Functional Roles of Chelated Magnesium Ions in RNA Folding and Function

Abstract

RNA regulates myriad cellular events such as transcription, translation, and splicing. To perform these essential functions, RNA often folds into complex tertiary structures in which its negatively charged ribose–phosphate backbone interacts with metal ions. Magnesium, the most abundant divalent metal ion in cells, neutralizes the backbone, thereby playing essential roles in RNA folding and function. This has been known for more than 50 years, and there are now thousands of in vitro studies, most of which have used ≥10 mM free Mg²⁺ ions to achieve optimal RNA folding and function. In the cell, however, concentrations of free Mg²⁺ ions are much lower, with most Mg²⁺ ions chelated by metabolites. In this Perspective, we curate data from a number of sources to provide extensive summaries of cellular concentrations of metabolites that bind Mg²⁺ and to estimate cellular concentrations of metabolite-chelated Mg²⁺ species, in the representative prokaryotic and eukaryotic systems Escherichia coli, Saccharomyces cerevisiae, and iBMK cells. Recent research from our lab and others has uncovered the fact that such weakly chelated Mg²⁺ ions can enhance RNA function, including its thermodynamic stability, chemical stability, and catalysis. We also discuss how metabolite-chelated Mg²⁺ complexes may have played roles in the origins of life. It is clear from this analysis that bound Mg²⁺ should not be simply considered non-RNA-interacting and that future RNA research, as well as protein research, could benefit from considering chelated magnesium.

Graphical Abstrct

Introduction

As a single-stranded polymer with diverse structures, RNA plays critical roles in many aspects of cellular function. For the past two decades, researchers have revealed that RNA has numerous functions beyond serving as the conduit of genetic information from DNA to proteins. For instance, RNA synthesizes proteins, regulates transcription and translation, and perform splicing.¹ Equally important, RNA structures are associated with disease; for example, the COVID-19 pandemic is caused by an RNA virus, and mRNA vaccines are critical tools for fighting emerging viruses like SARS-CoV-2.^2–5 It is thus essential to understand how RNA folds and functions in cells.^6,7

To fold into compact structures, the negatively charged ribose–phosphate backbone needs to be neutralized. Both divalent and monovalent metal cations contribute to neutralization. As a small divalent metal cation, Mg²⁺ can interact strongly with the phosphate backbone, allowing RNA to adopt compact tertiary structures.^8–12 RNA folding and function have been historically examined under non-physiological in vitro conditions, often with ~10 mM free Mg²⁺ and 0.1 – 1 M Na⁺, as thousands of previous studies show (see, for instance, refs 13–23). However, these conditions are not representative of the complex cellular environment, which contains smaller amounts of Na⁺ and higher concentrations of metabolites, proteins, and other RNAs, with very little free Mg²⁺ (Figure 1).^24,25

Figure 1 |. Exposure of RNA to small biological molecules in cells.

(A) Illustration of a cellular environment. (B) Mg²⁺ ions interact with metabolites and bind to RNA such as the HDV ribozyme (Protein Data Bank entry 3NKB). (C) Crystal structure of glutamate–chelated Mg²⁺ on a representative coordinated Mg²⁺ ion.²⁶ Panel A is reproduced with permission from ref 24. Copyright 1969 Elsevier.

In vivo-like conditions that simulate various aspects of the cell have been used to provide insight into how metal ions, metabolites, and crowders might affect RNA folding and function in real cells.²⁵ The concentration of free Mg²⁺ ions, defined as Mg²⁺ ions that are fully coordinated by water molecules²⁷, is much lower in vivo than typically used in vitro; for instance, in vivo concentrations of free Mg²⁺ are estimated to be only 0.5 – 1 mM in eukaryotic cells and 2 – 3 mM in bacterial cells, whereas typical in vitro concentrations of Mg²⁺, all of which are free, have been ~10 mM.^25,28–35 In contrast, the concentration of bound Mg²⁺ ions in vivo is quite high, estimated to be 20 – 80 mM. In other words, in vitro conditions typically overrepresent the concentrations of free Mg²⁺ ions and underrepresent the concentrations of bound Mg²⁺ ions.

This imbalance between free and bound Mg²⁺ ions under traditional in vitro conditions has consequences for studying RNA folding and function. First, the effects of free Mg²⁺ ions are likely exaggerated under traditional in vitro conditions. Free Mg²⁺ ions stabilize RNA structures via diffuse, nonspecific interactions that neutralize the negatively charged RNA backbone.³⁶ Additionally, free Mg²⁺ ions can also shed their hydration shell and form strong interactions with specific sites in RNA, which promote local structural rearrangement.³⁷ Second, the effects of metabolite-chelated Mg²⁺ ions are likely underrepresented by traditional in vitro conditions. Many of the Mg²⁺–metabolite complexes we describe below are weak, potentially allowing the bound Mg²⁺ to act much like free ions. Moreover, weakly bound Mg²⁺ ions may actually offer advantages to RNA folding and stability as compared to free Mg²⁺ ions, as our research suggested. Weakly bound Mg²⁺ ions stabilize RNA folding, protect RNA from chemical degradation, and promote nucleic acid catalysis.^27,38 Herein, we define weakly bound Mg²⁺ ions as those ions having a K_D of > 2 mM, the approximate free concentration of Mg²⁺ in Escherichia coli. To add to the complexity, biological small molecules such as urea and sucrose weakly destabilize RNA secondary structure and increase folding energy, typically by ~0.5 kcal mol⁻¹ m⁻¹.³⁹ Lastly, approximately 30% of the cell volume is occupied by molecular crowding from proteins and nucleic acids, and such crowding is known to enhance RNA folding and function.^{24,25,39–47} In sum, the cellular condition is a complex web of interactions among RNA, crowders, metal ions, and small molecules that themselves sometimes bind Mg²⁺. In an effort to gain insight into the role of biologically chelated magnesium ions in RNA function, we consider the network of interactions among RNA, metal ions, metabolites, and metal ion-metabolite complexes (Figure 1B).

