Abstract
It has been challenging to test how proteins acquire specific metals in cells. The speciation of metalation is thought to depend on the preferences of proteins for different metals competing at intracellular metal-availabilities. This implies mis-metalation may occur if proteins become mis-matched to metal-availabilities in heterologous cells. Here we use a cyanobacterial MnII-cupin (MncA) as a metal trap, to test predictions of metalation. By re-folding MncA in buffered competing metals, metal-preferences are determined. Relating metal-preferences to metal-availabilities estimated using cellular metal sensors, predicts mis-metalation of MncA with FeII in E. coli. After expression in E. coli, predominantly FeII-bound MncA is isolated experimentally. It is predicted that in metal-supplemented viable cells metal-MncA speciation should switch. MnII-, CoII-, or NiII-MncA are recovered from the respective metal-supplemented cells. Differences between observed and predicted metal-MncA speciation are used to refine estimated metal availabilities. Values are provided as blueprints to guide engineering biological protein metalation.
Similar content being viewed by others
Introduction
A purpose of this research is to test explanations of how proteins acquire different metals in cells: the speciation of metalation in biology (Supplementary Fig. 1). Metalloenzyme catalysis is mostly metal-specific. Yet metalloproteins typically bind one or more wrong metals in preference to the cognate metal(s)1,2,3. Nascent metal sites in proteins are flexible such that non-cognate metals can bind non-conservatively by using a subset of the native ligands, by recruiting additional ligands and/or by adopting non-cognate coordination geometries4. With such limited constraint the order of metal binding commonly follows the Irving-Williams series (Fig. 1a)2,5. An exception is where there has been prior structural organisation for example via cooperativity in di-metal sites6,7. Here we quantify the binding preferences of a protein (MncA) which kinetically traps metals during folding8. We then use MncA to establish if protein metalation can be correctly predicted, then predictably adjusted, and finally to refine estimates of intracellular metal availability.
a The Irving-Williams order of formation of complexes with metals, as the ionic states of exchangeable metals in the cytosol. From weak (left) to tight, noting reversal of arrow after copper. b Protocol for recovery of unfolded MncA from inclusion bodies after high-level expression in E. coli BL21(DE3) pLysS for 3 h using pET29aMncA at 37 °C, followed by folding denatured MncA via dropwise dilution into a large volume of urea-free buffer containing pairs of buffered and competing metal ions. Folded MncA was concentrated by anion exchange and eluted with single-step high-salt buffer. Metal-bound MncA was separated from unbound metal by SEC (PD-10 column), and metal analysed by ICP-MS. Gel inset shows, from left, size markers (from top, 100, 70, 55, 35, 25, 15 kDa), overloaded urea-solubilised MncA and MncA eluted from Q-Sepharose (full gel Supplementary Fig. 2a). c Metal contents of MncA-containing fractions after folding in excesses of competing metals buffered to different ratios of availabilities (as in panel b, buffers in Supplementary Table 1 prepared with assistance from Supplementary Data 1). Ratios of competing metals above each panel. The concentration of MncA approximated from A280nm (ɛ = 120,000 M−1). Stoichiometries approximate 2:1 but proportional metalation was calculated from total metal as in Table 1. Experiments with CuI and FeII performed in an anaerobic chamber, metal stocks prepared freshly, hydroxylamine (1 mM) used to sustain reduced copper. FeII and CuI confirmed to be > 95% reduced via reaction with excess ferrozine or BCA respectively. n = 3 independent experimental replicates shown. d Quantified preferences of MncA for different divalent metals, plus monovalent copper, relative to MnII (n = 3 independent experiments, triangles, squares, and circles depict replicates in (c), (I) and (II) denote forms of copper), determined from the elution profiles in (c). The order of preferences follows the Irving-Williams series as in (a). MncA prefers all non-cognate metals over MnII. Source data are provided as a Source Data file.
In combination with the metal-binding preferences of proteins, a second factor determining metal-protein speciation is metal availability8. Metal partitioning can occur directly from pools of labile, exchangeable and hence available metals to the protein of interest, or to an assembly pathway for a small molecule cofactor (such as FeII into heme or CoII into vitamin B12), or to a metallochaperone that delivers to an assembly pathway or to the protein of interest9,10,11,12,13,14,15,16,17,18. Here we explore the initial partitioning step from labile, available intracellular metal pools.
The significance of intracellular metal availability to metal-protein speciation emerged over several decades. When expressed in a cyanobacterial cell a NiII- and CoII-responsive DNA-binding metal-sensor from Mycobacterium tuberculosis solely responded to CoII19. This change in specificity was attributed to NiII being insufficiently available inside viable cyanobacteria to metalate NmtR. Similarly, FeII-responsive DtxR from Corynebacterium diphtheriae gained responsiveness to MnII when expressed in Bacillus subtilis, again attributed to different intracellular MnII availabilities in the different cells20. Two metal-binding cupins, MnII-MncA and CuII-CucA, were discovered in the periplasm of a cyanobacterium (Synechocystis PCC 6803), notably binding metals from different ends of the Irving-Williams series yet exploiting similar folds and identical metal-binding residues8. When folded in vitro in similar amounts of competing copper versus MnII, or ZnII versus MnII, the non-cognate metals, copper or ZnII, bound to MncA8. While the copper-cupin is secreted unfolded via the Sec-system, the MnII-cupin is a Tat-substrate which folds and kinetically traps the less competitive metal in the cytosol before secretion. Thus, metal availability at the site of protein folding determines metal speciation, and MnII must be more available than CuI or ZnII in the cyanobacterial cytosol8. MncA is an MnII-dependent oxalate decarboxylase (inactive with copper or ZnII), and the related cytosolic OxdC from Bacillus subtilis trapped either MnII or CoII in E. coli depending on media metal supplementation21. MncA thus offers enticing opportunities to directly interrogate mechanisms and predictions of metalation because (1) kinetically trapped metals are unlikely to exchange during purification, and so MncA should faithfully report its in vivo metalation state, and (2) precedent suggests it could be possible to switch the speciation of MncA metalation in E. coli and discover if various metalation states are predictable.
We recently developed a metalation calculator which accounts for inter-metal competition within cells22. The calculations use estimations of intracellular metal availabilities derived from thermodynamically calibrated responses of the cells’ DNA-binding metal-sensing transcriptional regulators (Supplementary Fig. 1)23,24. Availabilities are described as free energies consistent with bound but labile metals capable of rapid associative ligand exchange with the protein of interest23,25,26. By using estimates of intracellular availabilities standardised to the mid-points of the sensor ranges (representing idealised cells) the cognate metals of four exemplar proteins have been correctly decoded22,23. These data encourage a view that our understanding of the speciation of metalation is (broadly) correct. However, unlike in idealised cells, metal sensors will be at different positions in their ranges in actual cells depending on growth conditions. The abundance of transcripts encoded by metal-sensor regulated genes was estimated by qPCR and then calibrated to estimate metal availabilities inside E. coli grown in un-supplemented medium27. To date, calculations of protein metal speciation have only been experimentally tested indirectly in E. coli engineered to manufacture vitamin B1222,27,28. CoII-dependent production of B12 was thus measured as a proxy for the metalation states of the CoII metallochaperone CobW, and the CoII chelatase CobNST. This work now tests predictions directly by using cyanobacterial MncA to read out in vivo metalation when expressed in E. coli cells under different growth conditions.
Here we determine whether metalation calculations correctly predict mis-metalation speculated to occur in engineered cells where metal availabilities are mismatched to the metal preferences of heterologously expressed proteins. In cells supplemented with cobalt, nickel and manganese, residual differences between predicted and observed MncA metalation are also used to refine estimated availabilities of other metals, and we test if changes in metal atoms cell−1 align with these refinements. We explore the mechanism by which exposure to one metal (eg cobalt) can change availability of another (eg FeII). Calculators are included to enable use of MncA to probe metal availabilities under other growth conditions and in other cell types. Efforts are being made to engineer proteins that overcome the Irving-Williams series29. These approaches place constraints on, for example, bi-metallic catalytic centres. In contrast, here we show how biology can be exploited in a predictable way to overcome the challenge presented by the Irving-Williams series, thus expanding the repertoire of metal-driven biocatalysis that can be predictably utilised. Calculators are provided to guide the optimisation of protein metalation with different metals.
Results
Metal preferences at folding and trapping can also follow the Irving-Williams series
The first objective was to measure the metal-binding preferences of MncA in vitro. It is not feasible to measure affinities because MnII is entrapped within the folded protein such that off-rates become negligible8. Instead, relative preferences during folding were determined. Rapid, high-level, expression of MncA (minus secretion signal peptide) in E. coli produces MncA-containing inclusion bodies from which unfolded apo-protein can be recovered8. MncA was thus expressed in E. coli BL21(DE3) pLysS, and urea-solubilised MncA, shown in Fig. 1b, was refolded by dilution into urea-free buffer. Refolding solutions contained pairs of competing metals buffered with NTA (or histidine for NiII competitions) as in Supplementary Table 1, formulated via Supplementary Note 1 using the provided calculator (Supplementary Data 1). Competitions involving FeII and CuI were performed in an anaerobic chamber with N2-purged buffers and metal stocks confirmed > 95% reduced immediately prior to use. Metals were unbuffered in competitions between CuI and ZnII. Refolded MncA (Fig. 1b), recovered by anion exchange chromatography, was resolved from unbound metal by size exclusion chromatography (SEC) with fractions (0.5 mL) analysed for MncA by UV absorbance and metals by ICP-MS.
The proportion of each metal acquired by MncA was determined from the chromatograms in Fig. 1c. Challenges in generating NTA-buffered NiII competitions initially led to NiII competitions being performed without buffer before employing histidine-buffers. An extra replicate of histidine-buffered NiII-competition was also performed (Supplementary Fig. 2, Supplementary Table 2). Competition between MnII and bicinchoninic acid (BCA) buffered CuI confirmed that MncA has < 4 × 107-fold preference for CuI, consistent with the determined 4 × 104-fold preference (Supplementary Fig. 2d, Table 1). Preferences of metal-binding to MncA at folding relative to MnII were calculated from Table 1 to generate Fig. 1d. The order of binding follows the Irving-Williams series (Fig. 1a). The exchangeable forms of metals in the cytosol are thought to be divalent except copper, which is monovalent. Fig. 1d illustrates the challenge to predict the metalation states of proteins in vivo and to decode cognate metals, since here MnII seems least likely.
