fig. S5). Although the proteins seem to be structurally similar, the potential B₁₂-binding residues are not conserved in PhrR regulators. The histidine residue located in the beginning of the PF02310 domain, which was found to be functionally important in the previous studies of LitR and CarH proteins, is not conserved in any of the studied PhrR proteins. These observations suggest that the identified PhrR proteins in Gammaproteobacteria are characterized by highly diverged B₁₂-binding domains (often not detectable by Pfam search) that use a different pattern of residues for interaction with B₁₂. Future structure-functional studies will be required to identify these functionally important residues in the PhrR family.

Fig. S5.

Multiple alignment of Gammaproteobacterial PhrR regulators and homologous LitR and CarH regulators. The sequence alignment was constructed using ClustalX. Gene locus tags and species names are listed in Dataset S2. PhrR proteins are characterized by an N-terminal DNA-binding domain from the (A) MerR family and (B and C) two C-terminal B₁₂-binding domains. Secondary structure elements in the B₁₂-binding domain according to the known 3D structure of the T. thermophilus LitR (PDB ID code 3WHP) are shown by pink boxes (α-helices) and yellow arrows (β-strands). Residues involved in B₁₂ binding in 3WHP (as calculated in the PDB database) are shown by green circles.

Comparison of PhrR- and LitR-Binding DNA Motifs.

The predicted lineage-specific PhrR-binding motifs demonstrated significant conservation across the analyzed taxonomic groups of Gammaproteobacteria (Fig. S4). The conserved consensus of palindromic PhrR motifs is TRTACAa-(flexible linker)-tTGTAYA. However, the length of internal linker between two conserved half-sites in PhrR motifs showed a remarkable flexibility. The PhrR half-sites are separated by 3 bp in the Halomonadaceae, Chromatiales/Thiotrichales, 9 bp in Vibrionales and Marinomonas spp., 10 bp in Pseudomonadales, and 11 bp in Aeromonas and Shewanella spp. Interestingly, the consensus half-site DNA motifs of PhrRs are similar to the experimentally determined DNA sites of the LitR regulators from B. megaterium and T. thermophilus (Fig. S4), however, the linkers between half-sites in the latter operators are 14 bp in length (30, 33). The similarity between half-site DNA motifs of PhrR operators correlates with high conservation of DNA-binding domains in PhrR and LitR (see above).

Functional Analysis of PhrR.

To validate the computationally predicted DNA-binding motif of PhrR, we heterologously expressed and purified the PhrR protein from Halomonas. The recombinant PhrR protein exists partially as a dimer in solution (Fig. S6), supporting the hypothesis that the PhrR dimer binds to its cognate palindromic DNA motif. Additional spectrometry of the monomer fraction of PhrR incubated with fourfold molar excess of vitamin B₁₂, a prospective ligand of PhrR, demonstrated specific ligand binding to the protein (Fig. S7).

Fig. S6.

Gel-filtration analysis for the purified recombinant PhrR protein from Halomonas sp. HL-48 after refolding. Retention volumes of 60 and 70 mL correspond by size to the dimer and monomer fractions of the protein, respectively.

Fig. S7.

Spectrometry of recombinant PhrR protein binding to B₁₂. UV spectrum of the monomer fraction of PhrR incubated with fourfold molar excess of vitamin B₁₂ (cyanocobalamin), a prospective ligand of PhrR, is shown by red line. As a control, UV spectrum of the recombinant PhrR protein in the absence of ligand is shown by black line.

The interaction of the purified PhrR protein with its cognate DNA motif was assessed using a fluorescence polarization assay. The results show that PhrR specifically binds to the synthetic DNA fragment containing the consensus PhrR-binding site, TTGTACAAtttTTGTACAA (Fig. 4A). The apparent dissociation constant (K_d) value for the apo-PhrR protein interacting with the tested DNA fragment was 77 nM. The addition of potential ligands, CNB₁₂ (2 µM) and adenosylcobalamin (AdoB₁₂, 4 µM), improved the PhrR–DNA complex formation, resulting in decreased K_d values, 57 ± 9 nM and 25 ± 8 nM, respectively. Titration of the effect of B₁₂-ABP and CNB₁₂ on the interaction between PhrR (20 nM) and the DNA fragment revealed the apparent K_d values 4 ± 0.8 μM and 0.4 ± 0.2 μM for B₁₂-ABP and CNB₁₂, respectively (Fig. 4B), demonstrating that the probe binds B₁₂-dependent proteins akin to native vitamin B₁₂. We further tested the effect of illumination on the interaction of AdoB₁₂-PhrR with DNA. The dark-incubated AdoB₁₂-PhrR protein demonstrated specific binding to the consensus DNA motif, whereas illumination with white light for 5 min results in failure of PhrR to bind to the same DNA fragment (Fig. 4C). We also tested the interaction of AdoB₁₂-PhrR with six DNA fragments containing the predicted PhrR operators in Halomonas sp. HL-48 (Fig. 4 D and E). All tested DNA fragments demonstrated a concentration-dependent increase of fluorescence polarization, confirming specific interaction between the regulator and DNA fragments. As a negative control, we assessed the binding of PhrR with a DNA fragment containing a TrpR-binding site in Halomonas sp. HL-48. The apparent K_d values for the PhrR protein interacting with two high-scored PhrR binding sites from candidate promoter regions of the CY41DRAFT_3441 and CY41DRAFT_2637 genes were in the range of 30–40 nM, whereas the low-scored PhrR sites found upstream of the pyp, phrR, folK/M, and folE genes have showed the apparent K_d values 60 ± 10 nM, 77 ± 36 nM, 103 ± 30 nM, and 16 ± 6 nM, respectively. Finally, further confirming the interaction between B₁₂ and PhrR, B₁₂-ABP labeling of purified PhrR is inhibited by addition of CNB₁₂ (Fig. S2 and Dataset S1).

