A Cobalamin Activity-Based Probe Enables Microbial Cell Growth and Finds New Cobalamin-Protein Interactions across Domains

We demonstrate that a cobalamin chemical probe can be used to investigate in vivo roles of vitamin B₁₂ in microbial growth and regulation by supporting the growth of B₁₂ auxotrophic bacteria and archaea, enabling biological activity with three different cell macromolecules (RNA, DNA, and proteins), and facilitating functional proteomics to characterize B₁₂-protein interactions. The B₁₂-ABP is both transcriptionally and translationally able to regulate gene expression analogous to natural vitamin B₁₂. The application of the B₁₂-ABP at biologically relevant concentrations facilitates a unique way to measure B₁₂ microbial dynamics and identify new B₁₂ protein targets in bacteria and archaea. We demonstrate that the B₁₂-ABP can be used to identify in vivo protein interactions across diverse microbes, from E. coli to microbes isolated from naturally occurring phototrophic biofilms to the salt-tolerant archaea Haloferax volcanii.

ABSTRACT

Understanding the factors that regulate microbe function and microbial community assembly, function, and fitness is a grand challenge. A critical factor and an important enzyme cofactor and regulator of gene expression is cobalamin (vitamin B₁₂). Our knowledge of the roles of vitamin B₁₂ is limited, because technologies that enable in situ characterization of microbial metabolism and gene regulation with minimal impact on cell physiology are needed. To meet this need, we show that a synthetic probe mimic of B₁₂ supports the growth of B₁₂-auxotrophic bacteria and archaea. We demonstrate that a B₁₂ activity-based probe (B₁₂-ABP) is actively transported into Escherichia coli cells and converted to adenosyl-B₁₂-ABP akin to native B₁₂. Identification of the proteins that bind the B₁₂-ABP in vivo in E. coli, a Rhodobacteraceae sp. and Haloferax volcanii, demonstrate the specificity for known and novel B₁₂ protein targets. The B₁₂-ABP also regulates the B₁₂ dependent RNA riboswitch btuB and the transcription factor EutR. Our results demonstrate a new approach to gain knowledge about the role of B₁₂ in microbe functions. Our approach provides a powerful nondisruptive tool to analyze B₁₂ interactions in living cells and can be used to discover the role of B₁₂ in diverse microbial systems.

IMPORTANCE We demonstrate that a cobalamin chemical probe can be used to investigate in vivo roles of vitamin B₁₂ in microbial growth and regulation by supporting the growth of B₁₂ auxotrophic bacteria and archaea, enabling biological activity with three different cell macromolecules (RNA, DNA, and proteins), and facilitating functional proteomics to characterize B₁₂-protein interactions. The B₁₂-ABP is both transcriptionally and translationally able to regulate gene expression analogous to natural vitamin B₁₂. The application of the B₁₂-ABP at biologically relevant concentrations facilitates a unique way to measure B₁₂ microbial dynamics and identify new B₁₂ protein targets in bacteria and archaea. We demonstrate that the B₁₂-ABP can be used to identify in vivo protein interactions across diverse microbes, from E. coli to microbes isolated from naturally occurring phototrophic biofilms to the salt-tolerant archaea Haloferax volcanii.

INTRODUCTION

Microorganisms are foundational to diverse ecosystems, including host-associated (1), soil (2), aquatic (3), and radioactively or chemically contaminated sites (4). To thrive in these diverse environments, microbes must respond to environmental changes and interactions by regulating nutrient influx and efflux, metabolism, and gene expression. Despite considerable efforts, our understanding of these regulatory mechanisms is limited. This is especially true of uncultured members of microbial communities, where function can only be predicted using metagenomics (5). To delineate the regulatory strategies employed by microbes, approaches are needed that enable in situ characterization of microbial metabolism and gene regulation with minimal impact on cell physiology and independent of microbial cultivation.

Chemical probe-based approaches have made it possible to assign enzyme function in diverse biological systems, including microbes (6, 7). Activity-based protein profiling (ABPP) is a method that facilitates the measurement and identification of enzyme activity from any biological sample depending on the structure of the activity-based probe (ABP) (8, 9). ABPs resemble biological substrates that bind an enzyme-active site in an activity-dependent manner either in vitro or in vivo. ABPs contain a reporter tag, often an alkyne or azide residue, for the visualization and/or detection of labeled proteins. The application of ABPs requires a two-step labeling methodology; first, the ABP is added in vivo to the biological system to enable labeling, and in the second step, a reporter tag, such as a fluorophore or biotin, is attached after cell lysis via a biorthogonal click chemistry reaction (10). Here, we deploy an ABPP method that obviates the need for the addition of excessive amounts of probe to microbial cells but rather itself promotes microbial cell growth, enabling transcription, translation, and enzyme function while also eliciting measurable outputs in a biologically meaningful manner. This approach also reveals that a single probe can be designed to naturally interact with and regulate multiple biopolymers, DNA, RNA, and proteins.

In many microbiomes, microbial function and growth rates are controlled by the availability of essential growth factors, such as nucleosides, amino acids, or cofactors (11). Understanding the role B vitamins play in microbial community composition and function is essential, since their abundance can regulate gene expression, enzyme activity, and organism abundance (12, 13). Cobalamin (vitamin B₁₂) is a tetrapyrrole that is synthesized exclusively by bacteria and archaea (14). Here, we refer to four cobalamin family molecules, cyanocobalamin (CN-B₁₂), adenosylcobalamin (Ado-B₁₂), a B₁₂ activity-based probe (B₁₂-ABP) (15), and adenosyl-B₁₂-ABP, all of which have the lower axial ligand of 5,6-dimethylbenzimidazole (Fig. 1A to D). Vitamin B₁₂ derivatives are required as enzyme cofactors for the growth of many microbes and can regulate transcription and translation by serving as ligands of transcription factors and RNA switches (i.e., riboswitches) (16). Here, we focus on vitamin B₁₂-diazirine-based chemical probe and its demonstration as a chemical substitute for natural B₁₂ in facilitating bacterial and archaeal growth and its use in monitoring gene regulation and enzyme activity and identifying protein partners.

