Characterization of the nickel-inserting cyclometallase LarC from Moorella thermoacetica and identification of a cytidinylylated reaction intermediate

Abstract

LarC catalyzes the CTP-dependent insertion of nickel ion into pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN), the final biosynthetic step for generating the nickel-pincer nucleotide (NPN) enzyme cofactor. In this study, we characterized a LarC homolog from Moorella thermoacetica (LarC_Mt) and characterized selected properties of the protein. We ruled out the hypothesis that enzyme inhibition by its product pyrophosphate accounts for its apparent single-turnover activity. Most notably, we identified a cytidinylylated-substrate intermediate that is formed during the reaction of LarC_Mt. Selected LarC_Mt variants with substitutions at the predicted CTP-binding site retained substantial amounts of activity, but exhibited greatly reduced levels of the CMP-P2TMN intermediate. In contrast, enhanced amounts of the CMP-P2TMN intermediate were generated when using LarC_Mt from cells grown on medium without supplemental nickel. On the basis of these results, we propose a functional role for CTP in the unprecedented nickel-insertase reaction during NPN biosynthesis.

Graphical Abstract

Graphical Abstract.

The LarC-catalyzed nickel insertion step to produce the mature NPN cofactor involves a cytidinylylated intermediate.

Introduction

The nickel-pincer nucleotide (NPN) is a recently discovered cofactor of lactate racemase that also functions in other racemase and epimerase reactions.^1,2 This complex, with nickel coordinated to pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN), is covalently tethered by one thiocarboxylic acid to a lysyl residue in some, but not all, NPN-containing proteins. The biosynthesis of NPN is best characterized in Lactobacillus plantarum (Fig. 1). The pathway initiates from nicotinic acid adenine dinucleotide (NaAD),³ with LarB catalyzing the addition of a second carboxyl group to the pyridinium ring and hydrolyzing the phosphoanhydride bond to form pyridinium-3,5-biscarboxylic acid mononucleotide (P2CMN).⁴ Two molecules of LarE sequentially catalyze ATP-dependent sacrificial sulfur insertion reactions, resulting in P2TMN.^5,6 Finally, LarC completes the synthesis of the mature NPN cofactor by generating nickel–carbon and nickel–sulfur σ bonds in a CTP-dependent reaction.⁷ Inorganic chemists use the term cyclometallation to describe metal insertion reactions that form a metallacycle in which the metal becomes coordinated to carbon and an electrophilic atom.⁸ LarC creates such a metallacycle (indicated by the thicker lines in Fig. 1) and represents the first cyclometallase identified in nature.⁷

Fig. 1.

Biosynthesis and structure of the nickel-pincer nucleotide (NPN) cofactor. LarB catalyzes both pyridinium ring C5 carboxylation of nicotinic acid adenine dinucleotide (NaAD) and the hydrolysis of the phosphoanhydride, releasing AMP, to produce pyridinium-3,5-biscarboxylic acid (P2CMN). LarE uses ATP to activate the pyridinium ring carboxyl groups of P2CMN by adenylylation, and then transfers a cysteinyl sulfur atom to this substrate to release AMP and produce dehydroalanine. Two molecules of LarE are needed to produce each molecule of pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN). LarC transfers a protein-bound nickel ion into P2TMN in a CTP-driven reaction producing the NPN cofactor. The metallacycle generated by this reaction is highlighted by bold lines.

The sequences of LarC proteins are not homologous to other proteins of known function.⁹ LarC of L. plantarum (LarC_Lp) is among the ∼8% of LarC homologs that are encoded by two open reading frames, larC1 and larC2, and separated by a programmed ribosomal frameshift (PRF).⁹ The PRF can be eliminated by gene fusion without compromising the activity of the enzyme. The N-terminal sequence (LarC1) contains a His-rich region that is presumed to bind nickel, and ∼90% of LarC_Lp as purified from nickel-enriched growth medium is loaded with this metal ion.⁷ The full-length protein undergoes apparent proteolysis during crystallization, with only the C-terminal portion (LarC2) being crystallizable. The crystal structure of this protein fragment, a hexamer containing two domains, was solved at a resolution of 2.0 Å [protein database (PDB) ID: 6BWO].⁷ A full-length LarC protein structure has not yet been reported.