NiII availabilities defined by NiII-RcnR refine mid-range metal availabilities
Figure 1d presents the metal preferences of MncA after accounting for competition from other ligands (NTA or histidine). The speciation of protein metalation in the crowded cytosol is similarly thought to result from competition with diverse ligands binding labile, exchangeable metals at different availabilities23,30,31,32,33,34,35,36,37,38. Metal availabilities in the cytosol of bacterial cells (Salmonella and E. coli) have been estimated as free energies (∆GM): formally free energies for complex formation with a notional half-metalated ligand at the respective availability22,23. DNA-binding metal sensors detect changes in ∆GM and their responses have been calibrated for these bacteria22,23,39. However, the NiII responses of high NiII-sensing RcnR were previously overlooked, and its calibration requires determination of NiII affinity plus DNA affinity of NiII-RcnR (Supplementary Fig. 3a)23,24,40,41.
RcnR absorbance changes with NiII and a difference spectrum saturates at one equivalent per monomer, or 4:1 NiII:RcnR4 (Fig. 2a). RcnR co-elutes with one NiII atom per monomer by SEC (Supplementary Fig. 3b). EGTA competes for NiII with RcnR enabling determination of RcnR NiII affinity 2.36 (± 0.13) × 10−12 M from a simultaneous fit to n = 4 titrations in different EGTA concentrations (Supplementary Fig. 3c). A representative data set confirms the fitted value is within the limits of the assay from dashed lines simulating values ten times tighter and weaker (Fig. 2b). Binding of apo- and of CoII-RcnR to DNA was previously monitored by fluorescence anisotropy using hexachlorofluorescein-labelled rcnA operator-promoter fragments, to determine DNA affinities39. Here analogous titrations confirm that NiII similarly weakens DNA-binding with fitted affinity of 3.09 (± 0.04) × 10−6 M (Fig. 2c). Using these values, with previous apo-RcnR DNA affinity of 1.5 × 10−7 M, plus known RcnR molecules cell−1, the relationship between intracellular NiII availability and rcnA operator-promoter occupancy was calculated as for CoII (Supplementary Data 2)23. rcnA transcripts increase as RcnR DNA occupancy decreases with elevated intracellular NiII following the relationship in Fig. 2d. Metal availabilities at DNA occupancies of 0.99, 0.90, 0.10 and 0.01 are indicated. Analogous relationships for other metal sensors are in Supplementary Fig. 4. The bars on Fig. 2e show ranges of availability (as ∆GM) corresponding to occupancies of 0.1–0.923. For NiII the range in Fig. 2e now combines those for NiII-NikR and NiII-RcnR. The mid-point of the combined range is shown along with the mid-points for other metals, annotated as concentrations (M) and free energies (kJ mol−1) (Fig. 2e). The Irving-Williams series is ambiguous about the order of ZnII versus NiII, but both are weaker than copper (Fig. 1a). Including NiII-RcnR, Fig. 2e reverses the order of intracellular availabilities of ZnII and NiII shown in previous iterations23, but still the sensors maintain availabilities to the inverse of the Irving-Williams series2.
a Apo-subtracted difference spectra of RcnR (17.2 µM RcnR monomer, ɛ calculated from total protein) titrated with NiII, inset showing peak wavelength confirming 1:1 stoichiometry of NiII to RcnR monomer, or 4:1 to RcnR4 (n = 1). RcnR (20 μM monomer) also migrated with one equivalent of NiII by SEC (Supplementary Fig. 3b). b Representative NiII titration of RcnR (31.5 μM monomer) in EGTA (464 μM), solid line representing calculated KD for NiII from simultaneously fitted RncR-EGTA competitions (n = 4 independent experiments) at varied EGTA concentrations (Supplementary Fig. 3c, fitting models in Supplementary Software). Dashed lines, simulations with affinities 10-fold tighter and weaker than calculated KD. c RcnR binding to hexachlorofluorescein-labelled rcnRA operator-promoter (10 nM) by fluorescence anisotropy. Solid line, best simultaneous fit to n = 3 experimental replicates (circles, triangles, squares) for NiII-RcnR (fitting model in Supplementary Software), dashed line simulates apo-RcnR using published KD 1.5 × 10−7 M and maximum ∆robs 0.111539. d NiII and DNA affinities determined above used with previously measured RcnR molecules cell−1 to calculate (via Supplementary Data 2) relationship between intracellular NiII availability and RcnR DNA occupancy (θD), as for CoII (circles show θD 0.99, 0.90, 0.10 and 0.01). Combined mid-point of ranges for NiII-RcnR and NiII-NikR also shown (blue arrow). e Metal availabilities (squares and inset text) as activities/concentrations and free energies, ∆GM, at mid-points (50% DNA occupancies, representing idealised cells) of metal sensors (bars are sensor ranges, 10% to 90%), now including NiII-RcnR. Pale bars show individual ranges where two cognate sensors. MncA metal preferences (pale blue circles, CuI triangles, from Fig. 1d) as free energies (∆GMP) from pseudo-affinities giving 99% MnII metalation at mid-range MnII availability (∆GM) without competing metals (Supplementary Fig. 5 simulates alternative values for MnII-MncA ∆GMP). Inset shows occupancies of MncA predicted from free energy differences between MncA and labile metal (∆∆G = ∆GMP − ∆GM). MnII has the largest favourable gradient annotated ∆∆Gmax. Supplementary Data 3 enables similar predictions of cognate metals for other proteins. Source data are provided as a Source Data file.
NiII-RcnR refined mid-range metal availabilities decode correct metalation
Metalation within a cell should be predictable from the binding preferences of MncA in Fig. 1d relative to how tightly available (exchangeable) intracellular metals are bound, for example at the mid-points of sensor ranges (Fig. 2e). To make this comparison, MnII-MncA was assigned a pseudo-affinity (2.6 x 10-8 M) giving 99% MnII-metalation at the ∆GM mid-point for MnII, and free energies of metal-MncA complex formation were designated ∆GMP. Values (∆GMP) for other metals were then calculated using Fig. 1d (Table 1). The gradients from exchangeable cytosolic sites to MncA (∆∆G, formally ∆GMP - ∆GM)22,23, were calculated and inter-metal competition accounted for using the NiII-RcnR-refined calculator in Supplementary Data 3, to predict occupancies (Fig. 2e). The cognate metal MnII was correctly decoded (largest favourable ∆∆G). Pseudo-affinities ten times tighter or weaker generated the same proportional occupancies (Supplementary Fig. 5a, b). Thus, the speciation of metalation is a function of relative metal-binding preferences and availabilities, but total metal occupancy does vary with absolute ∆GMP values for MnII-MncA. Using the NiII-RcnR-revised mid-point availabilities with four exemplar proteins whose affinities have been measured22,23, decodes their cognate metals despite all preferring copper (Supplementary Fig. 5c–f). The NiII-RcnR revised calculator in Supplementary Data 3 can be used to decode cognate metalation.
Non-cognate metal is also kinetically trapped by MncA
The labile character of protein-bound metals creates a challenge to define the in-cell metalation states of metalloproteins. Metals can be lost, gained, or exchanged at cell lysis and/or during purification and analysis. MncA is attractive because MnII becomes trapped in the folded protein. However, MncA might adopt non-native folds with non-cognate metals. Abnormally coordinated metals might not be kinetically trapped. NiII often prefers four-coordinate, planar geometries, and NiII-MncA was chosen for structural analysis. A crystal structure of NiII-MncA (1.6 Å resolution, Supplementary Table 3, Supplementary Fig. 6a), shows the bi-cupin fold of MnII-MncA with a metal atom in each cupin domain (Fig. 3a)8. Each MnII atom is coordinated to three histidine and one glutamate residue with non-protein ligands completing hexacoordinate MnII coordination spheres. NiII-MncA has similar coordination environments with no channel to the amino-terminal NiII and a narrow channel to the carboxy-terminal site analogous to MnII-MncA (Fig. 3b, c)8. The channel is presumed to allow substrate access to catalytic MnII. The substrate site was gratuitously occupied by acetate in the MnII structure and with glycine in the crystalised NiII-form (Supplementary Fig. 6, Fig. 3c). The narrow hydrophobic channel is unlikely to allow NiII exchange and thus both NiII atoms appear trapped.
a Ribbon representation of crystal structure of (NiII)2MncA at 1.6 Å resolution showing characteristic bi-cupin fold and metal sites (boxed) modelled from residue 39 of the full protein sequence including signal peptide which is absent from the expressed protein (data collection and refinement statistics shown in Supplementary Table 3). Amino-terminal domain (pink), carboxy-terminal domain (blue) from residue 238 (in the full sequence). b, c Cross sections of solvent accessibility surfaces (modelled at 1.1 Å solvent radius to encompass dynamics) surrounding the metal sites show no channel to amino-terminal NiII and narrow lipophilic (hydrophobic) channel to carboxy-terminal NiII. NiII becomes trapped in the folded protein in a near octahedral geometry (yellow bonds; Supplementary Fig. 6) analogous to cognate MnII. The MncA model illustrates kinetic trapping of non-cognate metals suggesting MncA may be used to faithfully report in-cell metalation. Supplementary Fig. 7 shows that post-folding NiII did not exchange with CuII in vitro.
MncA has > 100-fold preference for CuII over NiII at folding (Fig. 1d). To test if NiII is trapped, NiII-MncA was incubated for 24 h in a two-fold molar excess of CuII, bound and free metal separated by SEC followed by ICP-MS (Supplementary Fig. 7). The protein remained exclusively bound to NiII confirming NiII is kinetically trapped.