FIG 1.

Examples of cobalamin structures and biological targets. Cobalamin molecules used in this study are cyanocobalamin (A), cyanocobalamin activity-based probe (i.e., B₁₂-ABP) (B), adenosylcobalamin (C), and adenosylcobalamin activity-based probe (i.e., B₁₂-ABP adenosylated in vivo) (D). (E) B₁₂-ABP interactions investigated in this paper include import by BtuB, enzyme catalysis of MetH, NrdJ, and EtuBC, adenosylation by BtuR, RNA regulation of the btuB riboswitch, and transcriptional regulation of EutR. The red dots represent the B₁₂-ABP, while the green dots represent adenosine.

To enable in situ physiological understanding of the role B₁₂ plays in microbial cells, we tested if vitamin B₁₂-auxotrophic bacteria and archaea can grow and be viable using the B₁₂-ABP as the sole source of B₁₂. Proteomic analysis of B₁₂-ABP-grown prototrophic Escherichia coli and auxotrophic Rhodobacteraceae sp. strain HL-91 and Haloferax volcanii is used to identify B₁₂ targets during exponential growth. The B₁₂-ABP was also evaluated for its ability to regulate gene expression via the transcription factor EutR and translation of the btuB riboswitch similarly to B₁₂. The ability to grow microbes on biologically relevant concentrations of B₁₂-ABP enables the measurement of B₁₂-dependent transcriptional and translational regulation, protein-vitamin interactions, and cofactor-enzyme associations (Fig. 1E). Our approach fills a key gap in enabling in situ characterization of microbial metabolism and gene regulation, with minimal impact on cell physiology, and it greatly improves our understanding of the factors that regulate microbe function. In the near future, the synthesis of new probes and new probe applications will facilitate illumination of the mechanisms of microbe function and microbial community assembly, function, and fitness.

RESULTS

The B₁₂-ABP is actively transported and adenosylated, and it functions akin to native B₁₂.

Bacteria repress the translation of several vitamin B₁₂ biosynthesis or salvage pathway proteins by the binding of B₁₂ to a regulatory motif in the mRNA transcript, termed a riboswitch (17). Potential riboswitches may be fused to reporter genes, such as β-galactosidase (lacZ), to assess the internal concentrations of binding metabolites and corresponding extent of translational repression in vivo. Here, we utilize the B₁₂-binding btuB riboswitch to determine if (i) the B₁₂-ABP is entering the cell through the outer membrane B₁₂ transporter BtuB, (ii) there are differences in the affinities of the btuB riboswitch between B₁₂ and B₁₂-ABP, and (iii) the B₁₂-ABP is adenosylated in vivo.

To determine whether the B₁₂-ABP binds to an RNA riboswitch akin to Ado-B₁₂, we used a btuB reporter construct containing the full btuB riboswitch, including the complete 5′-untranslated region and coding sequence for the first 70 amino acids of btuB fused in-frame to lacZ (Table 1) (18). The binding of B₁₂-ABP or B₁₂ to the btuB riboswitch resulted in the repression of lacZ expression, from which we determined dose responsiveness and compared the binding affinities of B₁₂-ABP versus CN-B₁₂ for the btuB riboswitch. As shown in Fig. 2, no significant difference was observed in the affinities of the btuB riboswitch for Ado-B₁₂ (50% effective concentration [EC₅₀], 2.68 nM ± 0.36 nM), CN-B₁₂ (EC₅₀, 3.12 nM ± 0.91 nM), or the B₁₂-ABP (EC₅₀, 6.39 nM ± 2.01 nM). These results reveal that the B₁₂-ABP structure can interact with RNA and repress translation of the btuB riboswitch akin to natural B₁₂.

TABLE 1.

Escherichia coli, Salmonella Typhimurium, and Haloferax volcanii strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Description or sequence (5′ to 3′)a	Source or reference
Strains
E. coli
BL21	fhuA2 [lon] ompT gal (λDE3) [dcm]ΔhsdS λDE3 = λsBamHIo ΔEcoRI-B int::(lacI::PlacUV5::T7 gene1) i21 Δnin5	New England BioLabs
DH5α	huA2 Δ(argF-lacZ)U169 phoA glnV44 Φ80dlacZΔM15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17	Invitrogen
BW25113	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− rph-1 Δ(rhaD-rhaB)568 hsdR514	23
BW25113	ΔbtuB	18
JW1262-5	ΔbtuR778	23
TOP10	F− mcrA Δ(mrr-hsdRMS-mcrBC) Φ80dlacZΔM15 ΔlacX74 recA1 araD139 Δ(ara leu)7697 galU galK rpsL (Strr) endA1 nupG λ−	Invitrogen
GM2163	F− ara-14 leuB6 fhuA31 lacY1 tsx78 glnV44 galK2 galT22 mcrA dcm-6 hisG4 rfbD1 rpsL136 dam13::Tn9 xylA5 mtl-1 thi-1 mcrB1 hsdR2	New England BioLabs
S. Typhimurium
AR2680	ΔmetE cbiB	58
Rhodobacteraceae sp.	HL-91	34
Erythrobacter sp.	HL-111	34
H. volcanii
DS70	Wild-type isolate DS2 cured of plasmid pHV2	59
H26	DS70 ΔpyrE2	60
SH122	H26 ΔcobNΔcbiCΔcbiE	This study
Plasmids
pRsBs-ACbl-βgal-HC	RiboSwitch BioSensor, AdoCbl, β-galactosidase, high copy	18
pEutS-lacZ	B12 transcriptional reporter plasmid	This study
pTA131	Apr; pBluescript II containing Pfdx-pyrE2	60
pJAM3357	pTA131-based predeletion of cobN-cbiC-cbiE	This study
pJAM3358	pTA131-based deletion of cobN-cbiC-cbiE	This study
Primers
eutS-FW	gcgtaccaTTTTTTTTAACTCTCACGCTTATC	This study
eutS-RV	cgacgttgtaaaacgacTTCTTTATCCATGAGTCGC	This study
eutR-ori-FW	ggtgatgccggccacCGTAACGGAGGTTGCCGA	This study
eutR-RV	gttaaaaaaaaTGGTACGCTGGGAATATGATTATC	This study
ori-FW	GTCGTTTTACAACGTCGTGACTGGGAAAAC	This study
ori-RV	GTGGCCGGCATCACCGGC	This study
preKO_HVO_B0050-48_HindIII_F	AACAAAGCTTCTCGTCCACGTAGAGGATGC	This study
preKO_HVO_B0050-48_EcoRI_R	ACAAGAATTCAACTTCTCGAACGCCTCGTA	This study
KO_HVO_B0050-48-FW	GCCGCCTCCTTCCCC	This study
KO_HVO_B0050-48-RV	TGATAAAGTGTTATTGTCTTACTGC	This study
HVO_B0050-48external_FW	ATAATCTCCACGCGGTCGTC	This study
HVO_B0050-48external_RV	TATCTTGGCTCACTGCCATATTTTT	This study
HVO_B0050-48internal_FW	GTCTACACGAACGAGGAGAAGGTC	This study
HVO_B0050-48internal_RV	ACCTTTCCGAGGAACTGGTAGC	This study