Fig. 4.

Experimental validation of the PhrR regulon in Halomonas sp. HL-48 by fluorescence polarization (FP) binding assay. (A) Interactions between the recombinant PhrR protein and a DNA fragments containing consensus PhrR-binding site (6 nM) shows that DNA binding is specific and enhanced in the presence of CNB₁₂ or adenosylcobalamin (AdoB₁₂). (B) Titration of the effect of B₁₂-ABP and CNB₁₂ on the interaction between the recombinant PhrR protein (20 nM) and its predicted consensus DNA binding site (2.5 nM). (C) Light disrupts AdoB₁₂-dependent binding of PhrR to DNA. The PhrR protein (3 µM) was preincubated with AdoB₁₂ (66 µM) in the dark or irradiated with light for 5 min and then 0.7, 1.5, and 3 µM of the resulting PhrB:AdoB₁₂ complexes were checked for interaction with its consensus DNA motif (6 nM). (D and E) Effect of increasing concentrations of PhrR mixed with 33-bp DNA fragments (10 nM) containing candidate PhrR binding sites in the presence of AdoB₁₂ (4 µM). (F) Sequences of validated DNA fragments containing consensus PhrR-binding site and natural PhrR sites from Halomonas sp. HL-48 genome. Sequence logo represents the consensus PhrR-binding motif in the Halomonadaceae.

New Mode of Light-Dependent Global Gene Regulation Controlling Fatty Acid and Folate Biosynthesis.

The comparative analysis of upstream promoter sequences in multiple Halomonas genomes (Fig. S4) suggests that PhrR functions as a repressor of its target genes. Thus, the PhrR regulon genes are predicted to be de-repressed after exposure to light. To validate this bioinformatics prediction, we evaluated the gene-expression pattern of selected genes from the PhrR regulon by quantitative RT-PCR (qRT-PCR) analysis. The expression profiles of three folate biosynthetic genes (folE, folK, and folM) from two different growth conditions (either constant light or dark) were tested. All three genes tested showed up-regulation of expression when cells were grown in the light compared with those grown in the dark (Fig. 5A). These results indicate that the regulation of PhrR is similar to the recently described CarH (7), in which B₁₂ serves as a light sensor to modulate its activities, thus resulting in light-dependent gene regulation.

Fig. 5.

Effect of light vs. dark growth conditions on Halomonas HL-48 gene expression and intracellular metabolite production. (A) Relative gene-expression levels of PhrR regulated folate genes; rpoB is a control not regulated by PhrR. n = 3. (B) Cyclopropane fatty acid levels determined by metabolomic analysis in WT Halomonas HL-48 and a PhrR mutant. n = 3. (C) Levels of folate derivatives; 5mTHF, 5-methyl tetrahydrofolate; 5fTHF, 5-formyl tetrahydrofolate; DHF, dihydrofolate. n = 3. Statistically significant differences in measured gene or metabolite levels in A–C were evaluated by t test (n = 3): one asterisk (*): 0.05 < P < 0.1; two asterisks (**): 0.05 < P < 0.001.

To further validate the unexpected linkage between B₁₂ and folate biosynthesis, we measured the intracellular concentrations of folate and its derivatives in wild-type Halomonas and a deletion mutant that lacked phrR grown in light or dark conditions. As anticipated, under light conditions wild-type Halomonas produced higher intracellular concentrations of tetrahydrofolate (THF) (Fig. 5B and Dataset S4). Light-responsive THF production was lost in the mutant and nearly all of the metabolite detected was THF, indicative of uncontrolled production of THF.

Our regulon analyses suggest that expression of a cyclopropane fatty acid (CFA) synthase gene is controlled by PhrR (Dataset S2). To confirm this prediction, cellular levels of CFAs were measured in wild-type and mutant Halomonas. Our results demonstrate that production of CFA in the wild-type was indeed higher during light growth (Fig. 5C) and that the mutant produced significantly more CFA than the wild-type, regardless of the growth condition.

Discussion

Previous to this study, vitamin B₁₂ was recognized as a facilitator of enzyme function, a sensory component of transcriptional regulators that control functions required in light conditions, and an effector of riboswitches that repress protein synthesis. Using our B₁₂-ABP in a nonphotosynthetic organism, we validated that the probe captures proteins expected to interact with B₁₂. We also discovered a transcriptional regulator, PhrR, which uses B₁₂ as a light sensor and identified genes and processes that are under its control, including several that are not obviously linked to light stress response. We also captured proteins with the B₁₂-ABP not expected to bind B₁₂. A remarkable overlap was revealed in processes that were under control of PhrR or captured by the B₁₂-ABP; both were linked to ubiquinone biosynthesis, folate metabolism, and chemotaxis. The unprecedented connection between B₁₂ and these processes suggest that B₁₂ plays an even greater role in coordinating cellular metabolism than previously recognized.