Mis-metalation of MncA with FeII predicted and observed in E. coli
Figure 4 (unlike Fig. 2e) shows estimated metal availabilities in actual E. coli BW25113 (elsewhere E. coli) grown aerobically27, and the metalation of soluble MncA in E. coli. Transcripts regulated by metal-sensors were previously quantified by qPCR, then related to promoter occupancies and hence metal availabilities via the relationships in Supplementary Fig. 427. Notably, intracellular NiII availability in aerobically grown cells is below the range for NiII-RcnR. The largest favourable gradient from available exchangeable metal to MncA, ∆∆G, is for FeII not cognate MnII (Fig. 4c).
a Purification of soluble MncA folded in vivo, minus secretion signal, by anion exchange chromatography (5 mL Q-Sepharose, and 1 mL Q-Sepharose) both eluted using high-salt buffer, with intervening SEC (Superdex 200 or 75). MncA was recovered from a soluble protein fraction of E. coli after low-level expression overnight at 18 °C using pBAD30-mncA, induced with 0.02% w/v arabinose. Representative of n = 18 biologically independent purifications. Full gel image in Supplementary Fig. 8a. b Representative (n = 3 independent biological replicates) chromatogram showing in-cell acquired metals in MncA-containing fractions determined by ICP-MS revealing mis-metalation with FeII and traces of MnII. MncA-containing fractions were confirmed by SDS-PAGE (Supplementary Fig. 8d) and quantified by A280nm (using ɛ = 120,000 M−1). Metal contents of MncA-containing fractions (Supplementary Table 4a, b, Supplementary Fig. 8) were used to calculate fractional occupancies (%). Supplementary Table 4 shows similar outcomes via a modified protocol as used in subsequent experiments. c Metal availabilities (∆GM, squares and text inset) in the cytosol of E. coli grown aerobically in LB, estimated as in Foster et al. 27 by calibrated qPCR with genes regulated by cognate metal sensors (bars are sensor ranges from 1% to 99%). MncA metal-preferences (∆GMP, pale blue circles), CuI (triangle). Inset shows occupancies of MncA predicted from the free energy differences using Supplementary Data 4. The inset shows the resulting predicted MncA mis-metalation with FeII (dark blue columns) in the heterologous host based on the largest favourable gradient (∆∆Gmax in main figure). Inset also shows mean ( ± SD) in-cell metalation from the n = 3 biological replicates (pale blue columns, squares, circles, triangles, Supplementary Table 4b), closely matching predictions. Source data are provided as a Source Data file.
High-level protein expression in heterologous host cells can deplete cofactors, and likely contributed to formation of MncA-containing inclusion bodies (Fig. 1b). To monitor in-cell metalation, MncA was thus expressed at a low level from a tuneable promoter using low inducer, low temperature, and prolonged slow growth overnight in E. coli (Fig. 4a). MncA was enriched from a soluble protein extract via anion exchange and SEC followed by ICP-MS. Metalation was quantified in three biologically independent experiments (Supplementary Fig. 8, Supplementary Table 4a, b). Absorbance at 280 nm largely reports MncA concentration due to its high extinction coefficient (ɛ = 120,000 M−1 cm−1), evident from stoichiometry approximating 2:1, and suggesting no apo-protein (Supplementary Table 4a, b). Some metals might alter the extinction coefficient and traces of interfering proteins could introduce variation between experiments. Proportional occupancies have thus been calculated from total metal in one or more MncA-containing fraction(s) rather than MncA concentration, notably with similar outcomes when duplicated (Supplementary Table 4b). The experiment was repeated using a single anion exchange step and analytical HPLC SEC with similar outcome (Supplementary Table 4). The mean ± SD MncA-speciation is shown (inset Fig. 4c). Speciation accounting for inter-metal competition was calculated using Supplementary Data 4. Predicted FeII mis-metalation closely matches observed mis-metalation in E. coli (Fig. 4b, c).
Substantial cognate MncA metalation predicted and observed in 4 mM manganese
We were eager to know if intracellular MnII could be increased to a tolerable availability inside viable E. coli, sufficient to predominantly form MnII-MncA. Fig. 4c shows metal preferences (∆GMP) following a similar trend to availabilities (∆GM), predicting that modest changes could switch the speciation of metalation. Using only manganese supplementation maximum intracellular MnII availability was detected in media plus 4 mM manganese27. Notably, high manganese and hydrogen peroxide defined the 0.99 sensor boundary (Supplementary Fig. 4)27, because manganese import is modulated by OxyR42,43. Here, 4 mM manganese did not inhibit cell density after prolonged culturing (Supplementary Fig. 9). A switch to predominant MnII metalation was observed in three independent E. coli cultures supplemented with 4 mM manganese (Fig. 5a, inset Fig. 5b, Supplementary Table 5a, b, Supplementary Fig. 10).
a Representative (n = 3 independent biological replicates) chromatogram showing in-cell acquired metals in MncA-containing fractions determined by ICP-MS, as in Fig. 4a, b but using a single anion exchange step (Q-Sepharose), HPLC SEC (TSK SW3000) and showing increased cognate metalation with MnII. MncA was identified by SDS-PAGE in Supplementary Fig. 10c. Metal contents of MncA-containing fractions (Supplementary Table 5) from independent biological replicates (Supplementary Fig. 10) were used to calculate fractional occupancies (%). b Metal availabilities (∆GM, squares, and text inset) in E. coli cytosol grown aerobically as in Fig. 4c, except MnII replaced with estimates from calibrated qPCR of MntR target mntS in cells cultured in 4 mM manganese (Supplementary Fig. 4). MncA metal-preferences as ∆GMP (pale blue circles, CuI, triangle). Bars are sensor ranges from 1% to 99%. Inset shows occupancies of MncA predicted from free energy gradients using Supplementary Data 4 and substituting MnII availability with a value of 4.5 × 10−5 M (as in inset text). The inset shows predicted MncA metalation (dark blue columns) with MnII, based on the largest favourable gradient (∆∆Gmax in main figure), plus partial metalation with FeII. Inset also shows mean ( ± SD) in-cell metalation from the n = 3 independent biological replicates (square, circle, triangle, pale blue columns, as in Supplementary Table 5), largely matching predictions. Source data are provided as a Source Data file.
Calibrated mntS transcript abundance in 4 mM manganese reads out intracellular availability of 4.5 × 10-5 M (∆GM −24.8 kJ mol−1) (Supplementary Fig. 4)27. This value is shown in Fig. 5b plus availabilities for other metals in un-supplemented media. The largest favourable free energy gradient from exchangeable available metals to MncA becomes MnII (Fig. 5b). Metalation was predicted by substituting this elevated MnII availability into Supplementary Data 4 (inset Fig. 5b). A switch to predominant metalation with cognate MnII plus partial mis-metalation with FeII is thus predicted, as well as observed by MncA-trapping, in E. coli supplemented with 4 mM manganese.
MncA-trapped metals confirm negligible ZnII or CuI metalation in supplemented cells
We further wondered if MncA could be predictably metalated with other elements in metal-supplemented viable E. coli. Availabilities of ZnII and CuI at the upper (0.99) sensor boundaries (Supplementary Fig. 4)23, annotated in Fig. 6a, were entered into the metalation-calculator in Supplementary Data 4 as before. Negligible or no occupancy with either metal, but FeII mis-metalation, was predicted (Fig. 6a).
a In-cell metalation with NiII and CoII, not CuI or ZnII, switches from FeII-MncA in metal-supplemented (600 µM, 300 µM, 600 µM, 800 µM respectively as shown) media, qualitatively matching predictions but quantitatively greater NiII and CoII metalation than predicted. Metalation (dark blue bars) predicted using Supplementary Data 4 from availabilities in un-supplemented medium but with separately altered high availabilities of CuI, ZnII, NiII or CoII (as inset text). High availabilities correspond to the respective upper metal boundaries (θDM or θD, 0.99 or 0.01) shown in Supplementary Fig. 4. Mean ( ± SD) measurements of in-cell metalation (n = 3 independent biological replicates, square, circle, triangle) in high metal (pale blue bars) calculated using data in Supplementary Tables 6–9 based on Supplementary Figs. 11–15. b Selected intracellular metal availabilities (∆GM) refined (triangles) using Supplementary Data 5 for NiII (left) from observed in-cell FeII- and NiII-MncA occupancies in high NiII (relative to FeII in un-supplemented medium), and for FeII (right) from CoII and FeII occupancies in CoII (relative to elevated CoII as show in inset text). Metal preferences of MncA shown as ∆GMP (circles). Bars are sensor ranges from 1% to 99%. c Calculated metalation (dark blue bars) using Supplementary Data 4 but with refined availabilities as in panel (b), reduced residuals for MnII (which was not included in the refinement process) relative to observed in-cell metalation (pale blue bars included for comparison), showing mean ( ± SD) (n = 3 independent biological replicates, square, circle, triangle). MncA metalation can be used to refine relative intracellular metal availabilities and Supplementary Data 5 is provided to enable such calculations. Source data are provided as a Source Data file.
Predicted mis-metalation with FeII (negligible ZnII) was observed in cells grown in 800 µM zinc (Fig. 6a). This treatment was selected to approximate maximal abundance of zntA transcripts regulated by ZntR27, hence maximum ZnII availability with slight inhibition of growth (Supplementary Fig. 9). Chromatographic profiles for ZnII (unlike FeII and MnII) imperfectly align with absorbance at 280 nm or the distribution of MncA on SDS-PAGE (Supplementary Fig. 11). We did not identify the contaminating ZnII-protein by principal component analysis and MncA-metalation was determined from a single fraction showing least evidence of other proteins (Supplementary Fig. 11, Supplementary Table 6).
Predicted mis-metalation with FeII (not CuI) was also observed in cells grown in 600 µM copper (Fig. 6a, Supplementary Fig. 9). The profiles for copper (unlike FeII and MnII) again imperfectly align with MncA but correlated with a protein of Mr matching glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Supplementary Fig. 12). GAPDH is known to bind copper in copper-exposed cells and is removable using Blue Sepharose44. Including this step retained FeII-MncA and MnII-MncA but eliminated co-migrating copper (Supplementary Fig. 13). Metalation was quantified from three independent cultures after using Blue Sepharose (Supplementary Table 7a, b). In summary, intracellular availabilities of CuI or ZnII are insufficient to metalate MncA in supplemented cells, with metalation matching predictions.
MncA-trapped metals in high NiII and CoII refine availabilities
MncA trapped NiII or CoII in the respective metal supplemented media (Fig. 6a, Supplementary Fig. 9, 14, 15, Supplementary Table 8,9). Metalation was predicted as for high ZnII and CuI but using 0.01 boundary values for RcnR-DNA occupancy (Supplementary Fig. 4). In high cobalt, MncA bound more CoII, less FeII and more MnII, than predicted (Fig. 6a). The most parsimonious explanation is that FeII availability declines in high CoII. An MncA-residuals calculator in Supplementary Data 5, formulated in Supplementary Note 2, used the determined preferences of MncA to estimate the decrease in FeII availability relative to CoII (Fig. 6b). Figure 6c compares observed occupancies with predictions using the refined FeII availability. Residuals for MnII, not used in refinement, reduced as anticipated for the parsimonious solution.