^aStr^r, streptomycin resistance; Ap^r, ampicillin resistance. Lowercase sequence letters denote nontemplate overhangs created for Gibson Assembly, and restriction enzyme digestion sites are underlined.

FIG 2.

The affinity of the E. coli btuB riboswitch lacZ sensor for cyano-B₁₂ (circles), adenosyl-B₁₂ (diamonds), and B₁₂-ABP (squares). The data shown are the average fold repression values observed for BW25113 cells grown in the presence of the indicated B₁₂ molecule relative to the no-B₁₂ control. The standard error is from four biological replications. Kinetic parameters were determined by fitting a three-parameter equation to the data. The reporter construct is from Fowler et al. (18).

Cobalamin cannot passively cross cellular membranes (18, 19), requiring B₁₂ uptake systems, such as the btu system, which consists of the outer membrane TonB-dependent transporter BtuB in Gram-negative bacteria (19) and an inner membrane transporter of the ATP-binding cassette family BtuCDF (20). To determine whether the B₁₂-ABP is transported through BtuB and is adenosylated in vivo, the btuB-lacZ riboswitch reporter and a mutated version that cannot bind B₁₂ were tested in the E. coli BW25113 parent and an isogenic ΔbtuB mutant (18). The results showed that the addition of 1 μM CN-B₁₂ repressed lacZ translation in BW25113 but not in the ΔbtuB mutant, as expected. Strong translational repression of the btuB-lacZ reporter was also obtained with the B₁₂-ABP that was relieved by the ΔbtuB mutant (see Table S1 in the supplemental material), indicating that the B₁₂-ABP entered the cell through active BtuB transport. These results also suggested that the B₁₂-ABP was adenosylated in vivo, since the affinity of the E. coli btuB riboswitch has at least an order of magnitude less affinity for nonadenosylated B₁₂ molecules compared to Ado-B₁₂ (21).

Three nonhomologous cobalamin adenosyltransferases are known in bacteria, BtuR, EutT, and PduO. E. coli expresses a low level of BtuR constitutively but expresses EutT only in the presence of ethanolamine (22). To provide further support for in vivo B₁₂-ABP adenosylation, the btuB-lacZ reporter was transferred into a ΔbtuR mutant (JW2612-5) (Table 1) (23). In the ΔbtuR mutant, only Ado-B₁₂ significantly repressed btuB::lacZ expression, while CN-B₁₂ and the B₁₂-ABP did not (Table S2). These results support the conclusion that the B₁₂-ABP (i) is adenosylated in vivo by BtuR and (ii) binds and regulates the B₁₂-riboswitch similarly to Ado-B₁₂.

Vitamin B₁₂ and its derivatives function as ligands for several bacterial transcription factors, including CarH, PpaA/AerR, EutR, and PhrR (15, 24–26). We have previously demonstrated that the B₁₂-ABP binds PhrR but have not demonstrated that it forms a transcriptionally active B₁₂-ABP-protein-DNA complex (15). To extend this work, we examined B₁₂-ABP and its association with EutR, a transcription factor known to undergo a conformational change upon forming a complex with ethanolamine and Ado-B₁₂, which is needed for EutR to bind to the eutS promoter region (27). To determine whether B₁₂-ABP can regulate transcription via interactions with a transcription factor, a reporter was constructed that contained eutR and a transcriptional fusion of eutS with lacZ (eutS::lacZ) (Table 1 and Fig. 3A). We first determined appropriate concentrations of ethanolamine for further experimentation using LacZ reporter activity in the presence of 100 nM CN-B₁₂. The results revealed that the EC₅₀ of EutR for ethanolamine using this reporter construct was around 4.6 mM (Fig. 3B). Consequently, 20 mM ethanolamine was used for analysis of EutR B₁₂ binding affinity in combination with increasing concentrations of CN-B₁₂ or B₁₂-ABP. The affinities of EutR for CN-B₁₂ (26.4 nM ± 8.0 nM) and B₁₂-ABP (32.0 nM ± 12.2 nM) were found to be similar (Fig. 3C). These results demonstrate that B₁₂-ABP can function as an inducer of gene expression by a transcription factor (EutR) that requires the formation of a tertiary ligand-protein-DNA complex in vivo.

