Why Magnesium?

For the last few weeks I've been mulling over a particular question. It's been a puzzle, like a mental itch, and I've desperately wanted to figure it out. The last few days I finally got the time, so here’s the big question:

Why is magnesium used for its particular purposes in biology?

Cells use chemistry in different ways. Some are obvious, others not so much. Something that's been quite fascinating to me has been bridging the gap between the properties of a material's chemistry/physics, and its use within organisms. There's much more depth to this than one might expect. What I’m curious about is what makes magnesium distinct from other elements? What properties does it have that have caused it to stand out so much in the hierarchy of essential nutrients?

Before I continue, it’s important to clarify that in this case we’ll be discussing the properties of elements rather than molecules. In the past I’ve spent time discussing various molecular processes like methylation and acetylation, and the presence of such groups or structural motifs has a significant effect on how biology functions. Another similar example of this would be the alternating double and single bond structure of DHA, which allows it to act as a semiconductor.

When it comes to elements, or individual types of atoms, the metrics we’re discussing are very different. Rather than focusing on the influence of a double bond here, or a methyl group there, we’ll instead be looking at more fundamental properties like electronegativity, acidity, or atomic number. Let’s briefly explore the periodic table to get a flavor for these types of effects.

Introduction to elemental chemistry

If you haven’t taken a chemistry class in some time, the periodic table might seem off-putting, but for our purposes here it acts as an incredibly useful roadmap. The table is divided up into rows called groups, and columns called periods. The way the table was first organized was based on the number of electrons each element possessed in its valence shell, the outermost orbital of electrons an atom contains.

Each atom is composed of a nucleus, and an orbit of electrons. The vast majority of its mass comes from only a tiny portion of its volume, with the nucleus famously being compared to the size of the head of a pin sitting in the center of a cathedral. This should give you some idea of just how much of the atom is made up of empty space. The limits of the atom, or its atomic radius, are determined by its electrons which surround the nucleus. In our previous analogy the electrons are even more miniscule, something like motes of dust floating around the church.

The standard model of the atom, known as the Bohr model, loosely conceptualizes atoms as tiny solar systems, with the nucleus being like a star that electrons rotate around. This is where we get terminology like “electron orbitals” and “electron spin,” but the reality is that we’re using familiar words to describe something rather abstract.

Atoms are really bundles of what can best be described as particles with wave-like properties. The nucleus gets its mass from two types of particles, protons and neutrons, with only protons providing charge. The electrostatic attraction between the opposing positive charges of the proton with the electrons, which are negatively charged, is one of the fundamental forces holding the atom together.

Many of the numbers on the periodic table are tied to these basic particles. The atomic number, and more importantly what type of element we’re dealing with, is determined entirely by the number of protons the nucleus contains. For example, if carbon loses a proton, it’s no longer carbon, instead it’s been turned into boron which has an atomic number of 5. Thankfully the strong force, the force which holds together protons and neutrons in the nucleus, lives up to its name, and such “nuclear” reactions rarely occur in the elements life deals in.

As such, most of the properties of any particular element are determined more by its electrons than its protons. The protons set a baseline, the unchanging positive charge it contains, which the electrons are then able to structure around. One particularly “wave-like” property of electrons is that they exist ambiently. Rather than moving around the nucleus in a circular motion as is often depicted in movies, electrons are actually vibrations within a particular area that only behave as particles during measurement and certain interactions.

From our point of view, if we try to measure or pinpoint where an electron is, we get areas of probability around the nucleus that tell us where it’s most likely to find the electron. These areas are what we call its electron orbitals, and while they come in many forms, the first few look something like this:

These are the electron orbitals for the first and second “valence shell.” The valence shell is the outermost orbital of electrons, and it’s responsible for defining the radius or edge of the atom, as well as its binding properties.

The first layer, called the s-orbital, can only contain two electrons, while the second layer can contain 8 electrons. Each layer after the first has an additional s-orbital that must be filled before electrons can spill over into higher orbitals, so while the element neon has 10 electrons, it actually has two s-orbitals. If we filled it from 0 to 10 electrons, it would first fill 1s, then 2s, then 2p. As such, the p-orbital only contains 6 electrons, with its orbital clouds pointing out in six directions.