We speculate that our results reflect a role for B₁₂ as an allosteric regulator in Halomonas, controlling metabolic flux between B₁₂ and heme biosynthesis, biosynthesis of ubiquinone, interconversions between THF and 5meTHF, and metabolism of methionine. THF metabolites produced by enzymes bound by B₁₂-ABP are used in the biosynthesis of purines, DNA, CoA, and serine. Control of these enzymes would consequently have significant impact on host metabolism. Our previous genomic analysis of Halomonas HL-48 revealed that it only requires B₁₂ as a cofactor for ethanolamine biosynthetic genes and riboswitch control of the B₁₂ salvage system, thus making it surprising that it can synthesize such an energetically costly metabolite (12). The additional use of B₁₂ to control key metabolic processes could potentially justify the cost of making it and provide a means for this nonphototrophic organism to modulate its metabolism according to day/night cycles, like the cyanobacteria with which they coexist. Taken together, our findings weave an intricate web of B₁₂ regulation on metabolism within Halomonas, and points to a fundamental requirement for B₁₂ in cell metabolism, regulation, and protection.

As key research questions emerge from host-associated and environmental microbiomes, we believe our approach and results suggest that B₁₂ may be critical for coordination of individual microbe and community metabolism, and organismal interdependencies for B₁₂ may be pertinent to microbiome organization, stability, and overall function. We predict that these roles for B₁₂ may be generalizable in myriad communities.

Materials and Methods

Chemical Synthesis of B₁₂-ABP.

For full methodology and characterization, SI Materials and Methods.

B₁₂-ABP Binding Assays with Transcobalamin.

Transcobalamin (human transcobalamin 2, ACRO Biosystems), a known B₁₂-binding protein, was used to demonstrate probe selectivity and confirm that the probe and native B₁₂ bind at the same site on transcobalamin. Transcobalamin (10 µM) in PBS was labeled with B₁₂-ABP (in DMSO) at 37 °C for 1 h, and UV-irradiated on ice at 365 nm immediately following probe labeling. For inhibition studies, transcobalamin (10 µM) in PBS was incubated with native CNB₁₂ (10 or 20 µM) for 10 min, followed by addition of B₁₂-ABP (10 µM) for 30 min. The samples were then UV-irradiated on ice at 365 nm for 10 min. To characterize protein labeling and CNB₁₂ competition by SDS/PAGE, we first incorporated tetramethylrhodamine onto the probe via click chemistry. Azido-tetramethylrhodamine fluorophore (2.65 µM) was added to probe-labeled protein solutions followed by the addition of Tris(2-carboxyethyl)phosphine; (22 µM), TBTA (Tris[(1-benzyl-1H-1,2,3-traizol-4-yl)methyl])amine) in a 4:1 solution t-butanol:DMSO (44.8 µM), copper sulfate (45 µM), and proteins were separated using 10% (wt/vol) Tris-Glycine SDS/PAGE gels. Fluorescence imaging was performed on a Protein Simple FluorchemQ system (Fig. S1A).

Live-Cell B₁₂-ABP Labeling of Proteins in Halomonas HL-48 Detected by Fluorescence.

Halomonas strain HL-48 was grown at 30 °C in Hot Lake Heterotroph (HLH) broth, pH 8.0 (11). For B₁₂-ABP experiments, Halomonas was grown with shaking in 35 mL modified HLH defined medium lacking yeast extract and supplemented with 5 mM sucrose. At the mid-late log-growth phase, cells were collected by centrifugation (6,000 × g, 15 min) and used for live-cell probe labeling. To the pelleted cells was added 0.5 mL PBS with mild vortexing, followed by addition of B₁₂-ABP (10 μM, in 1 μL DMSO), which was incubated with the cells for 1 h at 30 °C. Following incubation, cells were UV-irradiated on ice at 365 nm for 10 min. Cells were washed with PBS (2 × 0.5 mL) and lysed by bead beating using a Next Advance Bullet Blender. Azido-tetramethylrhodamine-azide was added by click chemistry, as described above, and protein labeling was determined by fluorescence gel imaging (Fig. S1B).

Chemoproteomic Identification of B₁₂-ABP Binding Proteins, Competitive Inhibition Studies, and Global Proteomic Analysis of Halomonas HL-48.

For quantitative chemoproteomic identification of B₁₂-ABP binding targets, probe labeling of Halomonas HL-48 cells (n = 3) was conducted as described for fluorescence gel imaging. In addition, three replicate experiments of competition labeling with CNB₁₂ (50× probe concentration) were performed, as were three no-probe control experiments (addition of DMSO only), both to further confirm the identification of B₁₂-binding proteins (SI Materials and Methods for additional details). Probe-labeled whole cells were immediately washed with PBS, fractionated, lysed, and azido-biotin was appended via click chemistry for subsequent enrichment of probe-labeled proteins on streptavidin agarose resin. For chemoproteomic and global proteomic analyses, proteins were digested by trypsin digestion for LC-MS analysis using our recently described methods (22, 34–36).

LC-MS Proteomic Measurement and Quantitative Characterization of Halomonas Proteins Captured by the B₁₂-ABP.

Tryptic peptides from enriched proteins were separated on in-house manufactured reverse-phase resin columns by LC and analyzed on a Thermo Fisher Velos Orbitrap MS. Data were acquired for 100 min, beginning 65 min after sample injection into the LC. Spectra were collected from 400 to 2,000 m/z at a resolution of 100k, followed by data-dependent ion-trap generation of MS/MS spectra of the six most-abundant ions using collision energy of 35%. A dynamic exclusion time of 30 s was used to discriminate against previously analyzed ions. For full details of the data analysis, SI Materials and Methods.

Gene Cloning and Purification of Recombinant PhrR, FolD, MetE, and MetH Proteins.

The phrR gene (locus tag CY41DRAFT_2639) from Halomonas sp. HL-48 was amplified by PCR from genomic DNA using two primers containing the BamHI and HindIII restriction sites, 5′-gatcatggatccATGAGCAACAAGGCGACCCACCCACCCG, and 5′-gagtcgaagcttTCAAAGAACGCCCGATTCACGAAGCAACG. The PCR product was cloned into the pSMT3 expression vector and the recombinant PhrR protein was expressed with an N-terminal His₆-Smt3-tag in Escherichia coli BL21/DE3 under the T7 promoter.