FIG 3.

The eutR transcription factor construct used to determine B₁₂ cofactor binding affinity for the eutS promoter fused to lacZ. (A) Cartoon representation of the eutS-lacZ reporter construct. (B) Average LacZ activity under increasing concentrations of ethanolamine using 100 nM cyano-B₁₂ (circles). (C) Average LacZ activity under increasing concentrations of either cyano-B₁₂ (circles) or B₁₂-ABP (squares) using 20 mM ethanolamine. The standard error is from six biological replications. Kinetic parameters were determined by fitting a three-parameter equation to the data.

B₁₂-ABP promotes microbial cell growth at natural rates.

Demonstrating that the B₁₂-ABP can control gene regulation led us to hypothesize that the probe may retain its function as a cofactor in vivo. To determine whether auxotrophic bacteria and archaea can use the B₁₂-ABP for their B₁₂-dependent growth, four bacterial species and a mutant of the archaeon Haloferax volcanii were grown in media containing either CN-B₁₂ or B₁₂-ABP (Table 1). All five species are known or predicted to require B₁₂ for use as a cofactor for methionine synthase (MetH; EC 2.1.1.13) (28, 29), methylmalonyl-coenzyme A (methylmalonyl-CoA) mutase (MCM; EC 5.4.99.2) (30), ribonucleotide reductase (NrdJ/RNR II; EC 1.17.4.1) (14, 29), or ethanolamine ammonia-lyase (EutBC; EC 4.3.1) (31). E. coli BL21(DE3) and Salmonella enterica serovar Typhimurium AR2680 can utilize ethanolamine as a sole nitrogen source via the Ado-B₁₂-requiring enzyme activity of EutBC (32, 33). Here, we found E. coli and S. Typhimurium to have similar exponential growth in the presence of either CN-B₁₂ or B₁₂-ABP when grown on ammonia as the primary nitrogen source (Fig. 4A and B). Growth on ethanolamine in the absence of these B₁₂ compounds was not observed. While E. coli and S. Typhimurium grew more slowly on ethanolamine using the B₁₂-ABP than CN-B₁₂, the two species reached identical optical densities in stationary phase independent of the B₁₂ molecule type (Fig. 4A and B). These results provide evidence that the B₁₂-ABP is catalytically functional and fulfills the vitamin B₁₂ requirement for the growth of microbial B₁₂ auxotrophs.

FIG 4.

Growth analysis of Escherichia coli, Salmonella Typhimurium, Rhodobacteraceae bacterium, Erythrobacteraceae bacterium, and Haloferax volcanii on CN-B₁₂ compared to B₁₂-ABP. Triangles are the no-B₁₂ control, circles are growth with 10 nM CN-B₁₂, and squares are growth with 10 nM B₁₂-ABP, unless otherwise noted. (A) E. coli in the presence of 50 nM CN-B₁₂ with 18 mM NHCl₂ and 18 mM ethanolamine (○) or only 18 mM ethanolamine (♢), 50 nM B₁₂-ABP with 18 mM NHCl₂ and 18 mM ethanolamine (◽) or only 18 mM ethanolamine (+), and no-B₁₂ control with 18 mM ethanolamine (X). (B) S. Typhimurium in the presence of 10 nM CN-B₁₂ with 18 mM NHCl₂ (○) or only 18 mM ethanolamine (♢), 10 nM B₁₂-ABP with 18 mM NHCl₂ (◽) or only 18 mM ethanolamine (+), and no-B₁₂ controls with 18 mM NHCl₂ (△) or 18 mM ethanolamine (X). (C) Rhodobacteraceae bacterium HL-91. (D) Erythrobacteraceae bacterium HL-111. (E) H. volcanii Δhvo_b0050-4. Cell density was determined by measuring the optical density at 600 nm (600 nm Abs.). The plotted values are the average from six biological replicates, and error bars are omitted for clarity. A four-parameter sigmoidal curve was fit to the data.

To demonstrate the utility of this approach, we used the environmental isolates Rhodobacter sp. strain HL-91 and Erythrobacter sp. strain HL-111 (34). These organisms lack genes for the synthesis of several B vitamins, including thiamine (B₁), biotin (B₇), and B₁₂ (35), requiring B₁₂ as a cofactor for MetH and NrdJ. To test if B₁₂-ABP fulfills the B₁₂ vitamin requirement for growth, HL-91 and HL-111 were grown in the presence of (i) no B₁₂, (ii) CN-B₁₂, or (iii) B₁₂-ABP in defined medium supplemented with B₁ and B₇ but lacking methionine and nucleotides. We observed no growth in the absence of B₁₂; however, the two B₁₂ forms supported identical exponential-growth rates (Fig. 4C and D). Likewise, an H. volcanii mutant, SH122, was generated (Table 1) that cannot synthesize B₁₂, but must salvage it for growth from the environment, for presumed use by a NrdJ-type ribonucleotide reductase. The H. volcanii SH122 growth results revealed that in the absence of B₁₂, there was no growth, while the growth rate was similar on CN-B₁₂ and B₁₂-ABP (Fig. 4E). These results demonstrate that MetH and NrdJ are catalytically active when cells are grown in the presence of the B₁₂-ABP and further validate that the B₁₂-ABP functions similarly to B₁₂.

We employed a global proteomic and metabolomic approach to compare the effect of B₁₂-ABP on the physiological state of E. coli and Rhodobacteraceae sp. HL-91 when grown on CN-B₁₂ or B₁₂-ABP. A good linear correlation with protein and metabolite abundance was observed between the B₁₂-ABP- and CN-B₁₂-grown bacteria (Fig. S1). For HL-91, only 31 proteins differed between the two data sets, resulting in an excellent correlation for the linear fit (R² = 0.96; Fig. S1B). The linear correlation of E. coli proteins was weaker due to several B₁₂-ABP proteins having a higher accurate mass tag (AMT) value (36) than CN-B₁₂ proteins (R² = 0.40; Fig. S1A). An evaluation of metabolite abundances in E. coli when grown on CN-B₁₂ compared to the B₁₂-ABP produced a good linear correlation (R² = 0.72; Fig. S1C), while the global metabolite analysis of HL-91 produced an even better linear relationship (R² = 0.88; Fig. S1D). These results suggest that the B₁₂-ABP has minimal effect on protein abundance and metabolite pools in these two species compared to CN-B₁₂.