The reaction mechanism of this enzyme remains unclear, but several intriguing aspects of catalysis have been uncovered using LarC_Lp.⁷ Nickel incorporation into P2TMN requires the hydrolysis of CTP, forming CMP and presumably pyrophosphate (PP_i). Also required is the presence of Mg²⁺ or Mn²⁺, with a preference for the latter metal. Of great interest, the structure of LarC2∙Mn∙CTP was solved by soaking the protein fragment with this metal ion and nucleotide (PDB ID: 6BWQ).⁷ Surprisingly, LarC_Lp appears to be a single-turnover enzyme, with a single molecule of CTP undergoing hydrolysis for each molecule of NPN synthesized. Site-directed mutagenesis of the fused version of larC was used to replace several C-terminal domain residues involved in CTP binding, generally resulting in severe diminishment of LarC activity. Mutagenesis of larC also was used to delete the His-rich region or to substitute several acidic residues in the LarC1 region, demonstrating the importance of these components for the enzyme activity. The combined results led to a proposal that LarC utilizes a carboxylate-associated mechanism for transferring nickel into P2TMN in a CTP-dependent manner.⁷

In this study, we characterized a LarC homolog from Moorella thermoacetica (LarC_Mt) encoded by a gene lacking an internal stop signal and not subject to a PRF. We demonstrated that LarC_Mt is more resistant to proteolysis when compared to LarC_Lp, and we characterized selected properties of the protein. For example, we ruled out the hypothesis that enzyme inhibition by its product PP_i accounts for its apparent single-turnover activity. Of greatest interest, we identified a cytidinylylated (CMPylated)-substrate intermediate that is formed during the reaction of LarC_Mt. Selected variants with substitutions at the predicted CTP-binding site retained substantial activity, but exhibited greatly reduced levels of the intermediate. In contrast, use of LarC_Mt from cells grown on medium without supplemental nickel led to enhanced amounts of the intermediate. On the basis of these results, we propose a functional role for CTP in the unprecedented nickel-insertase reaction during NPN biosynthesis.

Methods

Materials

Carbenicillin, kanamycin, chloramphenicol, and β-D-1-thiogalactopyranoside were purchased from Gold Bio (St. Louis, MO, USA). Desthiobiotin and NaAD were acquired from Sigma (St. Louis, MO, USA). All other chemicals used were reagent grade or better.

Genes, plasmids, and cloning

The gene encoding LarC_Mt, flanked by NdeI and XhoI restriction sites, was chemically synthesized (Integrated DNA Technologies, Coralville, IA, USA). The DNA fragment was inserted into the vector pLW01¹⁰ resulting in the production of LarC_Mt with an N-terminal His₆-tag followed by a tobacco etch virus protease cleavage site. Site-directed mutagenesis of larC_Mt was carried out using the gap-repair method,¹¹ and the constructs were verified by Sanger sequencing (Azenta, South Plainfield, NJ, USA). The constructs were transformed into competent Escherichia coli BL21 (DE3) cells for gene expression and protein purification studies.¹² The strains, plasmids, and primers used in this study are provided in Table S1.

Gene overexpression and protein purification

We grew E. coli BL21 (DE3) strains containing plasmids with wild-type and mutant larC_Mt in an autoinduction medium¹³ amended with 100 mg/L carbenicillin. The cultures were grown at 20°C while agitating at 220 RPM and, except where indicated, 1 mM NiCl₂ was added after 4 h of growth. Cells were harvested after ∼20 h by centrifugation at 8000 rpm, resuspended in an equal volume of 100 mM Tris, pH 8.0, buffer containing 300 mM NaCl, and stored at −80°C until needed.

Thawed cells were lysed by use of a French pressure apparatus operating at 16 000 psi and 4°C. The debris was removed by centrifugation (45 min at 115 955 ×g) at 4°C. His-tagged LarC_Mt and its variants were purified using a His60 Ni Superflow resin by following the manufacturer's protocol (Takara Bio, San Jose, CA, USA). For native molecular weight determination, the sample was subjected to size exclusion chromatography (SEC) in 100 mM Tris-HCl buffer, pH 8.0, containing 300 mM NaCl on a Superdex 200 Increase 10/300 GL column (GE Healthcare, Chicago, IL, USA) while monitoring with miniDAWN TREOS multi-angle light scattering (MALS) and TRex refractive index detectors (Wyatt, Santa Barbara, CA, USA). The data were analysed with the ASTRA software (Wyatt).