The folD gene (locus tag CY41DRAFT_0662), partial metH gene containing B₁₂ binding domain (2634bp) (locus tag CY41DRAFT_0722), and metE gene (Locus tag CY41DRAFT_1840) from Halomonas sp. HL-48 were amplified by PCR from genomic DNA using primer sets containing the BamHI and HindIII restriction sites for folD (5′-ggattaggatccATGACCGCCCAACTCATCGATGG, and 5′-ggatataagcttTTAATGGTTTTCGCGATCGTGCTGTTCGG), metH (5′-ggattaggatccTCGCTGTTCGTCAACGTCGGTGAACGC, and 5′-ggtataaagcttTTAGCTCGGGTCGTAAGACAGCACCGG), and BamHI and XhoI restriction sites for metE (5′-ggattaggatccATGACAGTTTCTCATATTCTCGGC, and 5′-ggtatactcgagTCAGGCGTAACGCGCGCGCAGTTGC). The PCR product of each gene was cloned into the pSMT3 expression vector and the recombinant FolD, MetH, and MetE proteins were expressed with an N-terminal His6-Smt3-tag in E. coli BL21/DE3 under the T7 promoter.

Recombinant proteins were purified to homogeneity using Ni²⁺-chelation chromatography. SI Materials and Methods for full details of cloning and purification.

Fluorescence Gel Analysis of B₁₂-ABP Labeling of Purified FolD, MetE, MetH, and PhrR Proteins.

To further evaluate selectivity of the purified proteins PhrR, MetE, MetH, and FolD (each at 2 μM), proteins were labeled with B₁₂-ABP (2 μM) and competed with excess native CNB₁₂ (10×, 25×, and 50× the B₁₂-ABP concentration) for 30 min at 30 °C. Following labeling, UV irradiation, click chemistry, and gel electrophoresis were performed as described for the transcobalamin-labeling experiment (Fig. S2). SI Materials and Methods for additional experimental details.

DNA Binding Assays.

The interaction of the purified recombinant PhrR protein with its cognate DNA binding sites in Halomonas sp. HL-48 was assessed using a fluorescence polarization assay. The 29-bp DNA fragment containing the predicted consensus DNA binding motif of PhrR and 33-bp DNA fragments of promoter regions of three PhrR target genes from Halomonas sp. HL-48 (Fig. 4) were synthesized (Integrated DNA Technologies). The double-stranded DNA (dsDNA) fragments were obtained by annealing synthesized complementary oligonucleotides at a 1:10 ratio of 5′-labeled with 6-carboxyfluorescein to unlabeled complementary oligonucleotides. Another 29-bp DNA fragment containing candidate site of TrpR repressor upstream of the trpR gene was used as a negative control. The labeled dsDNA fragments were incubated at 24 °C with the increasing concentrations of the purified PhrR protein (5–300 nM) in a 100-µL reaction mixture in 96-well black plates (VWR) for 20 min. The binding buffer contained 50 mM Tris buffer (pH 7.5), 0.1 M NaCl. Herring sperm DNA (1 μg) was added to the reaction mixture as a nonspecific competitor DNA to suppress nonspecific binding. The fluorescence-labeled DNA was detected with the FLA-5100 fluorescent image analyzer. The effect of adenosylcobalamin (AdoB₁₂) and cyanocobalamin (CNB₁₂) was tested by their addition to the incubation mixture. The effect of light was tested by incubation of 3 μM PhrR with 66 µM AdoB₁₂ in the dark or under white light conditions for 5 min followed by fluorescence polarization assay testing of the PhrR–DNA specific interaction. The PhrR-AdoB₁₂ mixture was illuminated by a white lamp in transilluminator (115 V, 0.37 Amp) in a 96-well transparent plate.

Genomic Reconstruction of PhrR Regulons.

Orthologs of phrR (e.g., HELO_1271 in Halomonas elongata) were identified in all 17 Halomonas species with sequenced genomes. We applied the integrative comparative genomics approach to reconstruct the PhrR regulons in the genomes of Halomonas species (as implemented in the RegPredict Web server, regpredict.lbl.gov) (37). The approach combines identification of candidate regulator binding sites with cross-genomic comparison of regulons and with functional context analysis of candidate target genes (38). The upstream regions of all phrR-containing operons from 17 Halomonas species, as well as from three other representatives of the Halomonadaceae family, were analyzed using a DNA motif recognition program (the “Discover Profile” procedure implemented in RegPredict) to identify the conserved PhrR-binding DNA motif. The identified PhrR motif is a 21-bp palindrome. After construction of a positional-weight matrix for PhrR motif, we searched for additional PhrR-binding sites in the analyzed Halomonas genomes and performed a consistency check or cross-species comparison of the predicted PhrR regulons. The most conserved regulatory sites confirmed by phylogenetic footprinting approach (Fig. S4) were added to rebuild the positional weight matrix for PhrR sites. Scores of candidate sites were calculated as the sum of positional nucleotide weights. The score threshold was defined as the lowest score observed in the training set. Further genomic scans using the improved PhrR motif matrix resulted in final reconstruction of PhrR regulons in the Halomonadaceae spp.