Proteomic characterization of B₁₂-ABP targets.

After growth with CN-B₁₂ or the B₁₂-ABP, E. coli cells were exposed to UV light to excite the B₁₂-ABP-diazirine to form a covalent bond with the nearest macromolecule. To visualize the labeled proteins, cell extracts were prepared, and rhodamine-azide was attached via click chemistry to the alkyne moiety of the B₁₂-ABP-labeled proteins. The derivatized protein extract (15 μg) was separated by SDS-PAGE and visualized by fluorescence imaging. The gel results showed distinct fluorescent bands only in the extracts of cells grown in the presence of B₁₂-ABP versus the CN-B₁₂ control (Fig. S2A). Of the B₁₂-ABP-grown cells, a single dominant protein band was detected at ∼49 kDa in cells grown with ethanolamine versus ammonia as the sole nitrogen source (Fig. S2A). When cells were supplemented with an equimolar ratio of ethanolamine and ammonia, this 49-kDa band was similarly detected, along with additional protein bands (Fig. S2A). E. coli is a B₁₂ auxotroph when grown on ethanolamine as the sole nitrogen source, as ethanolamine ammonia-lyase (EutBC) is a B₁₂-dependent enzyme. The 49-kDa band separated at the expected molecular masses of EutB (49.4 kDa) and EutA (49.5 kDa) (the EutBC-reactivating factor), suggesting that one or both of these B₁₂-binding proteins are abundant when cells are grown on ethanolamine versus ammonia as the sole nitrogen source. This 49 kDa-band presumed to be EutA/B was detected during growth on ethanolamine irrespective of ammonia supplementation. Overall, these results reveal that the B₁₂-ABP can measure metabolism-induced (i.e., nitrogen source) shifts in the abundance of B₁₂-dependent enzymes in cells.

Using Rhodobacteraceae sp. HL-91 cells grown on CN-B₁₂ or the B₁₂-ABP, proteins were isolated and used in click chemistry to attach rhodamine-azide to image the proteins labeled by the B₁₂-ABP. The gel results did not show any observable protein bands specific to the B₁₂-ABP labeling compared to CN-B₁₂ (Fig. S2B). This finding may be due to insufficient signal from the low abundance of B₁₂-binding proteins for imaging. The lack of observable B₁₂-protein bands is consistent with the small quantity of B₁₂ required for cell growth and the benefit of being able to grow cells on B₁₂-ABP at low concentrations.

To identify proteins not visible by SDS-PAGE analysis, mass spectrometry was used after affinity enrichment of probe-labeled species. For proteomic analysis, two click chemistry reactions were applied for E. coli and H. volcanii protein extract, (i) a reaction that included biotin-azide and (ii) a no-biotin control reaction. The HL-91 samples only included the biotin-azide reaction, as background targets were expected to be lower. The attachment of biotin enabled the enrichment of labeled proteins on streptavidin-derivatized resin. Enriched proteins were digested into peptides by trypsin on the streptavidin resin and identified by liquid chromatography-mass spectrometry (LC-MS)-based proteomics. Application of filtering criteria, comparison to CN-B₁₂ and the no-biotin controls, resulted in the identification of 10 E. coli proteins that were significantly enriched >2-fold using the B₁₂-ABP over controls (see Data Set S1 in the supplemental material). The list of significantly enriched E. coli proteins included four proteins known to bind B₁₂ and 6 proteins not previously reported to bind B₁₂. We applied an additional filter to confirm B₁₂-ABP targets in which at least 1 unique peptide was required in the ABP data sets (i.e., not observed in controls or global protein data), which resulted in a reduction of the list of E. coli proteins down to five B₁₂-ABP binding proteins (Table 2).

TABLE 2.

Proteins that were significantly enriched by the B₁₂-ABP in Escherichia coli, Rhodobacteraceae sp. HL-91, and Haloferax volcanii

Organism	Gene	Locus tag	B12-ABP protein AMT tag valuea
E. coli	eutC	B21_02302	27.9
	eutB	B21_02303	27.6
	metH	B21_03851	25.0
	btuR	B21_01256	24.5
	nrdDb	B21_04070	21.5
Rhodobacteraceae sp. strain HL-91	pduO	Ga0058931_1472	27.1
	nrdJ2	Ga0058931_2479	26.3
	ecm	Ga0058931_2116	26.2
	ppaA	Ga0058931_2027	24.4
	metHb	Ga0058931_1643	23.6
	mut	Ga0058931_1592	22.7
	phaPb	Ga0058931_2637	21.7
	nrdJ1	Ga0058931_2752	21.6
	ccrMb	Ga0058931_2126	21.6
	Hypotheticalb	Ga0058931_2480	21.6
	cobA	Ga0058931_1645	21.0
	Hypotheticalb	Ga0058931_0532	20.6
	btuB	Ga0058931_0911	20.6
	metHa	Ga0058931_1644	20.6
	lonb	Ga0058931_3032	20.2
	bhmT	Ga0058931_2764	19.5
H. volcanii	btuF	HVO_1110	22.6
	mmcA1	HVO_0893	21.0
	dppAb	HVO_0062	20.8
	pepFb	HVO_1847	20.8
	mmcA2	HVO_1380	20.7
	psmBb	HVO_1562	19.0
	sppA1b	HVO_0881	18.6
	pduO	HVO_2395	18.2
	Hypotheticalb	HVO_1175	18.1

^aProtein AMT Tag values show the enriched abundance of each gene. Each protein had a t test value of <0.01 compared to the CN-B₁₂ values and had at least 1 unique peptide compared to the CN-B₁₂ and global data. All values shown are the average values for 4 biological replicates. The accurate mass tag (AMT) value represents the average log₂ abundance at the protein level, with each unique peptide having a normalized elution time and monoisotopic mass identification at a false-discovery rate (FDR) of <1.0%.