This means if we wrote each orbital out with each row being a valence shell, the orbitals would be filled in a diagonal pattern, adding one electron each time we took a step up in atomic number. Carbon has 6 electrons, and would fill the 1s and 2s orbitals, with two electrons in the 2p orbital. The chart below is often used to help illustrate this pattern for faster calculations:

Atoms reach a sort of equilibrium when they have a full valence shell of electrons, and this is what determines their chemistry. For example, sodium and potassium have only one valence electron each, in their 3s or 4s orbitals respectively. They easily give up this extra electron creating a cation, or positively charged form of the atom with a +1 charge.

Carbon on the other hand has 4/8 of its valence shell filled, so it more easily forms covalent bonds, where it shares 4 electrons of its own in exchange for 4 from other atoms to fill its valence shell. The fact that carbon can form four bonds is also why it’s used as a scaffolding atom in organic chemistry, with other elements like oxygen and hydrogen branching off from mostly carbon base structures. You can see this in methyl groups, which use 3 hydrogens to fill most of the carbon's bonds, and a fourth bond to attach it to various structures.

Another interesting trick is that we can correlate elements by group using this property. For example, elements in group 18 are profoundly unreactive, since they each already have a full valence shell. Group 1 will always tend to form a +1 cation, while group 2 will tend to form a divalent, or +2 cation. Group 14 elements like carbon and silicon can form four covalent bonds and make good scaffolding elements as a result, both in biology and crystal/computer chemistry. If we move further into elements like sulfur, oxygen, and chloride, they instead prefer to gain electrons to create negative ions or anions with -1 or -2 charges.

There are a few more properties which are relevant here. Atomic radius increases each time an electron shell is added, but decreases as the atomic number goes up within a given shell, since the electrostatic attraction holds the electrons closer to the nucleus. This means that the atomic radius increases as we move down and left on the periodic table. There’s also what’s known as the ionic radius, which is slightly different and influenced more by atomic number since different ions can form with the same number of electrons in their valence shell.

Ionization energy is the energy needed to kick an electron out of the atom’s valence shell creating an ion. This can occur as a result of extreme temperatures, such as in the sun, or due to ionizing radiation like UV or X-rays which possess short enough wavelengths to displace electrons. Ionization energy increases, meaning it is more difficult to ionize the atom, as we move up and right on the periodic table, so atoms like fluorine, helium, and neon, are the most difficult to ionize. Others on the left side ionize more easily, like sodium, potassium, calcium, and magnesium.

Electron affinity is a term used to describe the opposite effect, and describes how likely an atom is to gain an electron. Electron affinity increases as we move up and right on the periodic table, with the notable exception of all the elements in group 18, the noble gasses, which already have full valence shells. Electronegativity is a similar term that correlates closely with electron affinity. It refers to how tightly electrons are bound and attracted to an element, and increases as we move up and right as well.

The high electronegativity of oxygen is why oxygen is used as the terminal electron acceptor in chloroplasts and mitochondria, either in the form of NADP+ or oxygen gas. Each successive electron accepting protein in the electron transport chain is more electronegative, acting similar to a semiconductor junction which only allows electrons to move forward but not backward. However, this is also what makes reactive oxygen species, or oxygen-containing free radicals, so reactive, since they have a powerful attraction to electrons and can easily oxidize other nearby molecules. Now with these properties in mind, let’s move on to the interesting stuff.

The elements of life

I’ve tried to keep things engaging so far by pointing out a few interesting connections to biology along the way. Since the way life works is my primary interest, I find this stuff particularly fascinating. Another one of my personal favorite examples is the chemistry of the transition metals, especially those found in period 4 of the periodic table, which are found most abundantly in nature.

The transition metals are so named because of their ability to transition more readily between different electrons states. As we move up in orbital number, the size of the orbitals increases as well, and this gives us greater flexibility in electron configuration. This is quite pronounced in the 3d orbital, so that some transition metals can form a wide variety of cations and/or anions.

Iron and copper are two of the most notable examples of this in biology. Iron can lose several electrons from its d-orbital to form ferrous iron, a soluble ion with +2 charge, or ferric iron, an insoluble ion with +3 charge. Copper can behave similarly, however its higher atomic number means it will only readily give up 1 or 2 electrons in its d-orbital, so it has a +1 cuprous form and a +2 cupric form instead. Cells have made good use of these forms in various redox reactions.