Overexpression and purification of two other NPN biosynthesis proteins, LarB_Lp [from E. coli BL21 (DE2) cells] and LarE_Lp (expressed in E. coli ArcticExpress cells), were carried out as previously described.^4,5 The lactate racemase apoprotein from Thermoanaerobacterium thermosaccharolyticum, LarA_Tt, was obtained from Lactococcus lactis NZ3900 cells containing pGIR082, following the previously reported purification protocol.⁹

The protein concentrations were determined either using the absorbance at 280 nm (ε₂₈₀ = 23 740 M^–1 cm^–1 using the ExPASy protein parameter tool) or the Bradford protein assay reagent (Bio-Rad, Hercules, CA, USA) using bovine serum albumin as the standard.

LarC activity assay

The LarC_Mt substrate, P2TMN, was synthesized by incubating NaAD (0.2 mM) with LarB_Lp (10 µM) and NaHCO₃ (50 mM), to generate P2CMN, along with LarE_Lp (200 µM), ATP (2 mM), and MgCl₂ (20 mM) at room temperature for 1 h in 100 mM Tris-HCl buffer, pH 7.0.⁷ Conversion of P2TMN to NPN was achieved by incubation of an aliquot of the mixture mentioned earlier with an equal volume of CTP (0.2 mM), MgCl₂ (10 mM), β-mercaptoethanol (β-ME, 10 mM), and LarC_Mt or LarC_Lp (2.5 µM) in 100 mM 2-(N-morpholino)ethanesulfonic acid, pH 6.0, at room temperature for 30 min. Synthesis was terminated by heat treatment at 95°C for 10 min. A 5-µl aliquot of the resulting NPN was mixed with LarA_Tt apoprotein (0.8 µM) and L-lactate (45 mM) in 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (100 mM), pH 7.0, for 5 min at 50°C, and the reaction was terminated by incubation in 95°C for 10 min. The lactate racemase activity was measured using a commercial kit (Neogen) as previously described.⁹ Variations to these parameters were performed in specific experiments that are described in the “Results and discussion” section. In some cases, P2TMN was purified by chromatography on a Q-Sepharose column (5 mL column volume) in 30 mM Tris-HCl buffer, pH 8, with elution using a NaCl gradient (to 1 M) and detection at 254 nm, analogous to a previously described procedure.⁷

To assess whether the apparent single-turnover reactivity of LarC_Mt is due to inhibition by PP_i, we tested the effect of adding 10 mM PP_i (Avantor, Radnor, PA, USA) and of providing 2 units of pyrophosphatase (PPase; Sigma) to the LarC assay reaction.

Mass spectrometric (MS) analysis

The intact LarC_Mt protein mass was determined using a Waters G2-XS Q-TOF (time of flight) mass spectrometer by injecting 10 µl of sample onto a Thermo Hypersil Gold CN guard column (1.0 × 10 mm) for desalting. A gradient of water + 0.1% formic acid (solvent A) and acetonitrile (solvent B) was run as follows at a flow rate of 0.1 ml/min: initial conditions were 98% A/2% B, hold at 2% B until 5 min with the flow diverted to waste for the first 3 min, ramp to 75% B at 10 min and hold at 75% B until 12 min, return to 2% B at 12.01 min and hold until 15 min. Mass spectra were obtained using electrospray ionization in positive ion mode with a source temperature of 100°C, cone voltage of 35 V, desolvation temperature of 350°C, desolvation gas flow of 600 L/h, cone gas flow of 50 L/h, and capillary voltage of 3.0 kV. Data were acquired using a 0.5 s TOF MS scan across an m/z range of 200–2000 and the spectra were deconvoluted in Masslynx using the Max Ent I algorithm.