Extensive literature describes absolute concentrations of metabolites in several organisms.^48–51 Amino acids and nucleotides are the most abundant metabolites in the three organisms surveyed: E. coli, Saccharomyces cerevisiae, and BAX^−/−/BAK^−/−-immortalized baby mouse kidney epithelial cells (iBMK) (Figure 2A–C). These metabolites can chelate Mg²⁺ ions via carboxylate or phosphate groups, producing metabolite-chelated Mg²⁺ (Figure 1C).^27,38,52 In the top charts of Figure 2A–C, we used a pH of 7.5 and an ionic strength of 150 mM to estimate the chelated Mg²⁺ (CM) concentrations (see the Supporting Information). Indeed, we recently reported that cellular concentrations of metabolites and Mg²⁺ ions significantly promote RNA folding, stability, and catalysis.^27,38,52 In this Perspective, we discuss the estimated levels of metabolite-chelated Mg²⁺ ions present in various cells, the promotion of RNA function by metabolite-chelated Mg²⁺ ions, and the implications of metabolite-chelated Mg²⁺ ions on the origins of life.^{27,38,52–54}

Figure 2 |. Amino acids and nucleotides comprise the major metabolites across three organisms.

(A – C) Absolute concentrations of the top 30 metabolites (bottom chart in each panel) and estimated concentrations of chelated Mg²⁺ (CM) at pH 7.5 and an ionic strength of 150 mM (top) in (A) E. coli,⁴⁹ (B) S. cerevisiae, and (C) BAX^−/−/BAK^−/−-immortalized baby mouse kidney epithelial cells (iBMK).⁵¹ Metabolite concentrations in panels A – C are summarized in Table 1. (D) Total concentrations of metal ions in E. coli.³⁰ Panel D is reproduced with permission from ref 30. Copyright 2018 Springer Nature.

LEVELS OF METABOLITES AND METAL IONS IN CELLS

Metabolites are small molecules that have essential functions inside cells. Given how diverse life itself is, one question is how diverse metabolite levels are in various organisms. The metabolome analyses of Rabinowitz and co-workers in different cells (E. coli, S. cerevisiae, and iBMK) revealed a striking concordance in absolute metabolite concentrations among these diverse organisms.^49,51 While the metabolomes of these three phylogenetically diverse organisms are similar, one should exercise caution in extrapolating them to other organisms. In an effort to understand the levels of metabolite-chelated Mg²⁺ present in cells, we compiled data from these studies and others, including a compendium of physical constants that describe the interactions between small molecules and metal ions called Critical Stability Constants.^{27,30,49,51,55}

In E. coli, the 30 most abundant metabolites account for ~90% of all metabolites and are mostly amino acids, nucleotides, and their derivatives (Figure 2A). Indeed, approximately 50% of cellular metabolites are amino acids. Notably, the concentration of glutamate is exceptionally high, being 96 mM (Figure 2A). In mammalian and yeast cells, metabolites are sequestered in multiple organelles. Rabinowitz and colleagues provided absolute concentrations for metabolites in the cytoplasm and the mitochondria by constraining total metabolite concentrations by the free energy of formation of a metabolite in different compartments and total metabolite flux in a cell.⁵¹ The concentrations for the cytoplasm were compiled in our study. Even with compartmentalization, trends similar to those of E. coli are observed in S. cerevisiae and iBMK cytoplasms, where the 30 most abundant metabolites account for 95% of the total metabolites and are enriched with amino acids, nucleotides, and their derivatives (Figure 2B, C). Many metabolites can have one or more effects on biopolymers via their intermolecular forces.²⁴ For example, single-molecule FRET experiments and biochemical analyses with several different RNAs revealed that arginine can stabilize RNA secondary structure and destabilize RNA tertiary interactions.^56–58

Understanding the levels of metal ions in cells is motivated by the diverse roles metal ions, in particular Mg²⁺ ions, play in RNA functions. For example, hydrated magnesium hydroxide acts as a general base in the self-cleaving reaction catalyzed by the HDV ribozyme (Figure 3A).⁵⁹ In addition, Mg²⁺ ions coordinate a nonbridging oxygen in the catalytic site of the Tetrahymena ribozyme (Figure 3B), which assists in the formation of the catalytic site and the phosphoryl transfer reaction.⁶⁰

Figure 3 |. Mg²⁺ participates in RNA catalysis.