In high NiII, MncA bound more NiII, less MnII and FeII, than predicted (Fig. 6a). Here, the most parsimonious explanation is that NiII is more available than approximated. The MncA-residuals calculator estimated the further increase in intracellular NiII availability relative to FeII (Fig. 6b). Figure. 6c compares NiII-refined occupancies with those observed. Residuals for MnII again reduced. In summary, MncA metal-occupancies plus the MncA-residuals calculator can be used to probe and refine estimates of intracellular metal availabilities. These tools can assist predictions of in-cell metalation of proteins in other cell types and growth conditions. MncA-derived refinements also suggest FeII availability declines in E. coli in high cobalt and this is subsequently explored.
Shallow MnII pool is depleted in un-supplemented media
Although MnII is a most available metal (least negative ∆GM, Fig. 4c), the MnII pool in un-supplemented media is shallow, containing only a few thousand atoms per cell26. qPCR with primers to MntR-regulated mntS revealed increased expression and hence reduced available MnII (MnII-MntR co-repression being alleviated) during prolonged MncA expression in un-supplemented medium (Fig. 7a). This is consistent with observing less MnII-MncA than predicted (Fig. 4c inset). Fur-regulated fepD transcripts indicate a smaller decline in the available FeII pool in un-supplemented medium. These data also highlight a challenge with qPCR for estimating intracellular metal availabilities at upper transcript abundance, corresponding to lower availabilities for metal-dependent co-repressors MntR and Fur. The logarithmic nature of PCR as represented by the log2 scales in Fig. 7, means the upper 50% of the expression range corresponds to a single PCR cycle. The abundance of mntS transcripts reaches (and exceeds) the upper boundary, lowest available MnII, reported previously whereas fepD remains 3 to 4 cycles below (Fig. 7a, b)27. These data thus report depletion of the intracellular MnII pool overnight by ~18 h but a relatively modest change in FeII. The use of MncA to refine estimated availabilities is thus especially valuable for qPCR-based estimates in the upper 50% of expression ranges. This applies to some metal-supplemented cells but not un-supplemented cells as cultured previously27. Viewed together, Fig. 4c (inset) and Fig. 7a reveal that disparities between predicted and observed MncA-metalation can report on non-steady-state intracellular metal availabilities (∆GM).
qPCR of a mntS (regulated by MnII-MntR), b fepD (FeII-Fur), c zntA (ZnII-ZntR), d copA (CuI-CueR), e rcnA (CoII-RcnR), f rcnA (NiII-RcnR), as log2 change relative to respective lowest measured transcript abundance observed in Foster et al.27 using chelator, metal and/or H2O2 supplemented cells ( ~ 0). ∆Cq values in Source Data TXT for Fig. 7, data are mean ± SD of n = 3 independent biological replicates (square, circle, triangle) except where dagger n = 2. MntR and Fur target gene transcripts increase in low MnII and FeII (MntR and Fur are metal-dependent co-repressors), but other transcripts increase in high availabilities of cognate metals (regulated by activators or metal-dependent de-repressors). Previous log2(fold change) qPCR at high metal-boundaries (red dashed lines, 0 for Fur) and at low metal boundaries (black dashed lines for MntR and Fur, 0 for others) are shown. Arrows show previous log2(fold change) qPCR for cognate transcripts isolated from cells grown in LB27. Panels (a) and (b) show expression at 0 h, 1 h, 1.5 h, 18 h after addition of arabinose to un-supplemented medium to induce low expression of MncA at 18 °C. After 18 h MntR regulated mntS transcripts pass the low MnII boundary (DNA occupancy θD 0.01) (panel a) further quantified in Supplementary Fig. 18. The MnII pool is shallow and depleted in LB. For Fur-regulated fepD the log2 scale reports modest depletion of intracellular available FeII at 18 h in un-supplemented medium (Supplementary Fig. 18). The first two columns (panels c–f) show negligible change in availability of ZnII, CuI, CoII or NiII after 18 h MncA expression in non-supplemented medium. Final four columns (panels a, c–f) report log2(fold change) in metal-supplemented media either with or without expression of MncA as indicated. Steady-state MnII availability approximating upper MnII boundary sustained to 18 h in 4 mM manganese (panel a). Steady-state approximating the upper metal boundary is also sustained in high NiII and CoII, but not CuI and ZnII (panels c–f). Source data are provided as a Source Data file.
Steady-state availabilities sustained for high MnII, NiII and CoII, not CuI and ZnII
At the outset of MncA expression, cells exposed to elevated metals all reach or exceed the estimated transcript abundances defined as the qPCR boundaries for high intracellular metal availabilities (Fig. 7). For MnII-MntR co-repressed mntS transcripts, this reflects high occupancy of the mntS promoter in cells exposed to 4 mM manganese and low transcript abundance. In contrast this equates to high transcript abundance for ZnII, CuI, CoII and NiII responsive transcripts regulated by activators ZntR, CueR and metal-dependent de-repressor RcnR. At the end of MncA production (18 h) in metal-supplemented cells, MnII, CoII and NiII availabilities remain high, but they decline for ZnII and CuI (Fig. 7a, c–f). This is also true for cultures that are not expressing MncA excluding the formal possibility that turnover of metalated MncA sustained elevated steady-state MnII, NiII and CoII availabilities but not CuI and ZnII. Lack of steady-state elevated CuI and ZnII indicates a difference in metallostasis for these metals compared to MnII, CoII and NiII. It may be more challenging to metalate proteins in viable cells with ZnII or CuI simply by metal supplementation.
Fur causes FeII atoms cell-1 to decline in CoII
To investigate the MncA-predicted decline in intracellular FeII availability in high CoII (Fig. 6b), total atoms of iron cell-1 were measured by ICP-MS (Fig. 8a). Iron content declined in CoII consistent with the prediction (noting that available metal can sometimes trend differently to total metal). Cells lacking Fur contain less iron in standard media as explained previously45. Importantly, the effect of CoII on iron is absent in ∆fur (Fig. 8a). Fur binds CoII and CoII-Fur binds DNA with both affinities already known46. Fig. 8b calculates the Fur response to intracellular CoII simulated as for FeII23. DNA occupancy at the CoII availability inside cells grown here in elevated cobalt reveals an extreme cross-response of Fur. Non-cognate sensors are known to cross-respond to availabilities at the top of the cognate sensor ranges39. Crucially, cross-metalation of Fur with CoII explains the decline in total iron in Fig. 8a and in available intracellular FeII read out by MncA (Fig. 6b).
a Total iron atoms cell−1 determined by ICP-MS of cell digests, declines in cells grown in medium supplemented with high CoII (left panel). This is consistent with the decline in intracellular FeII availability estimated from the residuals in Fig. 6a, calculated using Supplementary Data 5 and shown in Fig. 6b. Elevated NiII does not affect total iron atoms cell−1 (right panel). Total iron cell−1 is less in ∆fur but does not further decline in elevated CoII. The decline in iron cell−1 in response to elevated CoII is Fur-dependent. b Fur binds CoII to promote DNA-binding: simulated using known CoII affinity of Fur, DNA affinity of CoII-Fur, plus apo-Fur DNA affinity and Fur molecules cell−1, as simulated for FeII-Fur23. Fur promoters will be aberrantly repressed in 300 µM cobalt (square and blue arrow) causing the predicted decline in intracellular available FeII shown in Fig. 6b and observed decline in total iron atoms cell−1 in panel (a). c Total manganese atoms cell−1 increase in high CoII consistent with slightly increased intracellular available MnII, modellable from the remaining residuals in high CoII in Fig. 6c, calculated using Supplementary Data 5, shown in Supplementary Fig. 16b. These effects of cobalt on MnII are independent of Fur. Total manganese atoms cell−1 decrease in high NiII, this trend is more pronounced in ∆mntR (Supplementary Fig. 17), the trend is not evident in ∆fur which contains low manganese (right). An analogous decrease in available intracellular MnII calculable (Supplementary Data 5) from the remaining residuals in high NiII, Fig. 6b, is shown in Supplementary Fig. 16a. Use of MncA as a probe of relative intracellular metal availabilities iteratively refines values shown in Supplementary Fig. 16, included in Supplementary Data 6−8, and provided as blueprints to assist manipulation of the speciation of in vivo protein metalation. Mean ± SD of n = 3 independent biological replicates (squares, circles, triangles) in (a) and (c). Source data are provided as a Source Data file.
MncA-trapped metals iteratively refine estimated availabilities
The remaining residual differences between observed and predicted occupancies in Fig. 6c have been used via the MncA-residuals calculator (Supplementary Data 5), to further refine intracellular metal availabilities (Supplementary Fig. 16). MnII availability is inferred to decrease in high NiII and increase in high CoII. Both predicted trends were reflected in the total atoms cell−1, and the effects of CoII on MnII were not Fur-dependent (Fig. 8c). The effects of high NiII on MnII are not dependent on MnII-sensing MntR (Supplementary Fig. 17a). The MntR-independent decline in MnII is less severe in wild-type where the sensor controls metallostasis dampening changes. The residuals in cells exposed to 4 mM manganese suggest a decline in available FeII and this matched a Fur-independent decline in total iron atoms cell−1 (Fig. 5b, Supplementary Fig. 17b). The MncA-residuals calculator was used iteratively to further refine high NiII, CoII and MnII blueprints (Supplementary Fig. 16). The refined availabilities and related high-metal calculators are provided (Supplementary Data 6–8). MncA and the MncA-residuals calculator in Supplementary Data 5, could be used to probe relative metal availabilities in other cell types including (rating) cells with metallostasis engineered to optimise the speciation of metalation for selected elements.