Orthologs of PhrR proteins from Halomonas spp. were identified by BLAST searches against the nonredundant set of sequenced bacterial genomes and were further checked for their genomic context using the IMG database. PhrR orthologs were identified in several lineages of Gammaproteobacteria. In most cases, the phrR orthologs are located in the vicinity of the phr or phrB genes encoding DNA photolyases. Phylogenetic analysis of the PhrR-like proteins detected major groups of genomes encoding closely related PhrRs and provided a basis for PhrR binding site identification. For each group, we collected training sets of upstream DNA regions from the phrR-containing loci, identified conserved PhrR binding motifs, constructed a positional weight matrix, and searched for additional PhrR-binding sites in the genomes analyzed. Cross-species comparisons of the predicted sets of potentially regulated genes allowed us to tentatively define regulon composition for each analyzed lineage. The details of the reconstructed PhrR regulon are captured and displayed in RegPrecise, a specialized database of bacterial regulons (regprecise.lbl.gov) (39), as a part of the PhrR collection.

phrR Mutant Construction.

A markerless in-frame phrR deletion mutant of Halomonas HL-48 was constructed by crossover-PCR, as described previously (40, 41). The 500-bp DNA fragments encoding sequence flanking phrR were generated by PCR using primer pairs (upstream 5′-CAAACCCGGGCGCGAGCATCAGCCCGAAAGCC-3′ and 5′-GCGACGGTGTCCATTAGTGGCGGGCGCCAGTGCCTCTTAGTTG-3′; downstream 5′-GCCACTAATGGACACCGTCGCCGCGGCGGTGCCTCTCTCTGAACCG-3′ and 5′-GTGTCCCGGGAGCAAAGGCCACGGCGTCGTGG-3′), joined by overlapping extension PCR, and then cloned into plasmid pDS3.0 SmaI site. E. coli WM3064 was used for mating with HL-48 and primers that flanked the region targeted for deletion were used to identify desired mutants (5′-ATGACAGCAGCACAGACGTCGTCGC-3′ and 5′-CTAGGTGGACAGCGCCTTGAAGGC-3′).

CFA Analysis.

Frozen Halomonas cell pellets were dried under vacuum. Next, 500 μL of 1.25 M HCl in methanol was added and the mixture incubated at 100 °C for 2 h to release free-fatty acids and convert them into fatty-acid methyl esters (FAMEs). Next, 500 μL of hexane was added to extract FAMEs, followed by addition of 500 μL water to the methanol layer to facilitate separation and extraction. After centrifuging for 5 min, the hexane layer collected and analyzed by GC-MS. For FAME data processing, peaks were matched and identified with two commercial databases, the NIST14 GC-MS library and Wiley GC-MS FAME Database. A mixture of FAME chemical standards (Sigma-Aldrich, C8-C28) was analyzed before sample analysis, and their retention time information was used to correct the retention time of FAME peaks in samples.

Measurement of Intracellular Folate and Derivatives.

The sample preparation procedure for folate extraction and analysis was performed with slight modifications as previously described (42). For full details SI Materials and Methods and Dataset S4.

SI Materials and Methods

Chemicals and Reagents for Synthesis.

Dry solvents were obtained from commercial sources or from an in-house LC-Technology Solutions, SP-1 dry solvent delivery system. l-photo-methionine, 1,1′-carbonyldi(1,2,4-triazole) (CDT), N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDCI), 1-hydroxybenzotriazole hydrate (HOBt), 4-methylmorpholine (NMM), cyanocobalamin (vitamin B₁₂), 1,1′-carbonyldiimidazole, N,N-dimethylformamide (DMF), ethyl acetate (ETOAc), methanol (MeOH), and dichloromethane (DCM) were purchased from Fisher Scientific or Sigma Aldrich.

General Methodology for Synthetic Chemistry.

All reactions requiring anhydrous conditions were carried out under an argon or nitrogen atmosphere using oven-dried glassware. A Biotage Dalton Isolera medium-pressure liquid chromatography system fitted with silica gel cartridges was used for chromatographic separations of reaction mixtures for isolation of desired compounds. Reactions flasks were oven-dried and cooled under vacuum. High-resolution mass spectra (HRMS) and low-resolution mass spectra (MS) electrospray ionization (ESI) accurate masses are reported for the molecular ion (M+1) or a suitable fragment ion. High-resolution MS were obtained on a Thermo Scientific Exactive Orbitrap mass spectrometer. Proton NMR (1H NMR) spectra were recorded with a Varian 500 MHz spectrometer. Chemical Shifts are reported in δ units, parts per million (ppm) downfield from a tetramethylsilane reference. Coupling constants are reported in Hertz (Hz). Carbon-13 NMR (¹³C NMR) spectra were recorded with a Varian NMR spectrometer (125.7 MHz). Chemical shifts are reported in δ units, parts per million relative to the center of the triplet at 77.00 ppm for deuterochloroform (CDCl₃), or 39.5 ppm for DMSO-D₆. ¹³C NMR spectra were routinely run with broadband.

Synthesis of Cyanocobalamin–CDT Complex, 2.

The synthesis schematic for the B₁₂-ABP is shown in Fig. 1. The cyanocobalamin-CDT (2) complex was synthesized as described previously (19). Cyanocobalamin (1, 1.00 g, 0.74 mmol) was dissolved in 2.0 mL of anhydrous DMSO and then CDT (0.425 g, 2.59 mmol) was added. The reaction mixture was stirred for 90 min and reaction completion was monitored by analytical HPLC. Once the starting material was consumed, the reaction mixture was slowly added to a rapidly stirring mixture of 1:1 diethyl ether:chloroform (400 mL). The solution was filtered by vacuum filtration and the precipitate washed twice with 30 mL of acetone. The red precipitate was collected and after drying under high vacuum a 65% yield was determined. Analytical HPLC of the cyanocobalamin–CDT complex indicated 90% pure, with the only impurity being cyanocobalamin. ES+ calculated for C66H₉₀CoN₁₇O₁₅P [M + H]⁺ 1450.5, found 1450.6 by ESI-MS. This product was in the next step used without further purification. HPLC conditions can be found below.