(A) General acid–base catalysis of the HDV ribozyme. Hydrated Mg²⁺ hydroxide acts as a general base to deprotonate the 2’-OH group of U–1, which results in in-line attack on the scissile phosphate. General acid C75 protonates the leaving group of G1. (B) Three Mg²⁺ ions assist in the formation of the transition state in the catalytic center of the Tetrahymena ribozyme.⁶⁰ Panel B is reproduced with permission from ref 60. Copyright 2001 American Chemical Society.

Moreover, RNA structural analysis by chemical probing and X-ray crystallography reveal that Mg²⁺ is required for folding of RNA tertiary structure.^61–63 In addition to its effect on RNA alone, Mg²⁺ assists in ribonucleoprotein functions. For example, RNase P, which cleaves the 5′-leader sequence of tRNA, uses Mg²⁺ ions to stabilize the structure of the RNaseP–tRNA complex and to promote catalytic activity.⁶⁴ Thus, Mg²⁺ has integral and diverse roles in biological reactions.

The metalome analysis of Outten and co-workers provides the total (free plus bound) concentrations of metal ions in E. coli, where Mg²⁺ is an abundant divalent metal ion and its total concentration is estimated to be an astounding 54 mM (Figure 2D).^30,32 In a different study, Moncany and Kellenberger reported that the cellular concentration of Mg²⁺ in E. coli is 90 – 110 mM, nearly double that of Outten et al., emphasizing the high total Mg²⁺ concentration in vivo and suggesting that the total Mg²⁺ concentration may depend on growth conditions and methods of cell collection, both of which differed in this study.²⁸ van Eunen et al. reported a total concentration of Mg²⁺ in S. cerevisiae as high as 51 mM⁶⁵, implying that the total Mg²⁺ concentration may be in a similar range in E. coli and S. cerevisiae.

CHELATION OF MG²⁺ BY VARIOUS METABOLITES IN CELLS

As mentioned in the previous section, cells contain a total concentration of Mg²⁺ ions ranging from 50 to 100 mM. The fractionation of metabolite-chelated Mg²⁺ complexes and their potential for influence on RNA structure and function are thus of great interest. Tyrrell and co-workers used an intracellular chelator of Mg²⁺ to estimate the free concentrations of Mg²⁺ in E. coli, where they determined it to be only ~1 mM, much lower than the total concentration of Mg²⁺ of 50 – 100 mM.³⁵ This difference implies that a large fraction of Mg²⁺ is chelated by cellular components. Amino acids, the most abundant metabolites, have at least one carboxylate, which can interact weakly with Mg²⁺ (Figure.1C).⁵⁵ Notably, the cellularly abundant amino acids aspartate and glutamate have two carboxylates, which facilitate Mg²⁺ binding at the pH at which the amine group of the amino acid is protonated.^66–68 The identity and estimated concentrations of the chelated Mg²⁺ ions in E. coli, yeast, and the mammalian cells are provided in Figure 2A–C and Table 1; the concentrations of chelated Mg²⁺ ions are calculated from chemical equilibria (Table 1), as described in detail previously^27,38 but taking into account the physiological pH and ionic strength (see the Supporting Information). Briefly, apparent dissociation constants for a metabolite binding to Mg²⁺ ions are affected by both pH and ionic strength. Protonation of metabolites at cellular pH can reduce the fraction of metabolite in a metal ion binding-competent protonation state, and screening via nonspecific ion pairing can reduce the activity of metabolites and Mg²⁺ ions in aqueous solution. The Supporting Information describes methods for accounting for ionic strength and pH and for calculating apparent K_D values, which are in reasonable agreement with isothermal titration experiments performed at physiological pH and ionic strength (Supplementary Table 2). All data and an apparent K_D calculating program are available at https://github.com/JPSieg/MetaboMgITC for reconstitution on any console. Even when the ionic strength and pH are taken into account, a majority of the 30 most abundant metabolites in E. coli, S. cerevisiae, and iBMK have an affinity for Mg^{2+ 55,69} and are estimated to form chelated Mg²⁺ complexes in cells at concentrations ranging from submillimolar to millimolar (Table 1). In E. coli cells, glutamate, glutathione, and nucleotides are the major chelators of Mg²⁺ (Figure 2A–C and Table 1).^27,38

Table 1 |.

Absolute concentrations of metabolites and chelated Mg²⁺.