Discussion
Here we discover that the metal-binding preferences of a protein which kinetically traps metals at folding, MncA, all follow the Irving-Williams series (Fig. 1). This may reflect the preferences of flexible sites in the folding pathway prior to kinetic trapping. It is also possible that preferred metals in the series increasingly encourage progression of MncA folding intermediates. Although copper is most preferred, the cognate metal is correctly decoded to be MnII by relating these preferences to idealised estimates of intracellular metal availabilities, standardised to the mid-points of the ranges of metal sensors (Fig. 2). MncA kinetically traps metals, such that those which co-purify are likely to reflect in vivo metalation states (Fig. 3, Supplementary Fig. 7). Estimates of metal availabilities inside E. coli grown aerobically predict that (cyanobacterial) MncA will be mis-metalated with FeII in E. coli, and this is established by ICP-MS analysis of purified soluble MncA (Fig. 4). Thus, the speciation of metalation is directly shown to be determined by relative binding preferences for metals competing at intracellular availabilities (Supplementary Fig. 5a), and mis-metalation can be predicted in heterologous hosts. Subtle changes in a metal availability (or preference) can switch metalation because intracellular availabilities follow the inverse of the Irving-Williams series and hence trend with preferences (Figs. 2c, 4c, 5b, 6a). Estimates of intracellular metal-availabilities can thus be used to predict the speciation of in vivo metalation to inform the engineering of natural and synthetic metalloproteins to optimise metalation (using Supplementary Data 3, 4, 6–8).
Unrefined metal availabilities derived from qPCR-based responses of metal-sensing transcriptional regulators correctly predicted FeII mis-metalation of MncA in E. coli in un-supplemented media (Fig. 4c). In metal-supplemented media, MncA occupancies with CoII, NiII and MnII exceeded (at least to some extent) predictions (Figs. 5, 6a). These occupancies were used with Supplementary Data 5 to refine estimates of intracellular availabilities predicting decreased available FeII in CoII and MnII, decreased available MnII in NiII, but increased available MnII in CoII (Fig. 6b, Supplementary Fig. 16). Encouragingly, in every case this coincided with analogous changes in total FeII or MnII atoms cell−1 (Fig. 8a, c, Supplementary Fig. 17a, b). The decline in cellular FeII in high CoII, but not in high MnII, was Fur-dependent and the former explained by responses to CoII of Fur (Fig. 8a, b, Supplementary Fig. 17b). The extent to which cross-metalation of Fur with CoII is disadvantageous mis-metalation, remains to be established. Together these data show the value of MncA as a probe of intracellular metal-availabilities, orthogonal to synthetic metal sensors47,48,49, and complementing calibrated endogenous metal sensors especially where qPCR is less reliable and/or crosstalk likely39.
MncA occupancies report lowered MnII availability inside cells after ~18 h growth in un-supplemented medium (Fig. 4). Here, the shallow MnII pool is depleted during MncA expression such that steady-state availability is not sustained (Fig. 7a, Supplementary Fig. 18). This is significant when recombinant metalloproteins are over-expressed in E. coli potentially depleting metals: metalation can then reflect depths of available metal pools rather than strengths of competition with other ligands. In manganese-, nickel- and cobalt-supplemented cells elevated intracellular availabilities were sustained for ~18 h but this was not the case in elevated zinc and copper (Fig. 7). Mechanisms of metallostasis for ZnII and CuI must adjust rates of import, export and/or consumption to restore pre-exposure steady-state intracellular availabilities, whereas elevated steady-states are maintained for the other metals (Fig. 7). Nonetheless, it is predicted that MncA would bind negligible CuI or ZnII even if elevated steady-state availabilities were sustained. Notably, more favourable gradients (∆∆G) for FeII and MnII largely exclude ZnII (Supplementary Fig. 19). Less FeII and MnII in the media, engineering cells to have reduced availabilities of FeII and MnII, or weakening MncA binding to FeII and MnII, could assist accumulation of ZnII-MncA.
Metalation of MncA inside E. coli switched between MnII, FeII, CoII, and NiII as predicted (Figs. 4–6). This becomes achievable because intracellular metal availabilities track with metal-binding preferences of proteins. This could indicate that protein mis-metalation is likely in vivo. Alternatively, mechanisms of metal homoeostasis could be exceptionally finely tuned to native metallo-proteomes minimising mis-metalation. In either eventuality, this raises the prospect in engineering biology of widespread mis-metalation because metal-availabilities will not necessary be correctly tuned to non-native metalloproteins. Notably, in E. coli, GTP-dependent metallochaperones YeiR and YjiA are predicted to be predominantly metalated with presumed cognate ZnII22, whereas heterologous CobW from Rhodobacter, CbiK from Salmonella, in common with cyanobacterial MncA become mis-metalated with ZnII, FeII and FeII respectively (Fig. 4, Supplementary Fig. 20). Quantification of B12 production in engineered E. coli provided indirect evidence of the predicted mis-metalation of CobW, and CbiK is known to insert FeII into siroheme in Salmonella missing the siroheme chelatase CysG9,22. Crucially, MncA now provides a direct read-out confirming predicted mis-metalation with FeII (Fig. 4).
MncA-refined intracellular metal availabilities along with the calculators provided here can guide optimisation of metalation via use of metal supplements or chelants, alterations to homeostasis by engineering host strains (chassis), or engineering proteins to match availabilities. Approximately half the reactions of life rely on the chemistries of the correct metals bound to metalloproteins1. Considerable research is being directed to the generation of various types of artificial metalloenzymes29,50,51. Inevitably, efforts in metabolic engineering, synthetic biology and directed evolution will often rely on metalloenzymes52,53,54. The blueprints and calculators can inform metalloenzyme engineered for in vivo bioprocessing applications. The optimisation of in vivo metalation presents opportunities as the engineering of biological systems for (sustainable) bio-manufacturing is prioritised55.
Methods
Expression and purification of unfolded apo-MncA
To examine the relative binding preferences of MncA in vitro, protein was expressed and purified8. Briefly, MncA minus TAT secretion signal was expressed from pET29a-mncA in E. coli BL21(DE3) pLysS. Isopropyl β-D-1-thiogalactopyranoside (IPTG, 1 mL 0.4 M) was added to a mid-log phase culture (1 L in a 2 L flask) ~ OD600nm 0.6–0.8 at 37 °C, to induce high-level expression for 3 h before harvesting cells by centrifugation (4000 × g, 4 °C) and freezing (−20 °C). The pellet was resuspended in 30 mL 100 mM Tris pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulphonyl fluoride (PMSF) and sonicated (4 × 10 s on ice, 1 min intervals). Lysate was cleared by centrifugation (27,000× g, 15 min, 4 °C), supernatant discarded, and the pellet resuspended in 100 mM Tris pH 7.5, 100 mM NaCl, 1% (v/v) Triton X-100 (30 mL), sonicated (3 × 10 s on ice, 1 min intervals) and centrifugation repeated. The pellet was washed in 30 mL 100 mM Tris pH 7.5, 100 mM NaCl to remove the Triton, then sequentially in 30 mL 50 mM HEPES pH 7.5, 1 M urea followed by 15 mL 50 mM HEPES pH 7.5, 2 M urea. During the final wash the lysate was split into 2 mL aliquots followed by centrifugation (15,890 × g, 10 min) to recover inclusion bodies, stored at −20 °C.
Preparation of metal stocks
Metal stocks (MnCl2, CoCl2, NiSO4, CuSO4 and ZnSO4) in ultrapure water were sterile filtered (0.2 μm) and quantified by inductively coupled plasma mass spectrometry (ICP-MS). When required for anaerobic experiments, (NH4)2Fe(SO4)2 stocks were prepared in N2-purged ultrapure water. Total Mn and Fe concentrations were then confirmed by ICP-MS. FeII stock was confirmed to be > 95% reduced by reaction with excess (10-fold) ferrozine (Fz) using ε562nm = 27,900 cm-1 M-1 for the FeIIFz3 complex. Reduced CuI stocks were prepared as described in specific experiments and cuprous state validated with bicinchoninic acid (BCA) and ICP-MS.
Production of buffered competing metals
MncA in vitro refolding buffers (in 50 mM MOPS pH 7.5) contained pairs of metals and ligand in varied amounts to achieve different buffered [available metals]. Buffers were prepared in acid-washed flasks with components added in the order, pH buffer, ultrapure water, ligand (typically NTA, or 1 mM L-histidine for competitions with NiII) followed by the two metals. Buffers were typically prepared in 100 mL volumes, filtered via 0.45 μm filters if any evidence of light scatter. Supplementary Table 1 shows the total amounts of each metal and ligand (NTA, histidine or neither) plus the buffered [available metals] used in each refolding solution. Supplementary Data 1 (derivation in Supplementary Note 1) was used and is provided here to assist in the production of such paired metal buffers. For an effective buffer the total metal concentration should substantially exceed the protein concentration ( > 100-fold generally, 10-fold for competitions with CuI). To achieve an effective buffer < 80% of the buffering agent should be metalated.
A preliminary experiment was performed with BCA to buffer CuI with no expected buffering of MnII. With 400 µM BCA and 50 µM CuI, expected free [CuI] is 10−15 M, which was competed against 10 µM MnCl2. Because < 1% copper co-migrated with MncA, unbuffered ZnII was competed against unbuffered CuI. Typical concentrations were 10 µM each and were verified by ICP-MS on the folding solution. Competitions between ZnII and CuI, used freshly prepared solution of CuCl (10 mM CuCl, 1 M NaCl, 0.1 M HCl) and refolding experiments included 1 mM hydroxylamine to maintain copper in reduced form. For buffers forming 2:1 metal-dependent complexes, buffered concentrations of each metal were determined using HySS software for His56. Logβ values for all equilibria are as follows: proton dissociation from histidine (HisH-9.08, HisH2-15.1, HisH3-16.8), His complexation of MnII (MnIIHis-3.3, MnIIHis2-6.3), His complexation of NiII (NiIIHis-8.67, NiIIHis2-15.54), pKw = 13.857. For CuI BCA complexes, derivation of buffering of [CuI] in Supplementary Note 1.