Synthesis of B₁₂-ABP.

Compound 2 (0.20 g, 0.14 mmol) was dissolved in 2.0 mL of anhydrous DMSO. The common “linker” (3), which was synthesized as previously described (22), was dissolved in 2.0 mL of DMSO and added dropwise to the cyanocobalamin–CDT complex solution. The solution was stirred for 20 h. The solution was then added slowly to a rapidly stirring mixture of 400 mL 1:1 diethyl ether:chloroform. The red precipitate was vacuum filtered using a large medium frit-filter and washed with 50 mL of acetone. The compound was then purified using semipreparative HPLC and desalted to obtain a pure red solid (0.035 g, 17, 5% isolated yield) (19); the HPLC semipreparative method is as follows: flow rate of 10 mL/min, 0–2 min: isocratic elution of 95.5 A:B; 2–10 min: linear gradient to 30:70 A:B; 10–15 min: linear gradient to 95:5 A:B; 15–18 min: isocratic flow of 95:5 A:B; 18–27 min: linear gradient to 5:95 A:B; buffer A: 0.05 M phosphoric acid, pH = 3.0 set with NH₄OH; buffer B: 9:1 acetonitrile:dd H₂O; a Waters 2489 dual wavelength detector and a Sun Fire prep C18 OBD 5 µM, 19 × 150-mm column, were used for HPLC. High-resolution ESI-TOF MS analysis of B₁₂-ABP in a solvent of 1:1 H₂O/CH₃CN observed m/z peak [M+H]²⁺ = 788.3403 (Calculated C₇₃H₁₀₀CoN₁₈O₁₆P, M = 1576.67); ¹H NMR (500 MHz, [D₆]DMSO): δ = 8.69 (s, 0.5H), 7.8 (s, 1H), 7.70 (s, 1.5H), 7.5 (s, 1H), 7.52 (s,1H), 7.34 (s, 1H), 7.32 (s, 1H), 7.20 (s, 1H), 7.15 (s, 1H), 7.09 (s, 1H), 7.02 (s, 1H), 6.95 (s, 1H), 6.82 (s, 1H), 6.78 (s, 1H), 6.75 (s, 1H), 6.52 (s, 1H), 6.43 (s, 1H), 6.40 (s, 1H), 6.24 (s, 1H), 5.90 (s, 1H), 5.76 (s, 1H), 4.70–4.60 (m, 2H), 4.27 (m, 1H), 4.20–4.05 (m, 3H), 4.00 (s, 2H), 3.93 (d, J = 10.6 Hz, 1H), 3.87 (s, 1H), 3.79 (s, 2H), 3.68 (s, 1H), 3.50–3.40 (m, 2H), 3.1 (d, J = 11.0 Hz, 1H), 3.0–2.9 (m, 3H), 2.67–2.58 (m, 2H), 2.38 (s, 3H), 2.19 (s, 3H), 2.16 (s, 3H), 2.06–1.96 (m, 3H), 1.80.-1.61 (m, 11H), 1.54–1.34 (m, 6H), 1.32 (s, 3H), 1.18 (s, 3H), 1.05(s, 6H), 0.98 (s, 3H), 0.85 (br s, 3H), 0.45 (br s 1H), 0.32 ppm (s, 3H); ¹³C NMR (125.7 MHz, [D₆]DMSO): 14.3, 15.6, 16.1, 16.8, 16.9, 19.1, 19.6, 20.30, 20.3, 22.5, 25.5, 26.0, 26.2, 27, 27.02, 28.4, 29.0, 29.02, 29.1, 29.3, 29.4, 29.5, 29.6, 30.6, 31.7, 32.1, 34.60, 35.6, 38.5, 40.7, 40.8, 42.2, 42.5, 47.0, 47.9, 50.7, 53.6, 54.6, 55.3, 59.0, 69.4, 70.7, 72.6, 73.1, 73.9, 75.2, 81.7, 84.8, 94.2, 103.5, 106.2, 111.7, 116.7, 118.1, 121.8, 127.8, 129.6, 130.0, 131.7, 132.9, 133.1, 136.5, 142.7, 155.9, 171.3, 171.5, 171.7, 172.8, 173.1, 173.7, 174.1, 174.6, 175.5, 178.6, 180.0 ppm.

B₁₂-ABP Labeling and CNB₁₂ Competition Studies of Pure Proteins for Fluorescence Gel Analysis, and HL-48 Proteins for LC-MS Proteomic Analyses.

Proteins MetE, the B₁₂-binding domain of MetH, PhrR, and FolD were recombinantly expressed and purified, then the proteins (2 μM in PBS buffer) were labeled with B₁₂-ABP (2 μM) for 30 min at 30 °C. Additionally, competition experiments were performed by carrying out protein labeling with the probe with simultaneous addition of excess CNB₁₂ (10×, 25×, and 50× concentrations versus the B₁₂-ABP). After the 30-min incubation, the samples were UV-irradiated for 10 min at 365 nm. Azido-tetramethylrhodamine fluorophore (2.65 µM) was added to probe-labeled protein solutions followed by the addition of Tris(2-carboxyethyl)phosphine (22 µM), TBTA (Tris[(1-benzyl-1H-1,2,3-traizol-4-yl)methyl])amine) in a 4:1 solution t-butanol:DMSO (44.8 µM), copper sulfate (45 µM), and proteins were separated using 10% (wt/vol) Tris-glycine SDS/PAGE gels. Fluorescence imaging was performed on a Protein Simple FluorchemQ system; the results are shown in Fig. S2.