Species	Metabolites	Cellular Conc. [mM]	KD for Mg2+a [mM]	Chelated Mgb [mM]	Expected Biological Effect (EBE)c (KD × [Chelated Mg])	Chelated Mg Source	Refs	Notes d
E.coli	L-Glutamic acid	96	31.8	5.7	180.6	This article	49,55
E.coli	Glutathione	16.6	21.5	1.4	30.4	This article	49,55
E.coli	Fructose 1,6-BP	15.2	9.2	2.7	25.0	This article	49,55
E.coli	ATP	9.63	0.1	9.2	0.92	ITC in this article	49
E.coli	UDP-GlcNAC	9.24	9.1	1.7	15.2	38	38,49
E.coli	UTP	8.29	0.2	7.6	1.3	This article	49,55
E.coli	Glucose 6-P	7.88	5.1	2.2	11.4	This article	49,55
E.coli	GTP	4.87	0.2	4.5	0.88	This article	49,55
E.coli	dTTP	4.62	0.2	4.3	0.84	This article	49,55
E.coli	L-Aspartic acid	4.23	135.4	0.062	8.3	This article	49,55
E.coli	L-Valine	4.02	5651.0	0.0014	8.0	This article	49,55
E.coli	L-Glutamine	3.81	28154.1	0.00027	7.6	This article	49,55
E.coli	6-P-gluconic acid	3.77	9.0	0.69	6.2	Predicted from chemical structure	49,55	Used P-enolpyruvate
E.coli	Pyruvic acid	3.66	92.0	0.078	7.2	This article	49,55
E.coli	diOH-ACN-P	3.06	22.3	0.25	5.6	Predicted from chemical structure	49,55	Used Glycerol 1-P
E.coli	CTP	2.73	0.2	2.5	0.5	This article	49,55
E.coli	L-Alanine	2.55	2606.4	0.0019	5.1	This article	49,55
E.coli	NAD	2.55	9.1	0.5	4.2	Predicted from chemical structure	49,55	Used UDP-GlcNAC
E.coli	Fructose 6-P	2.52	40.8	0.12	4.8	This article	49,55
E.coli	UDP-glucose	2.5	3.3	0.94	3.1	38	38,49
E.coli	Oxidized glutathione	2.37	21.5	0.20	4.3	Predicted from chemical structure	49,55	Used Glutahione
E.coli	uridine	2.09	NA	NA	NA	–	49
E.coli	Citric acid	1.96	0.6	1.5	0.9	This article	49,55
E.coli	UDP	1.79	1.1	1.1	1.3	This article	49,55
E.coli	(S)-Malic acid	1.68	26.2	0.12	3.1	This article	49,55
E.coli	3-P-D-glyceric acid	1.54	7.8	0.32	2.5	This article	49,55
E.coli	Glyceric acid	1.41	159.9	0.017	2.8	This article	49,55
E.coli	CoA	1.37	9.1	0.25	2.2	Predicted from chemical structure	49,55	Used UDP-GlcNAC
E.coli	Citrulline	1.35	2741.2	0.00098	2.7	This article	49,55
E.coli	L-Threonine	1.26	1384.5	0.0018	2.5	This article	49,55
Total conc. (E.coli)	–	225	–	48	–	This article	–	–
iBMK	L-Glutamic acid	63.8	31.8	1.0	31.4	This article	51,55
iBMK	L-Glutamine	17.2	28154.1	0.0003	8.6	This article	51,55
iBMK	L-Aspartic acid	14.9	135.4	0.055	7.4	This article	51,55
iBMK	UDP-GlcNAC	8.97	9.1	0.47	4.3	38	38,51
iBMK	L-Alanine	6.98	2606.4	0.0013	3.5	This article	51,55
iBMK	L-Threonine	6.69	1384.5	0.0024	3.3	This article	51,55
iBMK	Pyruvic acid	5.88	92.0	0.031	2.9	This article	51,55
iBMK	L-Serine	4.86	8696.5	0.00028	2.4	This article	51,55
iBMK	ATP	4.67	0.1	4.2	0.2	ITC in this article	51
iBMK	L-Glycine	3.71	2606.4	0.00071	1.9	Predicted from chemical structure	51,55	Used Alanine
iBMK	Glutathione	3.09	21.5	0.070	1.5	This article	51,55
iBMK	N-ACN-L-aspartic acid	2.9	135.4	0.011	1.4	This article	51,55	Used Aspartic acid
iBMK	Phenylpyruvic acid	1.77	92.0	0.0096	0.88	Predicted from chemical structure	51,55	Used Pyruvic acid
iBMK	L-Isoleucine	1.76	2358.4	0.00037	0.88	This article	51,55
iBMK	L-Leucine	1.76	2583.3	0.00034	0.99	This article	51,55
iBMK	UTP	1.76	0.2	1.3	0.25	This article	51,55
iBMK	diOH-ACN-P	1.63	22.3	0.036	0.80	Predicted from chemical structure	51,55	Used Glycerol 1-P
iBMK	UDP-glucose	1.53	3.3	0.20	0.66	38	38,51
iBMK	Fructose 1,6-BP	1.52	9.2	0.078	0.72	This article	51,55
iBMK	L-Valine	1.51	5651.0	0.00013	0.75	This article	51,55
iBMK	(S)-Malic acid	1.39	26.2	0.026	0.68	This article	51,55
iBMK	L-Proline	1.23	2606.4	0.00024	0.61	Predicted from chemical structure	51,55	Used Alanine
iBMK	L-Tyrosine	0.938	2608356.6	1.7e-7	0.47	This article	51,55
iBMK	CTP	0.897	0.2	0.64	0.13	This article	51,55
iBMK	L-Phenylalanine	0.84	2606.4	0.00016	0.42	Predicted from chemical structure	51,55	Used Alanine
iBMK	dihydro-orotate	0.735	376.1	0.00098	0.37	Predicted from chemical structure	51,55	Used Benzoic acid
iBMK	GTP	0.677	0.2	0.48	0.097	This article	51,55
iBMK	Glucose 6-P	0.675	5.1	0.060	0.31	This article	51,55
iBMK	L-Methionine	0.639	2606.4	0.00012	0.32	Predicted from chemical structure	51,55	Used Alanine
Total conc. (iBMK)	–	165.7	–	9.5	–	This article	–	–
S. cerevisiae	L-Glutamic acid	39.1	31.8	0.61	19.2	This article	51,55
S. cerevisiae	L-Glutamine	35.5	28154.1	0.00063	17.7	This article	51
S. cerevisiae	Citrulline	27	2741.2	0.0049	13.5	This article	51,55
S. cerevisiae	L-Alanine	22.3	2606.4	0.0043	11.1	This article	51,55
S. cerevisiae	L-Arginine	21.8	2535.0	0.0043	10.9	This article	51,55
S. cerevisiae	Pyruvic acid	9.4	92.0	0.051	4.7	This article	51,55
S. cerevisiae	Trehalose	8.4	NA	NA	NA	–	51
S. cerevisiae	L-Threonine	6.69	1384.5	0.0024	3.3	This article	51,55
S. cerevisiae	L-Aspartic acid	6.29	135.4	0.023	3.1	This article	51,55
S. cerevisiae	L-Asparagine	5.69	14753.3	0.00019	2.8	This article	51,55
S. cerevisiae	Glucose 6-P	5.31	5.1	0.47	2.4	This article	51,55
S. cerevisiae	L-Lysine	5.16	10776896	2.4e-7	2.6	This article	51,55
S. cerevisiae	CoA	4.9	9.1	0.26	2.3	Predicted from chemical structure	51,55	Used UDP-GlcNAC
S. cerevisiae	Ornithine	4.49	10776896	2.0e-7	2.2	Predicted from chemical structure	51,55	Used Lysine
S. cerevisiae	Glutathione	4.3	21.5	0.098	2.1	This article	51,55
S. cerevisiae	Fructose 1,6-BP	4	9.2	0.21	1.9	This article	51,55
S. cerevisiae	L-Serine	3.87	8696.5	0.00022	1.9	This article	51,55
S. cerevisiae	L-Valine	2.5	5651.0	0.00022	1.2	This article	51,55
S. cerevisiae	NAD	2.44	9.1	0.13	1.2	Predicted from chemical structure	51,55	Used UDP-GlcNAC
S. cerevisiae	Fructose 6-P	2.37	40.8	0.029	1.2	This article	51,55
S. cerevisiae	ATP	1.93	0.1	1.6	0.16	ITC in this article	51,55
S. cerevisiae	Citric acid	1.49	0.6	0.68	0.41	This article	51,55
S. cerevisiae	L-Proline	1.36	2606.4	0.00026	0.68	Predicted from chemical structure	51,55	Used Alanine
S. cerevisiae	UDP-GlcNAC	1.02	9.1	0.053	0.48	38	38,51
S. cerevisiae	(S)-Malic acid	0.925	26.2	0.017	0.45	This article	51,55
S. cerevisiae	diOH-ACN-P	0.823	22.3	0.018	0.40	Predicted from chemical structure	51,55	Used Glycerol 1-P
S. cerevisiae	Succinic acid	0.635	77.8	0.0040	0.32	This article	51,55
S. cerevisiae	3-P-D-glyceric acid	0.578	7.8	0.035	0.27	This article	51,55
S. cerevisiae	UTP	0.494	0.2	0.35	0.070	This article	51,55
Total conc. (S. cerevisiae)	–	231.6	–	5.5	–	This article	–	–