^bNot previously characterized as B₁₂ binding protein.

There were five enriched B₁₂-ABP proteins (BtuR, EutBC, MetH, and NrdD) that actively bound the B₁₂-ABP in E. coli in vivo. The two most abundant B₁₂-ABP binding proteins that were detected using B₁₂-ABP were the large (EutB, 49.4 kDa) and small (EutC, 31.7 kDa) subunits of ethanolamine ammonia-lyase, a central metabolic enzyme (Table 2). Surprisingly, the E. coli transcriptional regulator EutR, cobalamin adenosyltransferase EutT, and the ethanolamine ammonia-lyase-reactivating factor EutA were not significantly enriched using the B₁₂-ABP. We observed no peptides for EutR and EutT in the B₁₂-ABP data and only 1 peptide for each protein in the global CN-B₁₂ data, while EutA did not pass the filtering criteria (Data Set S1). These results suggest a possible technical issue with mass spectrometry detection or click chemistry efficiency in combination with possible low protein abundance. This central point is underscored by a lack of enrichment of any components of the B₁₂ transporter system, despite genetic evidence that the B₁₂-ABP is imported using this system (Table S1), while the likely more-abundant MetH was significantly enriched (Table 2). Interestingly, the anaerobic class III ribonucleotide reductase NrdD was enriched using the B₁₂-ABP, even though these cells were grown aerobically. All the E. coli proteins listed in Table 2, except for NrdD, have greater protein coverage (i.e., more peptides identified) in the B₁₂-ABP data set than in the controls and global proteomics (Data Set S1). These E. coli enrichment results demonstrate that the B₁₂-ABP is stable within cells and can selectively enrich B₁₂ targets when the probe is added as an initial component of the medium.

To demonstrate the applicability for identifying proteins that bind B₁₂ in a nonmodel bacterial species, the B₁₂-ABP was applied to Rhodobacteraceae sp. HL-91, an environmental isolate with a sequenced genome. To determine which proteins were significantly enriched by the B₁₂-ABP, a Student t test comparing enriched protein AMT values to CN-B₁₂ was used to determine a significant fold change. A total of 52 proteins had significant 2-fold enrichment due to the B₁₂-ABP (Data Set S2). Using the additional filtering criteria of requiring ≥2 unique peptides in the B₁₂-ABP data compared to CN-B₁₂ and both global data sets further reduced the data to 16 proteins (Table 2). Of the 16 proteins, 5 proteins did not have a conserved B₁₂-binding domain based on a homology search (Table S3). The most abundant B₁₂-ABP target in the HL-91-enriched data set was the cobalamin adenosyltransferase PduO, with an average AMT value of 27.1 (Table 2). The two most abundant catabolic enzymes were NrdJ2 and Ecm, with AMT values of 26.3 and 26.2, respectively. The antirepressor transcription factor PpaA was the fourth most abundant protein in the enriched B₁₂-ABP data set, with an AMT value of 24.4. The B₁₂-ABP was also able to enrich a protein of the btu system (BtuB) (Table 2). Three proteins (BhmT, CcrM, and Ga0058931_2480) that were enriched with the B₁₂-ABP did not have any identified peptides in the CN-B₁₂ data set (Data Set S2). Of these three proteins, only BhmT has a known B₁₂-binding domain (Table S3). Additionally, the three BhmT-identified peptides were only seen using the B₁₂-ABP, while CcrM and Ga0058931_2480 had more or identical peptide coverage in the global proteomics data set, respectively (Data Set S2). Interestingly, Ga0058931_2480 is in the same operon as NrdJ2 (not shown), suggesting that this protein may be interacting with the B₁₂-dependent ribonucleotide reductase, since this hypothetical protein does not have a conserved B₁₂-binding domain. B₁₂-ABP-mediated enrichment of 11 B₁₂-dependent enzymes in HL-91 demonstrates that the B₁₂-ABP binds diverse B₁₂ targets under physiologically relevant conditions.

To further expand the use of the B₁₂-ABP to environmental microbes, the B₁₂-ABP was applied to a B₁₂ auxotroph strain of H. volcanii. Using the same filtering methods, a total of 11 proteins had a significant 2-fold enrichment due to the B₁₂-ABP (Data Set S3). Using the additional filtering criteria of requiring at least 1 unique peptide in the B₁₂-ABP data compared to CN-B₁₂ and both global data sets further reduced the data to nine proteins (Table 1). Of the nine proteins, four are known B₁₂-binding proteins, while the other five proteins have not been previously described to bind B₁₂. The most abundant B₁₂-ABP target in the H. volcanii-enriched data set was the membrane-bound receptor for B₁₂ binding and uptake BtuF, with an average AMT value of 22.6 (Table 2). The B₁₂-ABP enriched subunits of methylmalonyl-CoA mutase (MutA) and cobalamin adenosyltransferase (PduO) (Table 2). The catabolic enzyme NrdJ2 did not pass the filtering criteria due to a fold change of 1.43 compared to the CN-B₁₂ with biotin, but it did have a significant t test (P < 0.0005) in a comparison of AMT tag values (Data Set S3) and had increased protein coverage due to enrichment (Data Set S3). Likewise, all the B₁₂-ABP-enriched H. volcanii proteins listed in Table 2 had the most protein coverage in the B₁₂-ABP data set (Data Set S3). These results provide a starting point for further investigation into novel/uncharacterized B₁₂ regulatory proteins and show that the B₁₂-ABP is functional across kingdoms.