LarC_Mt reaction samples were analysed using a Waters G2-XS Q-TOF mass spectrometer interfaced with a Waters Acquity UPLC. The 10-µl samples were injected onto a Waters Acquity UPLC BEH-C18 column (2.1 × 100 mm) that was held at 40°C. Compounds were separated by ion-pairing chromatography using a binary gradient as follows: initial conditions were 100% mobile phase A (10 mM tributylamine and15 mM acetic acid in a 97:3 water/methanol (v/v) mixture) and 0% mobile phase B (methanol), hold at 100% A for 1 min, linear ramp to 99% B at 7 min, hold at 99% B to 8 min, return to 100% A at 8.01 min and hold until 10 min. The flow rate was 0.3 mL/min. Mass spectra were obtained by electrospray ionization operating in negative ion mode with a capillary voltage of 2.0 kV, source temperature of 100°C, cone voltage at 35 V, desolvation temperature of 350°C, desolvation gas flow of 600 L/h and cone gas flow of 50 L/h. Data were acquired using a data-independent MS^e method (scans with fast switching between no collision energy and using a collision energy ramp of 20–80 V) across an m/z range of 50–1500. Daughter ion spectra were acquired for m/z = 715.02 using an MS/MS method with selection in the quadrupole and fragmentation using a collision energy ramp of 10–60 V. Lockmass correction was performed in MassLynx software using leucine enkephalin as the reference compound.

Results and discussion

Characterization of M. thermoacetica LarC

The genome of M. thermoacetica exhibits four widely dispersed sites of lar genes, with an isolated larB, two widely separated larA homologs of undefined roles, and larE grouped with larC; this organization contrasts with the situation in L. plantarum, where a single copy of larA encoding lactate racemase is located immediately adjacent to the three NPN biosynthetic genes (Fig. S1). LarC_Mt and LarC_Lp exhibit 38% sequence identity with only the latter protein containing a PRF (Fig. S2). The endogenous His-rich region is shorter in LarC_Mt compared to LarC_Lp and it has a smaller overall abundance of His (10 versus 23 residues), especially in the N-terminus. Only five His residues are conserved between the two proteins. Homogeneity of the His₆-tagged protein was established by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. S3A) and confirmed by ESI-MS, which yielded m/z = 44 202.5 (Fig. S3B), consistent with the calculated mass of the nickel-free, full-length protein subunit (monoisotopic M_r = 44 176.82).¹⁴ SEC comparison to standards provided an approximate M_r = 321 kDa (Fig. S3C), whereas SEC-MALS analysis indicated a 6- or 7-mer oligomeric state of LarC_Mt in solution (M_r = 287 500) (Fig. S3D).

The activity of LarC_Mt was assessed by an indirect assay that measured the lactate racemase activity of NPN-activated LarA_Tt apoprotein.⁷ The LarC_Mt substrate was generated from NaAD by the combined actions of LarB and LarE in the presence of CO₂/bicarbonate and Mg∙ATP. Transformation of P2TMN into the NPN cofactor was achieved by LarC_Mt in the presence of Mn∙CTP. Incubation of the resulting NPN with LarA_Tt was followed by measurement of lactate racemase activity. The general enzymatic properties of LarC_Mt were similar to those of LarC_Lp, but the activity of the M. thermoacetica enzyme (see following text) exhibited only approximately 3% of the L. plantarum enzyme activity⁷ when assayed using the reported standard conditions for the latter enzyme.

Relationship of PP_i to the reactivity of LarC_Mt

To assess whether the enzyme reaction product PP_i leads to the low activity of LarC_Mt and to investigate whether such inhibition accounts for the previously described apparent single-turnover reactivity of the enzyme,⁷ we examined the effects of adding PP_i or introducing pyrophosphatase (PPase) to the assay (Fig. 2). PP_i was demonstrated to be a potent inhibitor of LarC_Mt activity, but full activity was restored by the introduction of 2 units of PPase. Inclusion of PPase in a sample lacking PP_i, however, did not lead to greater levels of activity. Thus, while PP_i inhibits LarC_Mt, it does not account for the apparent single-turnover reactivity of this enzyme. As previously proposed,⁷ it is more likely that the stoichiometric nickel-insertion reaction is attributed to transfer of the inaccessible metal ion within the protein that cannot be replaced by adding nickel ions to the assay solution.

Fig. 2.

Pyrophosphate (PP_i) affects the activity of LarC_Tm. The nickel-inserting activity of LarC_Mt was assayed by an indirect assay that involved the activation of LarA_Tt apoprotein by enzymatically produced NPN and subsequent measurement of the conversion of L-lactate into D-lactate. The LarC_Mt substrate, pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN), was generated by the combined action of LarB_Lp and LarE_Lp acting on NaAD. Shown are the control reaction along with the effects of including 10 mM PP_i, both PP_i and pyrophosphatase (PPase), and only PPase (n = 2 biological replicates using separate enzyme preparations).