Metal-binding preferences of MncA at folding
Inclusion bodies containing MncA were solubilized in HEPES pH 7.5 with 8 M urea, and the concentration calculated from A280nm using an experimentally determined extinction coefficient (120,000 cm−1 M−1)8, typically in the range 100–500 μM. Refolding was achieved by dropwise dilution of urea-solubilised MncA in large volumes (100 mL) of solutions containing pairs of competing buffered metals prepared as described earlier. Solutions were thoroughly mixed before adding unfolded MncA and gently mixed between additions. Dilute refolded MncA was recovered by binding to a 1 mL Q-Sepharose (Cytiva or GE Healthcare) anion exchange column pre-equilibrated with low-salt buffer (50 mM Tris pH 7.5, 50 mM NaCl). The column was washed with 20 mL low-salt buffer before eluting MncA with high-salt buffer, 50 mM Tris pH 7.5, 500 mM NaCl. MncA was quantified by A280nm then resolved (0.5 mL at ~10 μM) by SEC (PD-10, GE healthcare, previously washed with 0.5 mL 5 mM EDTA followed by ultrapure water and equilibrated with the low-salt buffer). Fractions (0.5 mL) were analysed for protein via A280nm and metal by ICP-MS using corresponding matrix-matched calibration curves. To compete MnII versus FeII and CuI, stocks were prepared as described earlier and refolding done in an anaerobic chamber. The concentrations of MncA (A280nm) and metals (ICP-MS) were superimposed to identify metals co-eluting with MncA and estimate the metal:protein stoichiometry. The ratio of trapped metals was used to determine the relative binding preferences of the two metals at folding in buffer of known competing metal availabilities (Table 1, Supplementary Tables 1, 2).
(NiII)2MncA crystal structure
A 100 µM solution of MncA in 50 mM HEPES pH 7.5 and 8 M urea, prepared from pelleted inclusion bodies, was added dropwise, with stirring to 100 mL of 50 mM MOPS, pH 7.5, and 10 µM NiSO4, passed through a 0.22 µm filter and loaded on a 5 mL Q-Sepharose column (Cytiva) equilibrated with 50 mM Tris, pH 7.5, and 50 mM NaCl. The column was washed with the same buffer. Folded, concentrated (NiII)2MncA was eluted with buffer containing 500 mM NaCl. Buffer was exchanged via several cycles of dilution and concentration (Amicon Ultracel, 0.5 mL 10 kDa) to obtain MncA (10 mg/mL) in 10 mM Tris, pH 7.5, 50 mM NaCl. Crystals were grown in 100 mM sodium acetate, pH 4.0 and 8–10% PEG 80008. An aliquot of crystals redissolved in 50 mM Tris, pH 7.5, 50 mM NaCl were analysed by ICP-MS and A280nm measured confirming stoichiometric metalation with NiII. Following cryo-protection by the stepwise addition of glycerol to 20% v/v, crystals were flash-cooled and stored in liquid nitrogen prior to data collection.
Data were collected at beamline I04, Diamond Light Source (Supplementary Table 3). A highly redundant data set was collected by obtaining 360˚ of data in four separate scans along the axis of a long, hexagonal rod-shaped crystal. The data were processed using the xia2 package at Diamond58, which employed XDS to integrate59, and AIMLESS to scale and merge the data60. The structure solution was obtained by molecular replacement with the trimeric (MnII)2MncA structure (PDB ID 2VQA) as search model using Phaser61 implemented in Phenix62. Refinement continued in Phenix alternating with modelling in Coot63. Figures were prepared with ChimeraX64, and PyMOL (Schrödinger) software. MOLE 2.5 was used to identify channels present in the (NiII)2MncA structure65.
Expression and purification of RncR
Purification of RcnR overexpressed in E. coli BL21(DE3) from coding sequences cloned in pET29a has been described39,66. In common with the other regulators (MntR, Fur, NikR, ZntR, Zur, CueR), the sequence was from Salmonella enterica serovar Typhimurium strain SL1344 (referred to as Salmonella), and RcnR shares 100% sequence identity to E. coli RcnR. Anaerobic, reduced and apo-RcnR was prepared by applying purified, EDTA treated, apo-protein to a 1-mL HiTrap heparin column, transferred to an anaerobic chamber, washed with > 10 column volumes of Chelex-treated, N2-purged 240 mM KCl, 60 mM NaCl, 10 mM HEPES, pH 7.0, then eluted with 800 mM KCl, 200 mM NaCl, 10 mM HEPES, pH 7.0. RcnR was quantified by A280nm using experimentally determined extinction coefficient of 2,422 M−1 cm−1 obtained via quantitative amino acid analysis. Reduced thiol and metal content were assayed66,67, and all anaerobic protein samples (maintained in an anaerobic chamber) were ≥ 90% reduced and ≥ 95% metal-free. All in vitro experiments were carried out under anaerobic conditions using Chelex-treated and N2-purged buffers66,67.
NiII stoichiometry and affinity of RcnR
All experiments were conducted in 100 mM NaCl, 400 mM KCl, 10 mM HEPES pH 7.5. To determine stoichiometry, NiII (as NiCl2) was titrated into purified protein (17.2 µM) and absorption spectra recorded at equilibrium using a λ35 UV-visible spectrophotometer (Perkin Elmer Life Sciences). Additionally, an aliquot of RcnR (20 µM monomer) was incubated with NiII (30 µM) and bound metal resolved by SEC eluted with 100 mM NaCl, 400 mM KCl, 10 mM HEPES pH 7.5 (PD-10, collecting 0.5 mL fractions) and analysed for protein by Bradford assay standardised with known concentrations of RcnR and metal by ICP-MS. To determine NiII affinity, titrations were performed in the presence of EGTA using four RcnR monomer concentrations; 40.4 μM RcnR and 464 μM EGTA, 31.5 μM RcnR and 471 μM EGTA, 25.3 μM RcnR and 479 μM EGTA, 15.3 μM RcnR and 243 μM EGTA, monitoring a NiII-dependent feature of NiII-RcnR at 326 nm. A simultaneous fit was made to all data sets using Dynafit (fitting models in Supplementary Software)68.
NiII-RcnR DNA-affinity by fluorescence anisotropy
Fluorescently-labelled (hexachlorofluorescein) double-stranded DNA probes containing the identified RcnR-binding site upstream of the rcnA promoter were synthesised and annealed as described39,66. NiII-RcnR (1:1 NiII:RcnR tetramer) or apo-RcnR were titrated into 10 nM DNA in 60 mM NaCl, 240 mM KCl, 10 mM HEPES pH 7.5. Changes in anisotropy (Δrobs) were measured using a modified Cary Eclipse fluorescence spectrophotometer (Agilent Technologies) fitted with polarising filters (λex = 530 nm, λem = 570 nm, averaging time = 15 s, replicates = 3, and T = 25 °C), allowing the cuvette to equilibrate (3 min) before recording. A simultaneous fit was made to all data sets (fitting models in Supplementary Software using maximum ∆robs39,66).
Expression and purification of soluble MncA to determine in vivo metalation
The coding region of mncA as in pET29a-mncA was sub-cloned to create pBAD30-mncA to enable tuned expression dependent on [arabinose]. E. coli BW25113 (hereafter E. coli) containing pBAD30-mncA was inoculated into overnight cultures (10 mL LB + 0.2% w/v glucose + carbenicillin at 37 °C, 180 rpm) used the following day to inoculate 2 L flask containing 1 L LB medium and carbenicillin (no glucose) and incubated at 37 °C, 180 rpm, until OD reached mid-log phase ~OD600nm 0.6–0.8, ~3 h. Cultures were transferred to 18 °C and a low concentration (0.02% w/v) L-arabinose added to induce low-level gene expression followed by ~18 h culturing overnight (3 h + 18 h = ~21 h total) before harvesting cells by centrifugation (4000 × g, 4 °C). Purification of soluble in vivo metalated MncA involved protocols analogous to procedures used to recover native MncA from Synechocystis8. The entire cell pellet (from 1 L culture) was resuspended in 30 mL lysis buffer (20 mM Tris pH 7.5, 1 mM EDTA, 1 mM PMSF) sonicated (4 min pulsing) and centrifuged (27,000 × g, 4 °C) 45 min to remove cell debris. Supernatant was loaded onto a Q-Sepharose anion exchange column (5 mL, pre-equilibrated with 20 mM Tris pH 7.5) then washed with the same buffer. MncA was eluted using a 0–300 mM NaCl gradient in 30 mL, collecting 1 mL fractions. SDS-PAGE identified MncA-containing fractions which were pooled ( ~ 3–5 mL). Further purification and analysis used a more rapid analytical protocol in later experiments while in earlier experiments MncA was loaded onto Superdex 75 or Superdex 200 SEC columns (as specified) with fractions analysed by SDS-PAGE. MncA containing fractions were diluted to [NaCl] ≤ 50 mM, reapplied to Q-Sepharose (1 mL) washed then eluted with 300 mM NaCl and fractions analysed by SDS-PAGE, A280nm, and [metal] by ICP-MS. Proportional (%) occupancies of MncA with each metal was first calculated from the ratio of [metal]/[MncA] assuming 2 metal sites per MncA molecule, confirmed by mean experimental occupancy of 99% (replicates 1–3 in Supplementary Table 4a, b). The protocol was simplified to two steps with MncA first recovered via anion exchange (5 mL Q-Sepharose column, 0–300 mM NaCl gradient) followed by rapid analytical scale SEC using a SW3000 (TSK) column8. Fractions were again analysed by SDS-PAGE, A280nm and ICP-MS. A comparative fourth biological replicate of MncA extracted from cells grown without metal supplementation was purified via the simplified approach obtaining similar metal occupancies (Supplementary Table 4). Additionally similar occupancies were calculated as a proportion of the total metal content of MncA avoiding variation in MncA ɛ280nm with different metals (Supplementary Table 4b). In subsequent metal-supplemented cultures, respective metals were added at inoculation 1–3 h prior to L-arabinose addition. MncA isolated from cells supplemented with copper co-purified with a copper-protein tentatively identified as GAPDH. Fractions containing MncA were passed over a 5 mL Cibacron blue Sepharose (Blue Sepharose) column equilibrated with 50 mM Tris pH 7.5. MncA eluted rapidly while the contaminating protein was retained then eluted with 1 M NaCl in 50 mM Tris pH 7.5. Blue Sepharose-treated MncA-containing solution was subjected to analytical SEC as above.