In the electron transport chain, if we zoom in and look for how electrons actually travel through cytochromes I-IV, it turns out that in many cases the electron acceptors are actually heme units (a type of protein ring structure with an iron core), or copper atoms. The heme and copper can move up or down one charge (i.e. between +2 and +3 for iron) as they gain and give up electrons. This flexibility in their valence shell means many electrons can be passed through a single metal ion, without the ion being “used up” in the process.

Iron is also frequently converted back and forth between its ferric and ferrous forms. This is done primarily when iron needs to be transported within the body, since in its ferrous form it has a +2 charge which matches the -2 charge of an oxygen anion, making it strongly reactive with oxygen in the body. This is good in the case of proteins like hemoglobin or myoglobin, which use ferrous iron to bind oxygen, but it also means iron can catalyze the formation of harmful oxygen radicals via what’s called a Fenton reaction.

Copper acts as a cofactor for a class of enzymes called ferroxidases, which act as an “electron switch” for iron, donating or accepting electrons from it to swap between the ferric and ferrous forms. The cuprous form can give up an electron to convert ferric iron into ferrous iron, and vice versa. The vast majority of copper in the bloodstream is found in one such enzyme, ceruloplasmin, and this is useful since all iron must be in the ferric form to be transported safely. I’ve discussed this interplay in more detail here for those interested.

While I find the properties of transition metals interesting, as I said recently I’ve been drawn to a very different question. Elements from groups 1 and 2 are even more fundamental to biology, and they’re present in far greater quantities than metals like iron, but are far less reactive. When calcium or magnesium form an ion, they’re only really stable in the +2 divalent cation configuration, so what makes them useful to life? Obviously there’s more at play here than just the flexibility of a valence electron shell, but what?

To explore this idea, let’s create a hypothetical organism. We’ll give it some flexibility, maybe a few transition metals like iron and copper, and let’s say we give it one of each charge of the main ions as well. So it can use one +1 ion, like sodium, one +2 ion, like calcium, iron, copper, and that’s it.

The question is, could our hypothetical organism fulfill most of the same functions using only one set of ions, or is diversity important? What makes magnesium and these other elements so special beyond the number of electrons in their valence shell? To answer this, let’s explore the chemistry of magnesium in more detail, and see if calcium could be used to fill the same role.

Magnesium chemistry

Magnesium is found in group 2 of the periodic table, and its atomic number is 12. Because of this, it readily loses two electrons, dropping down a level so that only its n=2 valence shell is filled. Because of the ease of this drop, it tends to ionize fairly easily, and it tends to be stable in this state, so most of the magnesium in cells is found in the form of Mg2+ ions.

Magnesium is essential for life. It’s one of the top ten most abundant elements in biology, and is actually the fourth most abundant metal ion in cells by moles. It makes up around 0.1% of the human body by mass, and plays an instrumental role in its function. It’s estimated that the average human has around 25g of magnesium within their body at any given time, with a large percentage stored in bone.

Where magnesium shines most clearly is in its role as a cofactor for various enzymes. A common estimate that gets thrown around says that 300 enzymes in the body require magnesium, but this is laughably low, and only represents a fraction of its total uses. The reality is that it’s required for thousands of processes, including every time ATP is used within the body. The highest estimate I’ve seen is that around 40% of all enzymes in the body require magnesium directly in some form, and magnesium is the most abundant metal built into enzyme molecules. This makes it particularly striking, as it seems to straddle the line between the ions used in large amounts like sodium, and the ions used in small amounts like iron, with emphasis on both utility and quantity.

It turns out that there are several aspects of magnesium that make it useful, including its interaction with phosphate groups, enzymes, lipid membranes, and more. Let’s begin with its interaction with ATP, since this molecule is both the cornerstone and energetic currency of life.

Magnesium and ATP

The universal principle through which life functions is the creation of a charge gradient across a membrane. The cell’s processes revolve around maintaining this gradient, and in the case of our own mitochondria and chloroplasts in plants, this is done by constantly moving protons into a narrow intermembrane space to create a tightly packed build up of free positive charge. The dissipation of this charge is then coupled to various forms of work. This is seen throughout life, including in Prokaryotes and Archaea, though sometimes with slight variation. (For example, some Archaea use sodium ions instead of protons.)

ATP is also a universal principle. It allows cells to couple the far-from-equilibrium state of a charge gradient built up across a membrane to a far-from-equilibrium molecule as well. The key is in ATP’s structure. It consists of an adenosine molecule with a chain of phosphate groups added to it. Depending on the number of phosphate groups, it’s called AMP, ADP or ATP, with ATP being considered the active form.