Identification of a P2TMN-CMP reaction intermediate

Time-dependent MS analysis of metabolites associated with the LarC_Mt reaction revealed the expected decrease in P2TMN levels as NPN was synthesized (demonstrated by incorporating the cofactor into LarA_Tt and measuring the lactase racemase activity, Fig. 3A). The concentrations of synthesized P2TMN are not known because a standard is not available, so only the relative abundances are indicated for the representative data shown. Although NPN was able to be detected, the very weak intensity of the feature associated with the cofactor prevented its quantification using these conditions. Notably, an unidentified species (m/z = 715.02) was shown to be generated by LarC_Mt as P2TMN was consumed (Fig. 3A and B). The relative abundance of this new metabolite was based on comparison of its peak intensities to that of P2TMN at zero time; no metabolite was detected at the initial time point. Significantly, the novel species was not formed in the absence of CTP. The mass of this species is consistent with that of P2TMN linked to CMP. MS–MS fragmentation analysis of this species supported a structure in which the substrate forms a phosphoanhydride bond with the nucleotide (Fig. 4), i.e. a CMPylated P2TMN. In particular, a species with an m/z of 402.01 is consistent with a phosphoanhydride-containing molecule. The proposed formation and decay of this species (Fig. 5) is reminiscent of an intermediate formed by molybdenum insertase (Cnx1) during the synthesis of the molybdenum cofactor (Moco, Fig. S4). In that pathway, molybdopterin is thought to be adenylylated to properly position its dithiolene moiety near the molybdenum-binding site, followed by molybdate insertion and phosphoanhydride hydrolysis to release AMP and the Moco.^15–18 We speculate that a similar process occurs during NPN cofactor biosynthesis, i.e. P2TMN undergoes CMPylation to assist in orienting the pyridinium ring near the buried nickel-binding site followed by metal insertion and phosphoanhydride cleavage; however, additional studies (such as structure determination of the LarC_Mt∙CMP-P2TMN complex) are required to verify this hypothesis. Of additional interest, the Moco-forming enzyme is inhibited by PP_i,¹⁵ as shown above for LarC_Mt.

Fig. 3.

LarC_Mt forms a reaction intermediate. (A) The LarC_Mt reaction time course reveals a decrease in the concentration of pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN; solid gray bars), an increase in NPN synthesis (monitored by the ability to activate LarA_Tt to generate lactate racemase activity, line), and an increase in another metabolite (striped gray bars). The relative abundance of P2TMN and the novel metabolite are based on the intensities of the mass spectrometric (MS) peaks relative to that of P2TMN in the zero-time sample, which did not contain the other metabolite. (B) MS analysis of the intermediate species provides m/z = 715.02 and the relative intensities of the isotopic species indicate the presence of two sulfur atoms, consistent with a P2TMN-CMP linkage. These representative data are from a single experiment, but a replicate with another enzyme preparation showed the same trends.

Fig. 4.

The mass spectrometric (MS)/MS fragmentation spectrum of the LarC_Mt intermediate is consistent with cytidinylylated (CMPylated) pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN).

Fig. 5.

Two-step reaction of LarC. CTP-dependent CMPylation of pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN), with release of PP_i, is proposed to position the pyridinium ring in a proper orientation to allow nickel ion transfer accompanied by phosphoanhydride hydrolysis by LarC_Mt.

LarC_Mt variant activities

The D256A, E261A, E364A, and D256A/E261A variants of LarC_Mt were created and characterized for two reasons. First, the three side chains correspond to residues (Asp284, Glu289, and Glu387, Fig. S2) that coordinate the manganese within the LarC_Lp CTP-binding domain (PDB ID: 6BWQ).⁷ D256A and E364A variants of LarC_Mt were created to confirm their importance for activity, based on the significant effects of D284A and D387A variants of LarC_Lp (∼10% active and inactive, respectively).⁷ No variant of Glu289 was examined for LarC_Lp because that position is not conserved (e.g. Gln was noted in some LarC sequences). Second, these residues were targets for mutagenesis because their predicted positions in LarC_Mt (Fig. S5) were appropriate for facilitating hydrolysis of the intermediate. Thus, we wondered whether their substitution by alanine might increase the production of CMPylated-P2TMN.