Estimation of transcript abundance in E. coli
Two extracts (1 mL) were collected from E. coli, including from cells containing pBAD30-mncA expressing MncA immediately before arabinose addition, and secondly after overnight growth, and RNA stabilised using RNAProtect Bacteria Reagent, 2 mL (Qiagen). Samples were processed as described27. Briefly, RNA was extracted using RNeasy Mini Kit (Qiagen), [RNA] estimated from A260nm then treated with DNase I (Fermentas). ImProm-II Reverse Transcriptase System (Promega) generated cDNA, with parallel control reactions excluding reverse transcriptase. Transcript abundance was determined using primers for mntS, fepD, rcnA, nikA, znuA, zntA, copA and rpoD that amplify ~100 bp of DNA with sequences listed in Supplementary Data 9. Quantitative polymerase chain reaction (qPCR) analysis was executed in 20 μL reactions containing 5 ng of cDNA, 400 nM of each complementary primer and PowerUP SYBR Green Master Mix (Thermo Fisher Scientific). Three technical replicates of each biological replicate were analysed using a Rotor-Gene Q 2plex (Qiagen; Rotor-Gene-Q Pure Detection Software) with additional control reactions without cDNA templates (qPCR grade water used instead, supplied by Thermo Fisher Scientific) run for each primer pair, in addition to control reactions without reverse transcriptase for the reference gene primer pair (rpoD). Cq values were calculated with LinRegPCR (version 2021.1) after correcting for amplicon efficiency (Supplementary Source Data TXT, shown in Fig. 7). Change in gene abundance, relative to the control condition (defined as the condition where the minimum transcript abundance was observed for each target gene), was calculated using the 2–ΔΔCT method69 using rpoD as the reference gene and presented as log2(fold change).
Intracellular metal availabilities and predictions of in vivo metalation
Responses of metal sensors (θD for DNA occupancies of metal-dependent de-repressors and co-repressors, θDM for metalated activators) as a function of intracellular available buffered metal concentrations were calculated using sensor metal affinities, DNA affinities, protein abundances and numbers of DNA-binding sites for E. coli sensors as described23 (Supplementary Data 2 for NiIIRcnR). Transcript abundance was correlated with the response curves to enable estimations of intracellular metal availability expressed as a free energy for complex formation (∆GM) in E. coli grown aerobically in LB as described27. Using these availabilities, metalation of proteins was predicted in vivo, accounting for multiple inter-metal competitions including competition from the intracellular buffer as described by Young and coworkers22. Supplementary Data 3 and 4 perform these calculations for ideal cells (sensors at mid-range, including NiIIRcnR) and at intracellular metal availabilities in E. coli grown aerobically in LB respectively.
In metal-supplemented cultures, log2(fold change) relative transcript abundance in this work approximated values reported previously with some boundaries exceeded27 (Fig. 7). The reported values for high intracellular availabilities of supplemented metals were thus substituted into Supplementary Data 4, while retaining original values for all other metals, to make first predictions of MncA metalation at high intracellular MnII, CoII, NiII, CuI or ZnII. For MnII, CoII and NiII, residual differences between predicted and observed MncA metalation was used to iteratively refine intracellular metal availabilities according to Supplementary Note 2 and using Supplementary Data 5. Supplementary Data 6–8 contain the refinements and can be used to predict metalation of other proteins in E. coli in elevated MnII, NiII or CoII.
Elemental analyses by ICP-MS and metal atoms cell-1
ICP-MS was performed at the Durham University Bio-ICP-MS Facility (ThermoFisher iCAP RQ model) with matrix matched standard curves and internal silver standards. E. coli Δfur and ΔmntR strains were obtained from the Keio collection70. Individual colonies of E. coli or mutants were inoculated in LB (5 mL) shaking at 37 °C for 3–4 h, diluted into fresh medium (10 mL) in a 50 mL conical centrifuge tube supplemented with metals (MnCl2, CoCl2, NiSO4) where specified, to OD600nm 0.008. Cells were incubated with shaking at 180 rpm, 37 °C, overnight, 100 μL diluted 1:10 to measure OD600, and cells recovered (from remaining 9.9 mL) by centrifugation. Pellets were washed four times by resuspension in 1 mL wash buffer (20 mM Tris pH 8.5, 0.5 M sorbitol, 0.2 mM EDTA) followed by centrifugation. Ultrapure HNO3 (Merck) 65% v/v (0.4–0.5 mL) was added to each pellet and allowed to incubate for a minimum of 16 h until fully digested. The samples were then prepared for ICP-MS with matrix-matched calibration curves. An OD600nm of 1 equated to a cell count (CASY cell counter) of 6.47( ± 0.09) × 108 cells mL−1 using this strain, enabling atoms cell−1 to be calculated from calibrated ICP-MS data.
Statistics and reproducibility
Sample sizes were chosen based on prior experimental experience, and to give consistent results, following convention in the literature for equivalent analyses. Experiments designed to derive quantitative values used to predict or test and measure metalation, or to refine and evaluate estimates of intracellular metal availabilities, were performed in triplicate (n = 3) or more (n = 4) to enable calculation of SD (listed in tables or text or shown as error bars on figures) or SE for NiII-RcnR affinities. Predictions of metalation do not propagate SD from contributing values. Analogous chromatograms to the representative data in Figs. 4b and 5a were obtained on two further occasions (n = 3), data in Figs. 4a and 1b are representative of 18 and > 3 analogous purifications respectively. The number of independent experiments or biologically independent samples is otherwise shown in figure legends or footnotes of Tables.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All data are available within the article, its Supplementary Information files, plus PDB entry 9GOF and from corresponding authors on request. Source data are provided with this paper as Source Data files. Excel spreadsheets (with instructions) providing a calculator to formulate competing metal buffers, to calculate DNA occupancy as a function of NiII availability for metal-dependent de-repressor RcnR and providing a calculator to use in vivo recovered metal occupancies of MncA as a probe to refine estimates of intracellular metal availabilities, are provided as Supplementary Data 1, 2 and 5 respectively. Excel spreadsheets constituting calculators of metalation in NiII-RcnR-refined idealised cells, E. coli grown aerobically in LB, E. coli grown aerobically in LB supplemented with manganese, nickel, and cobalt, are provided as Supplementary Data 3, 4, 6–8 respectively. Supplementary Data 9 contains oligonucleotide sequences. Published structures used here for MncA and MntR are PDB entries 2VQA and 9C4D respectively. Source data are provided with this paper.
Code availability
Equation derivations are in Supplementary Notes 1 and 2 of the Supplementary Information. Dynafit scripts are provided in Supplementary Software.
References
Waldron, K. J., Rutherford, J. C., Ford, D. & Robinson, N. J. Metalloproteins and metal sensing. Nature 460, 823–830 (2009).
Irving, H. & Williams, R. J. P. Order of stability of metal complexes. Nature 162, 746–747 (1948).
Da Silva, J. F. & Williams, R. J. P. The biological chemistry of the elements: the inorganic chemistry of life. (Oxford University Press, 2001).
Robinson, N. J. & Glasfeld, A. Metalation: nature’s challenge in bioinorganic chemistry. J. Biol. Inorg. Chem. 25, 543–545 (2020).
Foster, A. W., Osman, D. & Robinson, N. J. Metal preferences and metallation. J. Biol. Chem. 289, 28095–28103 (2014).
Grāve, K., Griese, J. J., Berggren, G., Bennett, M. D. & Högbom, M. The Bacillus anthracis class Ib ribonucleotide reductase subunit NrdF intrinsically selects manganese over iron. J. Biol. Inorg. Chem. 25, 571–582 (2020).
Cotruvo, J. A. Jr & Stubbe, J. Metallation and mismetallation of iron and manganese proteins in vitro and in vivo: the class I ribonucleotide reductases as a case study. Metallomics 4, 1020–1036 (2012).
Tottey, S. et al. Protein-folding location can regulate manganese-binding versus copper- or zinc-binding. Nature 455, 1138–1142 (2008).
Raux, E., Thermes, C., Heathcote, P., Rambach, A. & Warren, M. J. A role for Salmonella typhimurium cbiK in cobalamin (vitamin B12) and siroheme biosynthesis. J. Bacteriol. 179, 3202–3212 (1997).
Schubert, H. L., Raux, E., Wilson, K. S. & Warren, M. J. Common chelatase design in the branched tetrapyrrole pathways of heme and anaerobic cobalamin synthesis. Biochemistry 38, 10660–10669 (1999).
Lacasse, M. J., Douglas, C. D. & Zamble, D. B. Mechanism of selective nickel transfer from HypB to HypA, Escherichia coli [NiFe]-hydrogenase accessory proteins. Biochemistry 55, 6821–6831 (2016).
Watanabe, S. et al. Structural basis of a Ni acquisition cycle for [NiFe] hydrogenase by Ni-metallochaperone HypA and its enhancer. Proc. Natl. Acad. Sci. 112, 7701–7706 (2015).
Banci, L. et al. The Atx1-Ccc2 complex is a metal-mediated protein-protein interaction. Nat. Chem. Biol. 2, 367–368 (2006).
Yang, X., Li, H., Lai, T.-P. & Sun, H. UreE-UreG complex facilitates nickel transfer and preactivates GTPase of UreG in Helicobacter pylori. J. Biol. Chem. 290, 12474–12485 (2015).
Crouzet, J. et al. Nucleotide sequence and genetic analysis of a 13.1-kilobase-pair Pseudomonas denitrificans DNA fragment containing five cob genes and identification of structural genes encoding Cob(I)alamin adenosyltransferase, cobyric acid synthase, and bifunctional cobinamide kinase-cobinamide phosphate guanylyltransferase. J. Bacteriol. 173, 6074–6087 (1991).
Blaby-Haas, C. E., Flood, J. A., de Crécy-Lagard, V. & Zamble, D. B. YeiR: a metal-binding GTPase from Escherichia coli involved in metal homeostasis. Metallomics 4, 488–497 (2012).
Weiss, A. et al. Zn-regulated GTPase metalloprotein activator 1 modulates vertebrate zinc homeostasis. Cell 185, 2148–2163. e2127 (2022).
O’Halloran, T. V. & Culotta, V. C. Metallochaperones, an intracellular shuttle service for metal ions. J. Biol. Chem. 275, 25057–25060 (2000).
Cavet, J. S. et al. A nickel-cobalt-sensing ArsR-SmtB family repressor: contributions of cytosol and effector binding sites to metal selectivity. J. Biol. Chem. 277, 38441–38448 (2002).
Guedon, E. & Helmann, J. D. Origins of metal ion selectivity in the DtxR/MntR family of metalloregulators. Mol. Microbiol. 48, 495–506 (2003).
Moomaw, E. W. et al. Metal dependence of oxalate decarboxylase activity. Biochemistry 48, 6116–6125 (2009).
Young, T. R. et al. Calculating metalation in cells reveals CobW acquires CoII for vitamin B12 biosynthesis while related proteins prefer ZnII. Nat. Commun. 12, 1195 (2021).