DISCUSSION

The application of chemical probes has traditionally occurred after cell growth in cell extracts, using nonphysiological concentrations and conditions, including the use of probes that can irreversibly destroy enzyme activity. In contrast, using chemical probes that mimic natural cellular metabolites (e.g., amino acids, nucleotides, sugars, or vitamins) and retain biological activity facilitates measurements during cell growth which best approximate native physiology. Only a few photoreactive probes based on metabolites (e.g., S-adenosyl homocysteine [37], B vitamins [15, 38], and amino acids [39, 40]) have been described in which a photoaffinity moiety enables covalent binding to molecules interacting with the probe. These probes have only been shown to interact with a single type of macromolecule, primarily proteins. Here, we demonstrate that a chemical probe can have biological activity with three different types of macromolecules (RNA, DNA, and protein) and that it facilitates live-cell studies of biological function and native physiology with minimal impact on cell physiology.

Vitamin B₁₂ is an essential cofactor for many prokaryotes (41), being involved in pathogen virulence (42), biofilm formation (43), and many other processes (44). Growth of microbes on a biochemically functional B₁₂ chemical probe, using biologically relevant concentrations, will improve our understating of how B₁₂ is used in microbial communities. The similar growth rates observed on CN-B₁₂ or the B₁₂-ABP for all species tested (Fig. 4) and the similar protein and metabolite abundances (Fig. S1 and Data Sets S1 to S3) suggest that the B₁₂-ABP has minimal effect on B₁₂-dependent reactions and cell growth. The slightly lower growth rate observed for E. coli and S. Typhimurium when grown on the B₁₂-ABP and ethanolamine possibly stems from steric hindrance of the larger B₁₂-ABP in either the ethanolamine ammonia-lyase reaction, the exchange of adenosyl-B₁₂-ABP by the reactivating factor EutA, or possibly the entry of adenosyl-B₁₂-ABP into the microcompartment. With all the essential enzymes for ethanolamine utilization residing in one multiprotein microcompartment (45), the selectivity of the B₁₂-ABP is demonstrated by the fact that only the B₁₂-dependent enzyme EutBC was significantly enriched with the B₁₂-ABP (Table 2). Taken together, our results show that the B₁₂-ABP is transported into the cell and adenosylated, functioning analogously to B₁₂. This suggests a new powerful method for identifying and monitoring B₁₂ interactions in vivo, a concept that could likely be extended to other cofactors.

The in vivo conversion of chemical probes is an important aspect of probe functionality and has been demonstrated before with the biotransformation of pantothenamide precursors into coenzyme A (46), conversion of methionine into S-adenosyl-l-methionine to monitor protein methylation (47), incorporation of dipeptide probes into peptidoglycans (48), and incorporation of sugar probes to study protein and lipid glycosylation (49, 50). This work shows that the B₁₂-ABP is being adenosylated in vivo by BtuR in E. coli, and PduO in HL-91 and H. volcanii, suggesting that the btuB riboswitch and cobalamin adenosyltransferases can tolerate variations in the lower nucleotide structure, such as the addition of a linker (Fig. 2 and Table 2). Since the B₁₂-ABP binds B₁₂-dependent riboswitches, a potential future application of the B₁₂-ABP would be to enrich corrinoid riboswitches to confirm predicted gene regulation, which is estimated at nearly 4,000 genes in the sequenced human gut microbiome alone (51). Understanding metabolite gene regulation has great utility, since these cofactors are predicted to mediate important aspects of microbial community dynamics. Although we have previously demonstrated that the B₁₂-ABP binds MetH, EutBC, and NrdJ in Halomonas spp. (15), here, we provide evidence that these enzymes are catalytically active when cells are labeled with the B₁₂-ABP by demonstrating that the photoaffinity probe promotes growth and thus can measure B₁₂ functions in vivo. A B₁₂-ABP that behaves analogously to microbially produced B₁₂ is an important tool to gain knowledge about the role B₁₂ plays in microbial physiology and microbial communities.

We demonstrate that B₁₂-auxotrophic bacteria and archaea can grow on a B₁₂ chemical probe at biologically relevant concentrations, demonstrate that the probe has little if any impact on cell physiology, and identify B₁₂-binding proteins. This work demonstrates that the B₁₂-ABP has similar interactions with proteins, RNA, and transcription factor-DNA targets analogous to cobalamin, demonstrating that a photoaffinity probe can have biological activity with multiple types of cellular macromolecules.

MATERIALS AND METHODS

Cell growth analysis.

Growth was monitored turbidometrically by measuring optical density at 600 nm intermittently with a Bioscreen C system (Growth Curves USA). Each experiment was repeated at least three times. Replicates were used for each condition.

E. coli or S. Typhimurium was initiated from a single colony grown on LB agar into M9 medium (47.7 mM Na₂HPO₄·7H₂O, 29.2 mM KH₂PO₄, 8.5 mM NaCl, 2 mM MgSO₄, 100 μM CaCl₂, and either 18.7 mM NH₄Cl or 18.7 mM C₂H₇NO˙HCl) supplemented with 22.2 mM d-glucose (Sigma) at 37°C. Overnight-grown cells were diluted down to a starting calculated assay optical density at 600 nm (OD₆₀₀) of 0.05 in M9 medium with 22.2 mM glucose and, unless otherwise noted, 10 nM B₁₂-ABP, CNB₁₂, or AdnB₁₂ (dimethyl sulfoxide [DMSO], <0.5%) was used for cell growth. Cells were used for analysis when the OD₆₀₀ was ∼0.6.