For HL-48, proteins were labeled as described in the methods, and for CNB₁₂ competition studies simultaneous addition of excess CNB₁₂ (50× concentrations vs. the B₁₂-ABP) was performed before UV irradiation and click chemistry to azido-biotin (see concentrations for click chemistry reagents listed just above). See Materials and Methods for additional details.

Analytical Processing of LC-MS Data.

Resulting MS/MS spectra from LC-MS analyses were searched using the MSGF⁺ algorithm against protein sequenced deduced from the Halomonas HL-48 genome sequence (43). We analyzed our chemoproteomics data by tag-free quantitative accurate mass and time (AMT) tag proteomics and spectral counting, as described previously (22), with the following modifications. Identified peptides of at least six amino acids in length having MS-GF score ≤1E-10, which corresponds to an estimated false-discovery rate (FDR) < 1% at the peptide level, were used to generate an AMT tag database (these same high-confidence peptides were used for spectral counting, in which observations of one or more peptides for a given protein are tabulated). For AMT tag analyses, MS spectra were deisotoped using the software tool Decon2LS, after which mass and elution time features were identified and matched with VIPER to peptides stored in the Cyanothece 51142 AMT tag database within mass measurement accuracy and elution time accuracy cut-offs of <2 ppm and <2%, respectively. Measured arbitrary abundance for a particular peptide was determined by integrating the area under each LC-MS peak for the detected feature matching to that peptide. Matched features from each MS dataset were then filtered on an FDR of less than or equal to 1%. Relative peptide abundance measurements in replicate samples were scaled and normalized to the dataset with the least information using linear regression in DAnTE (44). Normalized peptide abundance values were then rolled up to proteins using RRollup; a minimum of five peptides was required for the Grubb's test, with a P value cut-off of 0.05. Only peptides unique in identifying a single protein were used to estimate protein abundances. Additionally, proteins represented by ≤two unique peptides were removed. To confirm a protein as probe-labeled, we required the following: (i) the average of probe-labeled spectral counts must be ≥two; (ii) a t test calculated between probe-labeled and no probe controls (DMSO addition only) was generated for both spectral counting and AMT tag data, in which we required that a t test value of ≤ 0.05 for either spectral counting or AMT tag; and (iii) the AMT tag values for B₁₂-ABP labeling must be ≥1.5× (fold-change) the CNB₁₂ competition experiment values, and a t test calculated between probe-labeled and CNB₁₂ competition samples must be ≤0.01. All AMT tag values shown in Dataset S1 are in log₂; thereby, the fold-change measurements are calculated as = 2^{(probe value – control value)}. The 41 proteins passing these criteria are shown in Dataset S1.

Gene Cloning and Purification of Recombinant PhrR Protein.

The PCR phrR product was cloned into the pSMT3 expression vector and the recombinant PhrR protein was expressed with an N-terminal His₆-Smt3-tag in Escherichia coli BL21/DE3 under the T7 promoter. Cells were grown in LB medium (50 mL), induced by addition of 0.2 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and harvested after 18 h of additional shaking at 25 °C. Harvested cells were resuspended in 20 mM Hepes buffer (pH 7) containing 100 mM NaCl, 0.03% Brij-35, 2 mM β-mercaptoethanol, and 2 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich). Cells were lysed by incubation with lysozyme (1 mg/mL) for 30 min, followed by a freeze–thaw cycle and sonication. After centrifugation, the cell pellet remaining after removal of the supernatant was washed with lysis buffer. The insoluble protein fraction was refolded by resuspension in 8 M Urea-AT buffer, containing 100 mM Tris, pH7, 1 M NaCl and 0.3% Brij-35, β-mercaptoethanol, and centrifuged.

Recombinant PhrR protein was purified to homogeneity using Ni²⁺-chelation chromatography. The protein was loaded onto Ni-nitrilotriacetic acid (NTA) agarose minicolumn (0.3 mL) from Qiagen and washed with starting buffer in 7 M to 1 M gradient of urea. After washing with 1 M Urea buffer containing 1 M NaCl and 0.3% Brij-35, bound proteins were eluted with 0.3 mL of the 1 M Urea AT buffer supplemented with 250 mM imidazole. The purified protein was electrophoresed on a 12% (wt/vol) SDS/polyacrylamide gel to monitor size and purity (>90%). The protein concentration was determined using the Quick Start Bradford protein Assay kit from Bio-Rad. The PhrR monomer with N-terminal His₆-Smt3-tag has a predicted molecular weight 46.6 kDa. The molecular mass of the purified recombinant PhrR protein after refolding was calculated by gel-filtration. The calculated size corresponds to dimer and monomer state of the protein. The monomer fraction of PhrR was collected after initial gel filtration, concentrated by ultrafiltration (0.8 mg/mL), and 1 mL of PhrR of the resulting filtrate incubated with fourfold molar excess of cyanocobalamin (B₁₂) for 30 min in dark. The complex was injected into a gel-filtration column and then the collected monomer fraction of PhrR analyzed by spectrophotometer to determine the B₁₂ binding to the protein.

Gene Cloning and Purification of Recombinant FolD, MetH, and MetE Proteins.