ATPase enzymes are rotary proteins embedded in the inner membrane that spin as protons flow through them, similar to a water wheel in a hydroelectric dam. ATPases convert AMP and ADP into ATP by adding phosphate groups to it. It’s precisely this process of adding phosphates that leads to the anti-entropy state of ATP, and allows the cell to harness the charge gradient in chemical form.

The reason for this is that phosphate ions are very negatively charged, since they’re composed of one phosphorus atom and four oxygen atoms. If we check phosphorus’ group on the periodic table, we see that it’s in group 15, so phosphorus should only logically be able to form 3 covalent bonds similar to nitrogen. Strangely though, in phosphate ions it actually forms 5 covalent bonds! This seems to break the laws of chemistry, but can be explained by a peculiar quirk of the third valence shell.

Phosphorus normally contains a filled 3s orbital and half-filled 3p orbital, so by gaining three electrons it should be able to fill its 3p orbital. However, phosphorus can also excite the electrons from its 3s orbital to shift them up into its 3d orbital instead, allowing it to use the 3s orbital for two additional electrons, so that it can form five covalent bonds. The phosphate ion is a rare example where this state is stable long term.

This also makes phosphate ions especially packed with negatively charged oxygen atoms surrounding the phosphorus core, since oxygen is more electronegative and will tend to attract the electrons closer to it than the center. Since negative charges repel, this makes the initial jump of linking two phosphate groups together a reaction that rarely occurs spontaneously, so it requires an energetic catalyst like an ATPase.

This is where magnesium comes in. The equilibrium that we’re far from in this case could be viewed as the ratio between ATP and ADP. If we were to dissolve ATP in solution, it would naturally break down to ADP until it returns to equilibrium, but it’s kept at a far higher ATP level in living cells. Magnesium is what allows the cell to do this.

I’d compare this to trying to push a set of weak magnets together so that their repelling sides are all touching. It could be done temporarily, but if you wanted to keep them in this setup it would be far easier to use tape wrapped around them to hold them in place. In this case, magnesium acts as a positively charged “tape” that offsets the negative phosphate chain and holds the ATP molecule together. ATP does not exist in stable form in cells unless it’s complexed with a cation, and it uses magnesium for this purpose.

But why magnesium? Why not calcium? As it turns out, there’s a good reason for this. Magnesium occupies a unique position among the commonly available +1 and +2 ions used by cells. If we look at groups 1 and 2, magnesium is the farthest up and right out of calcium, magnesium, sodium, and potassium.

This means that out of these four, magnesium has both the highest electron affinity and electronegativity, so it tends to strongly attract and hold electrons. It also has the lowest atomic radius, while sharing the same +2 charge as calcium. It’s precisely these attributes that make it so useful. Magnesium’s higher charge/radius ratio favors stronger covalent and ionic interactions at closer ranges.

In the case of ATP, this allows magnesium to stabilize the phosphate chain far better than calcium, despite both having the same charge. I’ve heard a few people suggest that when calcium levels spike in the cell, it could begin to bind to ATP as well. Calcium ions and stimulation/stress are intimately linked, so in theory in cases of calcium influx this could destabilize ATP. That said, I’m skeptical that this occurs much, since it seems more likely any calcium that did bind with ATP would be easily replaced by the better suited magnesium ions. Magnesium deficiency could make the swap more likely.

I did find this paper, which looked at how lithium might interact with Mg-ATP complexes, and found that smaller +1 ions might change its structure overall, possibly filling the third phosphate’s binding site. The researchers speculated that one of lithium’s actions as a drug might be in stabilizing ATP. This paper suggests calcium can bind to ATP, though more weakly than magnesium, based on in situ results. They also suggest the interesting idea that beryllium’s toxicity may be due to its ability to displace magnesium from ATP, since it has an even greater charge/radius ratio.

Origin of life researchers have also theorized that there may have been an early precursor to ATP known as pyrophosphate, a “double phosphate” group in chemistry, that occurred in deep hydrothermal sea vents. At life’s inception, pyrophosphate/phosphate balance might have served a similar purpose to ATP/ADP balance today. This has been suggested as a component of the “iron-sulfur world hypothesis,” put forward by Günter Wächtershäuser.