^aReported K_D values are apparent dissociation constants at physiological pH (pH 7.5) and an ionic strength of 0.15 M. K_D values were either determined with isothermal titration calorimetry or calculated as described in the Supporting Information.

^b

Concentrations of chelated magnesium ions are estimated as described in ref 27 Briefly, free Mg²⁺ concentrations are 2 mM for E. coli and 0.5 mM for iBMK and S. cerevisiae. Interaction between metabolites and Mg²⁺ in aqueous solution can be represented by the following equilibrium:

$$\text{Metabolite}+{\text{Mg}}^{2+}\rightleftharpoons \text{Metabolite}\cdot {\text{Mg}}^{2+}$$

The binding affinity of metabolite for magnesium can then be represented as

$$K_{\text{D},\text{Metabolite}}=\frac{\left[\text{Metabolite}\right]\left[{\text{Mg}}^{2+}\right]}{\left[\text{Metabolite}\cdot {\text{Mg}}^{2+}\right]}$$

which can be shown to give

$$\left[\text{Metabolite}\cdot {\text{Mg}}^{2+}\right]={\left[\text{Metabolite}\right]}_0\frac{\left[{\text{Mg}}^{2+}\right]}{K_{\text{D},\text{Metabolite}}+\left[{\text{Mg}}^{2+}\right]}$$

where [Metabolite]₀ is the total concentration of metabolite in the cells, [Mg²⁺] is the concentration of free magnesium experienced by all metabolites in a given cell, and [Metabolite·Mg²⁺] is the concentration of the metabolite-magnesium complex.

^cChelated Mg²⁺ species that have large “Expected Biological Effect” (EBE) values are anticipated to have significant interactions with RNA because the Mg²⁺ is weakly chelated and the concentration of the chelated species is high.

^dMg²⁺ binding constants were not available for all metabolites in the data set. Binding constants for the closest chemical species to the metabolite of interest were used to estimate concentrations of chelated Mg²⁺. For example, the UDP-GlcNac constant was used as a stand in for NAD and CoA, as all three metabolites share the same nucleotide diphosphate linked to X (NUDIX) moiety, which is the basis of the NUDIX metabolites interacting with Mg²⁺.