The folD gene (locus tag CY41DRAFT_0662), partial metH gene containing the B₁₂ binding domain (2,634 bp) (Locus tag CY41DRAFT_0722), and metE gene (Locus tag CY41DRAFT_1840) from Halomonas sp. HL-48 were amplified by PCR from genomic DNA using primer sets containing the BamHI and HindIII restriction sites for folD (5′- ggattaggatccATGACCGCCCAACTCATCGATGG, and 5′- ggatataagcttTTAATGGTTTTCGCGATCGTGCTGTTCGG) and metH (5′- ggattaggatccTCGCTGTTCGTCAACGTCGGTGAACGC, and 5′- ggtataaagcttTTAGCTCGGGTCGTAAGACAGCACCGG), and BamHI and XhoI restriction sites for metE (5′- ggattaggatccATGACAGTTTCTCATATTCTCGGC, and 5′- ggtatactcgagTCAGGCGTAACGCGCGCGCAGTTGC). The PCR product of each gene was cloned into the pSMT3 expression vector and the recombinant FolD, MetH, and MetE proteins were expressed with an N-terminal His6-Smt3-tag in E. coli BL21/DE3 under the T7 promoter. Cells were grown in Luria–Bertani (LB) medium with antibiotic Kanamycin (50 µg/mL) until it reached OD₆₀₀ at 1.0, then induced by adding 0.2 mM IPTG, and harvested after 3 h of additional shaking at 37 °C for FolD, after 15 h at 22 °C for MetH and MetE. Harvested cells were resuspended in lysis buffer containing 10 mM Hepes (pH 7), 100 mM NaCl, 0.15% Brij-35, 2 mM β-mercaptoethanol, and 2 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich). Cells were lysed by incubation with lysozyme (1 mg/mL) for 30 min, followed by a freeze–thaw cycle and sonication. After centrifugation, supernatant were collected and the cell pellet remaining after removal of the supernatant was washed with lysis buffer. Recombinant FolD, MetH, and MetE proteins were purified using Ni2+-chelation chromatography. The protein was loaded onto NTA agarose minicolumn (0.3 mL) from Qiagen, and washed with AT-buffer containing 50 mM Tris, 300 mM NaCl, 0.3% Brij-35, and 2 mM β-mercaptoethanol. Protein-bound columns were washed further with a buffer containing 1 M NaCl and 0.3% Brij-35, then were eluted with 0.3 mL of the AT buffer supplemented with 300 mM imidazole.

Measurement of Intracellular Folate and Derivatives.

All chemicals were purchased from Sigma-Aldrich. Solvents used for LC-MS analysis were of HPLC-grade. Extraction buffer was prepared by dissolving 2% sodium ascorbate (wt/vol) and 0.1% 2-mercaptoethanol (vol/vol) in 0.1 M phosphate buffer (pH 6.1). Folate conjugate was prepared as follows: Halomonas lysate (5 mL) was dialyzed by using 10K dialysis membrane and 500 mL of 50 mM phosphate buffer, pH 6.1. The buffer was changed three times and dialysis was perform at 4 °C with stirring for 60 min for each time of buffer change. The dialyzed sample was frozen and stored at −20 °C.

The sample preparation procedure for folate analysis was adopted with slight modifications, as previously described (42). A 50-mL cell culture was centrifuged at 5000 × g for 20 min at 4 °C. The supernatant was removed and 4 mL of extraction buffer was added to the centrifuge tube. The tube was then flushed with CO₂ gas for 15 s and vortexed. Thereafter the centrifuge tube was placed in a boiling water bath for 12 min for extraction. The tube was rapidly cooled on ice and centrifuged at 15,000 × g for 15 min at 4 °C. The resulting supernatant was then transferred to another centrifuge tube and 50 µL dialyzed rat serum was added to per 1 mL of the cell extract. The deconjugation was performed by flushing the centrifuge tube with carbon dioxide gas for 15 s and incubated in a thermomixer at 37 °C for 3 h. The tube was then boiled in a water bath for 5 min, cooled on ice, and centrifuged at 15,000 × g for 15 min at 20 °C. The supernatant was collected, spiked with internal standard, filled to an exact volume of 5 mL, and then analyzed by LC-MS.

LC-MS/MS analysis was performed using a custom-built capillary LC system (45) coupled with a Finnigan TSQ mass spectrometer (Thermo Fisher Scientific). A C18 microsolid phase extraction column was installed on-line before a C18 reversed-phase capillary column. The capillary column was connected to an in-house manufactured nanoESI emitter. Three-hundred angstrom C18 particles with 5-µm particle size (Phenomenex) packed in a 4 cm × 150 µm i.d. fused silica capillary (Polymicro Technologies) was used as online micro-SPE for sample concentration. The reversed-phase capillary column was made by slurry packing 5-µm Jupiter C18 bonded particles (Phenomenex) in a 35 cm × 150 µm i.d. fused silica capillary. The solvents used included 0.1% formic acid in water (mobile phase A) and 0.1% formic acid in 90% acetonitrile (mobile phase B). Before LC gradient started, 5-µL samples were loaded onto micro-SPE column at 5 µL/min for 10 min with 2% B. LC separation was performed with a binary gradient of 2–95% B in 30 min and at 95% B for 10 min. The flow rate through the column was 300 nL/min. ESI was performed in the positive ion mode with the major MS parameters as follows: spray voltage, 2.4 kV; capillary temperature, 310 °C; scan width, 1.0 m/z; scan time 20 ms. All cell-culture samples were prepared and analyzed in three biological replicates. The quantification of folates was accomplished in selected reaction monitoring mode by using the quantification ion and confirmed by using confirmation ions (Dataset S4) with internal standard calibration (R² > 0.99). Xcalibur 2.2 software with a Genesis Peak Detection Algorithm was used for data analysis.