In this hypothesis, the creation of pyrophosphate may have been naturally catalyzed in hydrothermal vents, along with similar products such as acetyl-phosphate, through redox reactions on surface iron-containing crystals like mackinawite, greigite, and pyrite. Magnesium would have also been available in the presence of similar minerals like brucite as well, suggesting a link between phosphate chain stabilization and magnesium at the origin of life itself.

The stabilization of phosphate chains by magnesium isn’t just limited to ATP either. It occurs in a wide variety of proteins and molecules, in most cases where multiple phosphate groups are bound together. This also includes other nucleosides like guanosine triphosphate, and even the enigmatic thiamine triphosphate, which I proposed in a previous article might act as a phosphate/energy reserve, similar to the function of creatine phosphate.

Magnesium as a catalyst

Magnesium also plays a strong and significant role in enzymes involved in DNA regulation and other reactions. In some cases it acts as a stabilizing agent, similar to its role in maintaining ATP’s structure. However, it can also serve as a direct catalyst in different ways.

One property that we haven’t discussed yet is magnesium’s acidity. You may recall that acidity is usually used as a marker of proton concentration, with more protons making an environment more acidic. Beyond this, there’s also another type of acidity, which applies better to individual metal ions, known as Lewis acidity.

Rather than relying on increases or decreases in proton concentration, Lewis instead measures the inverse effect, how “electron-withdrawing” a material is. Both Lewis acids and standard acids will produce positive charge, but one may do so via releasing protons, while the other removes electrons. Interestingly, magnesium isn’t a strong Lewis acid, but it is still acidic enough for its presence to lower the energetic barrier for certain reactions dramatically.

Magnesium’s relationship with water in such enzymes is particularly crucial. Compared to other ions like calcium, magnesium is especially “water loving.” Since water is a molecular dipole, its hydrogen atoms are more positively charged while its oxygen atom is more negative. Because of this, water tends to form clusters around dissolved ions, known as a hydration shell. Magnesium forms a notably large and stable hydration shell compared to other ions, and this is responsible for many of its effects.

In many cases, the enzymes which magnesium is built into are actually relying more on the water molecules that it provides than on the mineral itself. Through a combination of its Lewis acidity and its association with water, magnesium can speed up “hydrolysis” reactions. These are reactions in which water is used to help break apart chemical bonds. Magnesium is able to provide both a linked water molecule, and lower the energetic barrier for things like proton release from nearby water.

Hydrolysis is one of the primary mechanisms in DNA regulating enzymes, so magnesium is commonly used in these processes. Most commonly it aids in phosphate ester hydrolysis, a process similar to the cleavage of the third phosphate in ATP, which it also enhances. It also promotes the transfer of phosphoryl groups, which are composed of a single -PO unit. It’s even been shown to play a role in the winding of DNA that produces its infamous double helix shape. If you want a more technical overview of the interplay between magnesium and DNA, I suggest this paper.

Magnesium vs. Calcium

Let’s return to our hypothetical organism and see how things look so far. It seems that we might be able to replace Mg-ATP with Ca-ATP if magnesium were absent, though it wouldn’t work nearly as well without magnesium’s unique ability to aid in hydrolysis and phosphate transfers. Nevertheless, maybe we could scrape by with most enzymes if we assume we could at least get ATP up and running. What other issues might we run into?

If we compare the uses of magnesium and calcium in the body, the difference is quite apparent. Calcium is actually more abundant within the body as a whole, we can group its functions into two broad categories. First, it's a component of hard structural tissue like bone and enamel in humans, and shells or exoskeletons in invertebrates. Calcium carbonate is the main crystal compound found in things like sea shells, and even pearls. In bone and enamel, it’s found in the form of another crystal known as hydroxyapatite, which is composed of both calcium and phosphorus.

The second main use of calcium is in signaling, specifically via channels that allow calcium to flow rapidly into cells, mitochondria, or neurons. This rapid influx of calcium triggers a variety of effects, but one of the most noteworthy is a shift in electric charge known as an action potential. By causing cells or neurons to polarize, calcium regulates processes like heart rate and muscle contraction, and plays a role in creating the body’s electric field. Sodium and potassium are also used as signaling ions in the same way, and often synergize with calcium.

As such, one of calcium’s main roles in cells is in regulating stimulation and stress. This isn’t always pathological, but it is in cases where excitation becomes overwhelming, a state known as excitotoxicity. This occurs most clearly in neurons, and is a direct result of over activation of the neuron by repeated calcium influx. Calcium influx also triggers responses in mitochondria, which can include greater release of free radicals and apoptosis.