Osman, D. et al. Bacterial sensors define intracellular free energies for correct enzyme metalation. Nat. Chem. Biol. 15, 241–249 (2019).
Grossoehme, N. E. & Giedroc, D. P. Energetics of allosteric negative coupling in the zinc sensor S. aureus CzrA. J. Am. Chem. Soc. 131, 17860–17870 (2009).
Foster, A. W. et al. A tight tunable range for Ni(II) sensing and buffering in cells. Nat. Chem. Biol. 13, 409–414 (2017).
Outten, C. E. & O’Halloran, T. V. Femtomolar sensitivity of metalloregulatory proteins controlling zinc homeostasis. Science 292, 2488–2492 (2001).
Foster, A. W. et al. Metalation calculators for E. coli strain JM109 (DE3): aerobic, anaerobic, and hydrogen peroxide exposed cells cultured in LB media. Metallomics 14, mfac058 (2022).
Young, T. R. et al. Two distinct thermodynamic gradients for cellular metalation of vitamin B12. JACS Au 3, 1472–1483 (2023).
Choi, T. S. & Tezcan, F. A. Overcoming universal restrictions on metal selectivity by protein design. Nature 603, 522–527 (2022).
Chivers, P. T., Basak, P. & Maroney, M. J. One His, two His… the emerging roles of histidine in cellular nickel trafficking. J. Inorg. Biochem. 259, 112668 (2024).
Krämer, U., Cotter-Howells, J. D., Charnock, J. M., Baker, A. J. M. & Smith, J. A. C. Free histidine as a metal chelator in plants that accumulate nickel. Nature 379, 635–638 (1996).
Ma, Z. et al. Bacillithiol is a major buffer of the labile zinc pool in Bacillus subtilis. Mol. Microbiol. 94, 756–770 (2014).
Murphy, J. T. et al. Histidine protects against zinc and nickel toxicity in Caenorhabditis elegans. PLoS Genet. 7, e1002013 (2011).
Nairn, B. L. et al. The response of Acinetobacter baumannii to zinc starvation. Cell Host Microbe 19, 826–836 (2016).
Morgan, M. T., Nguyen, L. A. H., Hancock, H. L. & Fahrni, C. J. Glutathione limits aquacopper(I) to sub-femtomolar concentrations through cooperative assembly of a tetranuclear cluster. J. Biol. Chem. 292, 21558–21567 (2017).
Brawley, H. N. & Lindahl, P. A. Low-molecular-mass labile metal pools in Escherichia coli: advances using chromatography and mass spectrometry. J. Biol. Inorg. Chem. 26, 479–494 (2021).
Stewart, L. J. et al. Role of glutathione in buffering excess intracellular copper in Streptococcus pyogenes. MBio 11, 02804–02820 (2020).
Hider, R. C. & Kong, X. L. Glutathione: a key component of the cytoplasmic labile iron pool. BioMetals 24, 1179–1187 (2011).
Osman, D. et al. Fine control of metal concentrations is necessary for cells to discern zinc from cobalt. Nat. Commun. 8, 1884 (2017).
Giedroc, D. P. & Arunkumar, A. I. Metal sensor proteins: nature’s metalloregulated allosteric switches. Dalton Trans. 29, 3107–3120 (2007).
Pennella, M. A., Arunkumar, A. I. & Giedroc, D. P. Individual metal ligands play distinct functional roles in the zinc sensor Staphylococcus aureus CzrA. J. Mol. Biol. 356, 1124–1136 (2006).
Aguirre, J. D. & Culotta, V. C. Battles with iron: manganese in oxidative stress protection. J. Biol. Chem. 287, 13541–13548 (2012).
Anjem, A., Varghese, S. & Imlay, J. A. Manganese import is a key element of the OxyR response to hydrogen peroxide in Escherichia coli. Mol. Microbiol. 72, 844–858 (2009).
Tarrant, E. et al. Copper stress in Staphylococcus aureus leads to adaptive changes in central carbon metabolism. Metallomics 11, 183–200 (2019).
Abdul-Tehrani, H. et al. Ferritin mutants of Escherichia coli are iron deficient and growth impaired, and fur mutants are iron deficient. J. Bacteriol. 181, 1415–1428 (1999).
Mills, S. A. & Marletta, M. A. Metal binding characteristics and role of iron oxidation in the ferric uptake regulator from Escherichia coli. Biochemistry 44, 13553–13559 (2005).
Carter, K. P., Young, A. M. & Palmer, A. E. Fluorescent sensors for measuring metal ions in living systems. Chem. Rev. 114, 4564–4601 (2014).
Morgan, M. T. et al. Ratiometric two-photon microscopy reveals attomolar copper buffering in normal and Menkes mutant cells. Proc. Nat. Acad. Sci. 116, 12167–12172 (2019).
Nolan, E. M. & Lippard, S. J. Small-molecule fluorescent sensors for investigating zinc metalloneurochemistry. Acc. Chem. Res. 42, 193–203 (2009).
Zou, Z., Higginson, B. & Ward, T. R. Creation and optimization of artificial metalloenzymes: Harnessing the power of directed evolution and beyond. Chem 10, 2373–2389 (2024).
Harnden, K. A., Wang, Y., Vo, L., Zhao, H. & Lu, Y. Engineering artificial metalloenzymes. Protein Eng., 177–205 (2021).
Helalat, S. H., Jers, C., Bebahani, M., Mohabatkar, H. & Mijakovic, I. Metabolic engineering of Deinococcus radiodurans for pinene production from glycerol. Microb. Cell Fact. 20, 1–14 (2021).
Shomar, H. & Bokinsky, G. Towards a synthetic biology toolset for metallocluster enzymes in biosynthetic pathways: what we know and what we need. Molecules 26, 6930 (2021).
Arnold, F. H. Directed evolution: bringing new chemistry to life. Angew. Chem. Int. Ed. 57, 4143 (2018).
United Kingdom Government, D. f. S. I. a. T. National Vision For Engineering Biology, https://www.gov.uk/government/publications/national-vision-for-engineering-biology (2023).
Alderighi, L. et al. Hyperquad simulation and speciation (HySS): a utility program for the investigation of equilibria involving soluble and partially soluble species. Coord. Chem. Rev. 184, 311–318 (1999).
Smith, R. M. & Martell, A. E. Critical stability constants. 1 (Amino acids, 1974).
Winter, G. xia2: an expert system for macromolecular crystallography data reduction. J. Appl. Crystallogr. 43, 186–190 (2010).
Kabsch, W. xds. Acta Crystallogr. D. Biol. Crystallogr 66, 125–132 (2010).
Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr. D. Biol. Crystallogr 67, 235–242 (2011).
McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674 (2007).
Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D. Biol. Crystallogr 66, 213–221 (2010).
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D. Biol. Crystallogr 66, 486–501 (2010).
Meng, E. C. et al. UCSF ChimeraX: Tools for structure building and analysis. Protein Sci. 32, e4792 (2023).
Pravda, L. et al. MOLEonline: a web-based tool for analyzing channels, tunnels and pores (2018 update). Nucleic Acids Res. 46, W368–W373 (2018).
Osman, D. et al. Generating a metal-responsive transcriptional regulator to test what confers metal sensing in cells. J. Biol. Chem. 290, 19806–19822 (2015).
Foster, A. W., Patterson, C. J., Pernil, R., Hess, C. R. & Robinson, N. J. Cytosolic Ni(II) sensor in cyanobacterium: nickel detection follows nickel affinity across four families of metal sensors. J. Biol. Chem. 287, 12142–12151 (2012).
Kuzmič, P. Program DYNAFIT for the analysis of enzyme kinetic data: application to HIV proteinase. Anal. Biochem. 237, 260–273 (1996).
Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods 25, 402–408 (2001).
Baba, T. et al. Construction of Escherichia coli K‐12 in‐frame, single‐gene knockout mutants: the Keio collection. Mol. Sys. Biol. 2, 2006 (2006).
Acknowledgements
The Diamond Light Source is acknowledged for time on beamline I04 under proposal MX32736, Prof. Timothy Blower, Prof. Liz Morris for assistance in crystallographic data collection and processing. This work was supported by Biotechnology and Biological Sciences Research Council awards BB/W015749/1 (N.J.R.), Understanding mis-metalation of native versus heterologously expressed protein, and BB/V006002/1 (N.J.R.), A calculator for metalation inside a cell, along with BB/S009787/1 (N.J.R.) supporting networking in Industrial Biotechnology. The authors acknowledge the contributions of Deenah Morton (née Osman) to whom the work is dedicated.
Author information
Authors and Affiliations
Contributions
T.R.Y., S.E.C., A.G., E.T., and N.J.R. contributed to in vivo MncA metal-binding experiments. T.R.Y., A.G., S.E.C., and E.T. contributed to in vitro refolding experiments, with T.R.Y. doing foundation work with S.E.C. and A.G. completing the bulk of these experiments. T.R.Y. derived calculations to formulate competing metal buffers and to use MncA as a probe to refine estimates of intracellular metal availabilities and designed related experimental protocols. A.G. generated the crystal structure of NiII-MncA. P.T.C. and A.J.P.S. analysed NiII and DNA binding properties of NiII-RcnR. A.G., S.E.C. and E.T. determined metal contents of E. coli cells. S.E.C. performed and analysed transcript abundance by qPCR. Relationships between DNA occupancies of metal sensors, intracellular metal availabilities, and qPCR data, used methods established by T.R.Y., N.J.R. T.R.Y. developed the metalation calculations in a form used here. N.J.R., S.E.C., with input from A.G. and E.T., wrote a first draft of the manuscript and generated graphics while all authors edited and approved the final version. T.R.Y., then S.E.C., then A.G., sequentially contributed the bulk of laboratory work at different phases of the programme, with T.R.Y. and S.E.C. making equivalently large contributions. N.J.R. with support of all authors, and especially T.R.Y., S.E.C., and A.G., interpreted the significance of the data. N.J.R. had overall responsibility for the design, finance, and management of the project throughout.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Erik Yukl and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Clough, S.E., Young, T.R., Tarrant, E. et al. A metal-trap tests and refines blueprints to engineer cellular protein metalation with different elements. Nat Commun 16, 810 (2025). https://doi.org/10.1038/s41467-025-56199-w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-56199-w
This article is cited by
-
Microbial metal physiology: ions to ecosystems
Nature Reviews Microbiology (2025)