Isolate HL-91 or HL-111 was initiated from a single colony grown on HLH agar plates (34) [10 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES) (pH 8.0), 400 mM MgSO₄, 80 mM Na₂SO₄, 20 mM KCl, 1 mM NaHCO₃, 5 mM NHCl, supplemented with 5 mM sucrose] into HLH medium. HL-91 or HL-111 was grown in the presence of either 10 nM CNB₁₂ or B₁₂-ABP. Overnight starter cultures were used to initial a larger culture used for mass spectrometry analysis. Cells were used for analysis when the OD₆₀₀ was ∼1.0.

Haloferax volcanii was initially grown from a single colony in 5 ml of complex ATCC 974 medium (52) supplemented with 446 nM uracil. The cells were washed two times with glycerol minimal medium (GMM) (53) and subcultured in GMM supplemented with 10 nM CN-B₁₂ or B₁₂-ABP. Once the cells reached log phase (OD₆₀₀, ∼0.7), the cell culture was harvested by centrifugation (8,600 × g, 2 min, at room temperature) for proteomic analysis.

Generation of an E. coli eutS-lacZ reporter strain.

All primers used to construct the EutR-dependent reporter plasmid that carries eutS-lacZ were synthesized by Integrated DNA Technologies, Inc. (Table 1). All parts used in the DNA assembly reactions for this reporter were first amplified using Phusion high-fidelity DNA polymerase (NEB) from genomic E. coli BW25113 DNA. The reactions were prepared per the manufacturer's recommendations using annealing temperatures recommended by the NEB Tm calculator. The primers used added the necessary adapter sequences to PCR fragments for Gibson assembly. The fragments were purified using the PureLink PCR micro kit, as per manufacturer's protocol (Invitrogen). The eutS-lacZ plasmid was assembled using the Gibson assembly kit, as per the manufacturer's protocols (NEB), and transformed into DH5α.

Generation of an H. volcanii B₁₂ auxotroph.

The H. volcanii B₁₂ auxotroph SH122 was generated by a markerless deletion strategy using a pyrE2-based pop-in/pop-out method (54, 55). The operon targeted for deletion from the H. volcanii H26 genome encoded homologs of aerobic cobaltochelatase subunit CobN (EC 6.6.1.2; HVO_B0050), cobalt-precorrin-8x methylmutase CbiC (EC 5.4.1.2, HVO_B0049), and cobalt-precorrin-6y C₅-methyltransferase CbiE (EC 2.1.1.-, HVO_B0048). Colonies were screened for deletion of the cobN-cbiC-cbiE operon by PCR and further assessed for auxotrophy by growth on glycerol minimal medium supplemented or not with vitamin B₁₂. For further details, see Table 1.

β-Galactosidase assay.

E. coli cells (100 μl) were added to 900 μl of Z-buffer (113 mM Na₂HPO₄, 45 mM NaH₂PO₄, 10 mM KCl, 2 mM MgSO₄, 50 mM β-mercaptoethanol, 0.02% SDS, and 0.04% chloroform) and vortexed. β-Galactosidase activity was measured by adding 10 mM ortho-nitrophenyl-β-galactoside (ONPG; Carbosynth) and stopped with 1 M NaHCO₃ after a yellow color developed (56). The results are expressed as the average β-galactosidase activity (change in ONPG absorbance/time × cell density). Experiments were run in triplicate, using the average value across six biological replicates. Cells were not UV cross-linked prior to use in assays using the B₁₂-ABP.

Click chemistry and protein enrichment.

After overnight growth in the presence of the B₁₂-ABP or CN-B₁₂, cells were placed in a glass petri dish on ice and UV-irradiated at 365 nm (UVP Bench XX-Series) for 10 min. Cells were spun down at 6,000 × g, resuspended in phosphate-buffered saline (PBS; pH 7.4) in the presence of EDTA-free protease inhibitor (Roche), and lysed by sonication (Fisher Scientific) using a total of 24 × 2-s pulses at 60% amplitude, followed by a 16,000 × g spin. Protein concentration was measured using a bicinchoninic acid (BCA) assay (Thermo Fisher). Protein was normalized to 3 mg · ml⁻¹ and used for click chemistry with 3 μM azido-tetramethylrhodamine fluorophore (Lumiprobe) or 30 μM N-(3-azidopropyl)biotinamide (TCI America), with both in the presence of 4 mM copper sulfate (Sigma), 2 mM tris(3-hydroxypropyltriazolylmethyl)amine (THPTA; Click Chemistry Tools), and 10 mM sodium ascorbate (Sigma). After incubation for 60 min at room temperature in the dark, proteins were precipitated with 2 volumes of methanol at −20°C. The proteins were pelleted at 21,000 × g and air-dried prior to resuspension in 1.2% SDS. Pellets were briefly sonicated with 6 × 1-s pulses at 60% amplitude and heated to 95°C for 5 min. Protein concentration was measured, and 3 mg of protein was enriched on streptavidin agarose beads (Thermo Fisher) for 3 h at 37°C. Agarose beads were washed 3 times each with 6 M urea prepared in 25 mM NH₄HCO₃, Milli-Q water, PBS, and 25 mM NH₄HCO₃ (pH 8.0), in that order. Enriched proteins were digested off the beads with trypsin (Promega) at 37°C for 16 h. For further proteomic and metabolomic analysis details, see Text S1 in the supplemental material.

Quantification and statistical analysis.

Data are presented as mean ± standard error of the mean (SEM) unless otherwise indicated in the figure legends. Sample number (n) indicates the number of independent biological samples in each experiment. Sample numbers and experimental repeats are indicated in the figures and figure legends or above. Data were analyzed using a paired Student's t test employing 2 tails. The probability for significance is reported at a P value of <0.05 unless otherwise noted.

For details regarding B₁₂-ABP synthesis, see Text S1 in the supplemental material.

Data availability.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (57) partner repository with the data set identifier PXD008595 and 10.6019/PXD008595. The metabolomics data have been deposited to the Open Science Framework and are accessible at https://osf.io/arp7s/.