The likelihood of a metabolite-chelated Mg²⁺ ion affecting RNA function is related to the strength of the binding interaction between Mg²⁺ and the metabolite and to abundance. Tightly chelated Mg²⁺ ions are unlikely to have any effect on RNA function. For instance, the apparent K_D values for interaction of Mg²⁺ with EDTA and citrate are ~3 μM and ~0.6 mM, respectively, at pH 7.5 and an ionic strength of 0.15 M.⁵⁵ Indeed, we showed that EDTA-chelated Mg²⁺, with six sites of chelation, has little effect on RNA thermostability and ribozyme activity.^27,54 Intermediate-strength chelation of Mg²⁺ with three or four sites of chelation, such as in citrate, also has little effect on RNA thermostability.²⁷ In contrast, weak chelation of Mg²⁺ ions with ligands such as glutamate or malate leads to a gain of RNA function.²⁷

As a semi-quantitative guide for the contribution of chelated Mg²⁺ to the promotion of RNA function, we provide a statistic we call the “expected biological effect (EBE)” in Table 1, which is the product of the dissociation constant and chelated species abundance (K_D [chelated Mg²⁺]). The EBE reflects the requirement that chelated Mg²⁺ be both weakly chelated and abundant to participate in interactions with RNA in biology. For example, if the K_D is too low, the Mg²⁺ will be too tightly sequestered to interact with the RNA, even if the concentration of the metabolite–Mg²⁺ complex is high. The low K_D will be reflected in a low EBE score. Likewise, if the metabolite is not abundant, there will be little metabolite-Mg²⁺ complex to interact with the RNA. This too will be reflected by a low EBE score. However, metabolites with a relatively high abundance in cells and a relatively high K_D (weakly chelated), will produce high concentrations of the metabolite-Mg²⁺ complex that can interact with RNA, resulting in a high EBE score. At the top of the EBE ranking in E. coli is L-glutamic acid (EBE of 181), followed by glutathione (EBE of 30), fructose 1,6-bisphosphate (EBE of 25), UDP-GlcNAc (EBE of 15), and glucose-6-phosphate (EBE of 11) (Table 1). The presence of amino acids glutamate (EBE of 181), aspartate (EBE of 8.3), valine (EBE of 8.0), and glutamine (EBE of 7.6) (total EBE of 205), along with the diphosphates UDP-GlcNAC (EBE of 15.2), NAD (EBE of 4.2), UDP-glucose (EBE of 3.1) and UDP (EBE of 1.3) (total EBE of 24), is consistent with our previous studies.^27,38 Additionally, it is possible that Mg²⁺ ions are weakly chelated by clusters of acidic amino acids on the surface of proteins^54,70, further increasing the concentration of weakly chelated Mg²⁺ in the cellular environment.

MECHANISTIC INSIGHTS INTO THE EFFECTS OF MG²⁺ CHELATION ON RNA: STABILIZATION, PROTECTION, AND PROMOTION OF RNA CATALYSIS

We recently investigated the effects of weakly chelated Mg²⁺ ions on RNA function. We focused our efforts on glutamate-chelated ions but also looked at the behavior of four of the top amino acids, as well as nucleotide diphosphates.^27,38 Our research revealed three effects of metabolite-chelated Mg²⁺ on RNA function (Figure 4):^27,38,47,52: (1) stabilization of RNA folding, (2) protection of RNA from degradation, and (3) promotion of RNA catalysis.

Figure 4 |. Chelated Mg²⁺ ions stabilize and protect RNA structure, and promote RNA catalysis.

(A) Amino acid- and nucleotide diphosphate-chelated Mg²⁺ ions have several favorable effects on RNA functions. (B) Chelated Mg²⁺ ions stabilize HDV-like drz-spur-3 ribozyme structure and increase folding cooperativity, where the red and blue lines are the chelated Mg²⁺ conditions (glutamate-chelated Mg²⁺ and amino acid-chelated Mg²⁺, respectively) and the gray line is the control Mg²⁺ condition. (C) The chelated Mg²⁺ protects HDV-like drz-spur-3 ribozyme from heat-mediated degradation. Successive thermal denaturation experiments show RNA degradation: (filled circles) first denaturation experiment and (empty circles) second denaturation experiment. The stronger transition peaks are observed in the second experiment in the presence of the chelated Mg²⁺ ions because of the RNA protection by the chelated Mg²⁺. “GluCM” represents glutamate-chelated Mg²⁺ and “Aa₄CM” represents amino acid-chelated Mg²⁺. Aa₄CM is a mixture of 96 mM glutamate, 4.2 mM aspartate, 3.8 mM glutamine, and 2.6 mM alanine. (D) Chelated Mg²⁺ increases the activity in the hammerhead ribozyme, HH16 (i.e., shorter reaction times). Red-filled circles show 1.63 mM NDPCM and 0.5 mM Mg²⁺_free, and blue-filled circles show 0.5 mM Mg²⁺total as the control. The SAXS experiment revealed that GluCM promotes HDV-like CPEB3 ribozyme compaction. The bead model represents the SAXS data, while structures are static placeholders to guide the eye. R_g is the radius of gyration, and D_max is the maximum linear dimension. Original data were obtained from our papers.^27,38