The activities of the LarC_Mt variants were compared to that of the wild-type enzyme, and to two of the corresponding LarC_Lp variants, by using the indirect lactate racemase-based assay (Fig. 6A). We found the D256A variant of LarC_Mt exhibited about 50% of the wild-type enzyme activity, compared to the 90% activity loss for the D284A variant of LarC_Lp. The E261A variant of LarC_Mt retained even greater levels of activity. Surprisingly, the E364A variant of LarC_Mt exhibited near wild-type activity levels, whereas the corresponding E387A variant of LarC_Lp was inactive. The basis of this difference is unclear, but may relate to protein folding issues, functional redundancy by another residue in close proximity, or other effects. The E261A variant of LarC_Mt was slightly more active than the D256A variant. The reductions in activity by the D256A and E261A variants were approximately additive for the D256A/E261A double variant. MS analysis of the metabolites associated with these reactions revealed substantial reduction in the amount of P2TMN for the wild-type enzyme, clear decreases of P2TMN for the more active variant enzymes, and less utilization of the substrate by the double variant (Fig. 6B), as expected. Notably, all of the variant proteins showed insignificant levels of the intermediate m/z = 715.02 species (Fig. 6C). These results suggest that the decreases in variant enzyme activities are primarily associated with reduced rates of synthesis of the CMP-P2TMN intermediate while not affecting the hydrolysis of this species.

Fig. 6.

LarC_Mt variant analysis and effect of nickel limitation. Relative levels of (A) lactate racemase activity, (B) remaining pyridinium-3,5-bisthiocarboxylic acid mononucleotide (P2TMN) substrate, and (C) CMP-P2TMN intermediate after 30 or 60 min of incubation of LarC_Mt samples with P2TMN and Mn∙CTP. Samples included wild-type LarC_Mt and its D256A, E261A, E364A, and D256A/E261A variants that were purified from cultures supplemented with 1 mM NiCl₂ as well as wild-type LarC_Mt that was isolated from cells grown in medium without supplemental nickel ions. The samples were incubated with enzymatically produced P2TMN for the times indicated, the relative abundance levels of P2TMN and CMP-P2TMN were quantified by comparing the intensities of these mass spectrometric (MS) peaks to that of P2TMN at zero time (n = 2, technical replicates), and the products were mixed with LarA_Tt apoprotein and the resulting lactate racemase activities were determined (n = 1).

Also shown in Fig. 6 are the activity and metabolite level results obtained using LarC_Mt that was purified from cells grown in medium without supplemental nickel addition. This form of the enzyme was active, demonstrating the ability of the enzyme to sequester trace levels of nickel ions from the medium during growth. Accordingly, the relative levels of P2TMN exhibited substantial decreases over time. Significantly, the relative level of the CMP-P2TMN intermediate was greater than that associated with enzyme purified from cells grown with excess nickel ions. These results suggest, but still require further verification, that the intermediate is generated prior to nickel insertion, and that limited nickel levels increases the amount of the intermediate.

Conclusions

The gene encoding LarC_Mt was expressed, the protein was purified, and several of its properties were determined. The enzyme converts P2TMN to NPN, but is inhibited by the product of the reaction, PP_i. This PP_i inhibition does not account for the apparent single-turnover reaction kinetics of the enzyme. We identified a novel intermediate in which the precursor, P2TMN is CMPylated. Substitution of residues that are predicted to be positioned at the Mn∙CTP-binding site resulted in only partial reduction of LarC_Mt activity, but a significant reduction in the formation of CMP-P2TMN, indicating the rate-determining step of NPN synthesis is associated with formation of the intermediate. By contrast, enhanced levels of CMP-P2TMN are produced when nickel ions are limiting, consistent with the metal ion binding to the CMPylated intermediate. Our discovery of the CMP-P2TMN reaction intermediate provides insight into the mechanism of the nickel insertion reaction and clarifies the role of CTP in this reaction.

Funding

Funding in support of this work was provided by the National Science Foundation (CHE1807073) and the National Institutes of Health (GM128959) to R.P.H. and J.H., and from the National Science Foundation Graduate Research Fellowship Program (DGE1848739) to A.T.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

The data underlying this article will be shared on reasonable basis by submitting a request to the corresponding author.