The mechanics of this response revolve around the creation of a protein aggregate known as the Mitochondrial Permeability Transition Pore (MPTP) in the outer mitochondrial membrane. MPTP activation triggers a “leak” in the outer mitochondrial membrane, making it permeable to any molecules less than 1500 Daltons in weight. This causes the charge gradient to dissipate as protons are lost, as well as allowing free radicals to diffuse out of the mitochondria into the cell. The mitochondria begins to swell and deform as small molecules and ions that leak into it cause it to retain more water.

Soon the mitochondria undergoes apoptosis, in some cases triggering a similar response in other mitochondria nearby. This loss of mitochondria is one of the main mechanisms that underlie calcium-driven cell stresses, like excitotoxicity in neurons. Drugs that have been researched for their ability to block MPTP activation may even work in part through calcium buffering, in other words by binding up excess calcium to offset its influx.

So what about magnesium? How does it relate to these processes? It turns out that in many ways magnesium works in opposition to calcium, despite having the same +2 charge. One of the most prominent differences is in the relationship between each ion and water. While magnesium’s higher charge/radius ratio causes it to form a large stable hydration shell, calcium is an entire valence shell larger than magnesium, and forms a much weaker hydration shell itself. The downside of this is that the exchange rate between magnesium and water is exponentially lower (10^4 times less) for magnesium than for calcium.

Put simply, this means that magnesium interacts closely with water, but is limited in its ability to rapidly diffuse through a cell. For this reason, it’s not used for depolarization the way calcium is. This doesn’t mean free magnesium ions aren’t used for signaling, but rather than calcium’s lower charge/radius ratio and weaker interactions with water make it ideally suited for this purpose.

The difference in hydration shells also allows magnesium to directly counteract calcium in certain pathways. Within the body they create a divalent cation yin and yang that balances stress, metabolism, and more. For example, rather than inducing action potentials itself, magnesium instead inhibits the various calcium channels that neurons use to trigger them. Higher levels of magnesium are associated with decreased frequency of action potentials and a higher activation threshold for neurons.

The reason for this is that while calcium is able to easily pass through channels like NMDA receptors or VGCCs, magnesium binds to the same proteins but is blocked from moving forward by its inability to give up its hydration shell. You could think of this as magnesium having the same “shape” as calcium, but being too wide to pass through, despite having a lower atomic radius!

This makes magnesium a potent muscle relaxant, analgesic, and neuroprotective substance. Its levels are tightly regulated by transporters in the kidneys, so it isn’t able to reach the same potency as a calcium channel blocking drug, but within the natural range it has a noticeable relaxing effect. This is why research into magnesium’s protective effects is especially promising with regard to excitotoxicity. Magnesium even directly counteracts the activation of MPTP. As one study noted “calcium is a potent activator of the PTP. Subsequently, it was demonstrated that Mg2+, ADP, H+, and NADH increased the Km of calcium-induced PTP formation such that higher levels of calcium are needed to induce PTP opening.”

Strangely, there is more interplay than we might expect between magnesium and calcium binding to different proteins. For example, common calcium-dependent proteins used for signaling in cells/neurons may actually contain empty binding sites, or be occupied by magnesium as a filler ion when the cell is in an unstimulated or dormant state. This includes many common proteins like calmodulin and parvalbumin. Once a pulse of calcium enters the cell, the magnesium is replaced with calcium since both are now present at similar concentration and the proteins have much higher affinity for calcium.

One of the reasons for the ability of magnesium and calcium to interchange is the structure of one of the most common binding sites for calcium in protein, known as an EF-hand motif. This is a fairly simple structure that forms a sort of “box” around the ion, through a pattern known as a helix-loop-helix structure. In each helix on either side of the ion, a series of glutamate and aspartate amino acids are present.

These amino acids are unique in the fact that they each contain two carboxyl groups, one of both sides of each molecule. Carboxyl groups are particularly useful in chemistry due to their ability to give up a proton from their hydrogen atom, leaving the extra electron spread out in a delocalized resonance state, similar to the concept of superposition in quantum physics. This is a state where the extra electron is shared between both oxygen atoms, creating a -1 charge on top of the lone electron pairs the oxygen also contains.