Absolute binding constants from Critical Stability Constants were used in our original investigations to calculate experimental CM conditions, without accounting for ionic strength or pH.^27,38 Our updated Mg²⁺ binding constant calculations and isothermal titration calorimetry experiments account for physiological pH and ionic strength, which weaken the interactions of Mg²⁺ with metabolites (Supplementary table 2), meaning that the actual free Mg²⁺ concentrations will be higher and the CM concentrations lower than previously reported. For nucleotide phosphates, the change in the level of free Mg²⁺ is very small, <0.1 mM or a 1.2-fold increase. For amino acids, the increase in the level of free Mg²⁺ is more significant. For example, 2 mM free magnesium with 11.3 mM glutamate CM, for a total of 13.3 mM total Mg²⁺ was reported using an uncorrected K_D of 15.1 mM.²⁷ The updated concentrations using a K_D of 32 mM, adjusted for pH 7.5 and an ionic strength of 0.1 M, are 3.6 mM free magnesium with 9.7 mM glutamate CM, a 1.8-fold increase in the level of free Mg²⁺. Further interpretation of the estimated amounts of free and chelated Mg²⁺ concentrations for these experiments, taking into account physiological pH and ionic strength, is provided in Supplementary Table 4. Overall, the increase in free Mg²⁺ concentrations when accounting for ionic strength and pH is relatively small, less than 2.1-fold for amino acids and 1.2-fold for nucleotide phosphates.

Stabilization of RNA Folding.

Biological small molecules generally destabilize nucleic acid structure in the absence of chelated Mg²⁺. For example, Lambert and Draper reported that osmolytes have a destabilizing effect on RNA structure via nonspecific interactions with nucleobases, which results in the accumulation of osmolytes at the nucleobase.^39,71 In contrast, metabolite-chelated Mg²⁺ ions stabilize RNA structure. For example, in the presence of glutamate-chelated Mg²⁺ and top four amino acid-chelated Mg²⁺, the melting temperature of the HDV-like drz-spur-3 ribozyme is increased by ~5 °C (Figure 4B), the folding free energy (ΔΔG) decreases by ~7 kcal/mol, and the folding enthalpy decreases by 7–20 kcal/mol. These observations indicate that metabolite-chelated Mg²⁺ ions strongly increase RNA stability and folding cooperativity.²⁷ Moreover, we observed the same trend in a similar experiment using tRNA.^47,52 On the basis of our in-line probing analysis of RNA structure, we hypothesize that the stabilizing effects of amino acid-chelated Mg²⁺, which weakens electrostatic repulsion in RNA structure, outcompete the destabilizing effects of the free amino acid, contributing to a net increase in RNA stability.²⁷ We observed similar effects across HDV-like, hammerhead, and glmS ribozymes, as well as tRNA, suggesting that the effects are likely general.

Protection of RNA from Degradation.

RNA is a chemically unstable molecule because the nucleophilic 2′-OH group on the ribose sugar can initiate self-cleavage with a metal ion that facilitates water deprotonation (Figure 3A).⁷² We found that amino acid-chelated Mg²⁺ ions significantly inhibit the background transesterification reaction, especially a mixture of four abundant amino acids (Figure 4C). The mechanism of the protection effect is not totally clear but it is possible that weakly coordinated Mg²⁺ ions are less likely to have a bound hydroxide ion due to repulsion with the bound anionic ligand(s). Our in-line probing experiments support this hypothesis; Mg²⁺ chelation protects the RNA from metal ion-dependent degradation.²⁷

Promotion of RNA Catalysis.

Our kinetics experiments on several self-cleaving nucleic acid enzymes including an HDV-like ribozyme, hammerhead ribozyme, glmS ribozyme, and DNAzyme, showed that glutamate-chelated Mg²⁺ and NDP-chelated Mg²⁺ species enhance ribozyme catalysis as compared with the control conditions with no chelated Mg²⁺ ions (Figure 4D).^27,38 With respect to nucleotides, tightly chelated Mg²⁺ ions such as ATP–Mg²⁺ (K_D ~10 μM), do not stimulate RNA catalysis.³⁸ Our biochemical and structural analyses led us to a model of the facilitation effect where metabolite-chelated Mg²⁺ ions do not participate directly in the catalytic mechanism but do so indirectly by increasing RNA compaction (Figure 4D, right panel).^{27,40,41,44,45} Notably, molecular crowders increase the fraction of the active ribozyme even in the presence of glutamate-chelated Mg²⁺,²⁷ suggesting that molecular crowders and chelated Mg²⁺ ions may have synergetic effects on RNA catalysis.

IMPLICATIONS OF CHELATED MG²⁺ IONS IN THE ORIGINS OF LIFE

Because it can both contain genetic information and catalyze reactions, RNA may have played important roles in the origins of life.⁷³ The three effects of metabolite-chelated Mg²⁺ on RNA function, elaborated in the previous section, may relate to the origins of life. In the RNA world, RNA must be protected from hydrolysis. As described, weakly chelated Mg²⁺ can participate in such RNA protection and extend the half-life of the RNA. Indeed, some of the weakly Mg²⁺-coordinating metabolites found in extant cells are potentially important in the origins of life.^74–76 For example, amino acids including glutamate are known to be synthesized under various simulated early Earth conditions.^77–79 Such protection could have been critical to the prebiotic accumulation of RNA and thus to the start and evolution of life.

Environmental pH is critical in evaluating conditions for the origins of life. Life as we know it requires hydrogen bonding, which requires a local pH range from ~5 to 9 so as to not protonate hydrogen bond acceptors and not deprotonate hydrogen bond donors. Indeed, modern acidophiles and alkaliphiles maintain their cytoplasmic pH in the range of 6–9.⁸⁰ We thus expect life began with an internal pH also near neutrality. The pK_a of glutamate, the predominate Mg²⁺ chelator in modern cells, and indeed of any other carboxylic acid, is near 4 and so would be largely in the functional deprotonated state over this pH range, suggesting that the Mg²⁺ chelation activity described herein is robust with respect to carboxylates. We previously demonstrated that diphosphates weakly chelate Mg²⁺ and favor RNA and DNA catalysis, which was performed at pH 7.5³⁸. However, at pH 6, a value still compatible with hydrogen bonding and thus life, the terminal phosphate of nucleotides would protonate. The diphosphate nucleotides would then shift from an overall charge of −3 to −2 and bind Mg²⁺ more weakly; at the same time, the triphosphates would shift from an overall charge of −4 to −3 and thus bind Mg²⁺ as weakly as diphosphate at higher pH. It is thus possible that NTP-chelated Mg²⁺ could take over the role of favoring RNA function. Indeed, the concentrations of NTPs are considerably higher than those of NDPs, at least in extant life, potentially leading to higher EBE values.

In addition, we and others have reported that compartmentalization by coacervates increases the local concentrations of RNA and metal ions, which enhances ribozyme folding and activity under otherwise suboptimal enzyme and ion conditions.^81–83 Weak chelation could help preserve the stability of the RNA phosphodiester backbone and enhance RNA function under such enriched conditions.

The high concentrations of free Mg²⁺ that are required for RNA function destabilize compartments created by fatty acid vesicles.^53,84,85 Adamala and Szostak revealed that Mg²⁺ chelation by citrate stabilizes the vesicles (Figure 5).⁵³ Likewise, in collaboration with the Keating lab, we showed that Mg²⁺-mediated loss of lipid vesicle membranes can be inhibited by the addition of a chelator.⁸⁶ Adamala and Szostak tested non-enzymatic template-directed RNA primer extension, in which citrate has a slight negative effect on the reaction as compared with the no chelator condition. This is likely because citrate has multiple inner coordination sites on Mg²⁺ resulting in a relatively low K_D value for Mg²⁺.²⁷ In the early Earth environment, weakly metabolite-chelated Mg²⁺ ions could have had a similar stabilization effect on lipid compartments while still promoting RNA catalysis and evolution inside primitive lipid compartments. As such, weak chelation could have been essential in the start of life.

Figure 5 |. Chelated Mg²⁺ stabilizes fatty acid vesicles.

Fatty acids form vesicles that encapsulate RNA inside the vesicles. Unchelated Mg²⁺ ions interact with fatty acids, which precipitates fatty acids. Such destabilized vesicles leak RNAs and small molecules, which leads to RNA unfolding and degradation, as well as loss of genetic information. On one hand, citrate strongly chelates Mg²⁺ ions and thereby stabilizes the fatty acid membranes and prevents RNA leakage, but on the other hand, it limits RNA folding. Glutamate-chelated Mg²⁺ might limit RNA hydrolysis in the vesicles, thus more fully promoting RNA function.

CONCLUSIONS

In cells, the vast majority of Mg²⁺ is chelated; however, some of these species can still promote RNA folding and function. The total concentration of metabolite-chelated Mg²⁺ ions reaches nearly 51 mM in E. coli cells, wherein 20% of ~240 mM total metabolites chelate Mg²⁺. The ability of such Mg²⁺-chelated species to support RNA function depends on both their concentrations being high and their Mg²⁺ affinities being low. In this Perspective, we focused on how the cytoplasm is a complicated milieu in which thousands of biological molecules are co-localized in the same compartment and how this messy environment generally promotes RNA function. We reported three surprising effects of amino acid-chelated Mg²⁺ and nucleotide diphosphate-chelated Mg²⁺ ions on RNA function, where the chelated Mg²⁺ stabilizes RNA structure, protects RNA from degradation, and enhances RNA catalysis. We also found that weakly chelated Mg²⁺ species (K_D ≥ 2 mM) can accumulate to relatively high values: ~18 mM in E. coli and ~2 mM in mammalian and yeast cells. Given the appreciable concentrations of weakly bound Mg²⁺ species as well as the protection they offer RNA from degradation in vivo, these conditions may promote RNA function as well as, or even better than, typical in vitro conditions, especially in E. coli. In many cases, detailed mechanistic insights into how metabolite-chelated Mg²⁺ ions promote RNA function are still needed. Recent advances in the chemical probing of RNA structure at the single-nucleotide level combined with a massively parallel sequencing technique (such as Structure-seq, SHAPE-MaP, and ILP in protocells)^87–89 and metal ion-specific dyes³⁵ will help address these questions. We suggest that the RNA field consider using weakly chelated Mg²⁺ ion conditions in their experiments to better mimic the cell and to promote RNA function. Finally, weakly chelated Mg²⁺ may influence the folding and function of proteins, especially in the hundreds of enzymes that use Mg²⁺ ions in their active sites.^90–92