The presence of these carboxyl groups in aspartate and glutamate is also why they sometimes occur as glutamic acid and aspartic acid, their proton-loaded forms. In their -ate forms in the EF-hand motif, they provide sites of negative charge which create a coordinated binding site that can fit calcium or magnesium. There are slight variations in the EF-hand sequence depending on the protein, so some are more selective for calcium than others. Parvalbumin has two EF-hand motifs, while calmodulin has four.

Motifs like this also help cells to differentiate between calcium, magnesium, and other +2 ions like zinc, copper, or iron. Other ions like zinc and copper have their own preferred amino acids to bind with, such as histidine or cysteine in the case of zinc. Magnesium and calcium also form different binding configurations at the same site as seen above. Magnesium tends to form an “octahedral” configuration with six bonds, often including one or more water molecules, while calcium prefers to form seven or eight bonds. This higher number of bonds causes the site to wrap more tightly around calcium, and is the main reason for its selectivity. This paper expands on these binding sites in detail.

This also highlights the interplay between calcium, magnesium, glutamate, and aspartate, which is seen throughout biology. I touched on the NMDA receptor, a calcium channel receptor which is activated by glutamate as a neurotransmitter, but can also be activated more weakly by aspartate as well. A significant portion of calcium within the body is complexed with these two amino acids at any given time. They play a role in the NMDA receptor’s selectivity for calcium, both structurally as building blocks and binding sites, and at the surface level by directly activating the receptor.

The ideal balance between calcium and magnesium in a high energy resting state within the cell is shifted toward higher magnesium and lower calcium. Magnesium fulfills a similar role here as it does in DNA and ATP, acting as a stabilizing agent or placeholder, so that calcium-dependent proteins can swap it out for calcium when the cell is stimulated. Inadequate levels of magnesium push this balance more toward over-stimulation and instability, and open the door to stress cascades like NMDA over-activation and MPTP opening.

Living in a green world

This brings us to my favorite use of magnesium in biology. We’ve spent quite a bit of time talking about magnesium’s role in enzymes. From a chemistry standpoint these enzymes are really just long chains of amino acids with metal ions like magnesium coiled into their structure. Discussing them in detail can lead to getting bogged down in minutiae, so I’ve focused more on general themes. Thankfully, there are other simpler magnesium-containing molecules which we can analyze more clearly.

The best example is Mg-ATP, but the second and my personal favorite is the molecule chlorophyll, which also contains magnesium. As you’ve probably heard, chlorophyll is used in the process of photosynthesis in plants and single celled organisms like cyanobacteria. The question is, how does this process actually occur? And what role does magnesium play?

Chlorophyll is found in a variety of forms, with the primary two known as chlorophyll A and chlorophyll B. The main difference is that in chlorophyll B, the methyl group in the top corner of chlorophyll A is swapped for an aldehyde group -COH. The molecule itself resembles heme, which is also composed of a four-ringed structure with a divalent metal ion in the center.

In the case of heme, this ion is ferrous iron, but in the case of chlorophyll, it’s magnesium. The ring structure itself is known as a “porphyrin” ring in heme, and a “chlorin” ring in chlorophyll, with the main difference being the addition of an extra pentagonal ring in the bottom corner of chlorophyll which is absent in heme. Chlorophyll is also distinct in its long and mostly saturated carbon chain known as a phytol chain.

To say that this porphyrin-based motif is useful in nature would be an understatement. It’s found in a wide variety of places in biology, including in the oxygen transport proteins hemoglobin and myoglobin, in chlorophyll and its various derivatives, in the structure of vitamin B12, and more. There are even porphyrins like heme synthesized in the myelin sheath around neurons, and they’ve been found to play a role in non-mitochondrial electron transport chains in myelin. When any of these porphyrin based systems are deranged, diseases emerge.

In essentially all cases these porphyrin-based structures include a divalent cation as the central atom, with two covalent bonds and four hydrogen bonds holding it in place. You may recall the “octahedral” binding of magnesium in various enzymes we covered earlier, with four sites extending within a plane and one site above and below creating an octahedron shape. The porphyrin binding of things like heme is similar, usually when these molecules are drawn the hydrogen bonds above and below the ion are omitted so that it’s easier to depict, but they are present. In the image above you can only see these vertical bonds depicted in cobalamin (B12).

This similarity is an ancient one, stemming from a prokaryotic ancestor of both mitochondria and chloroplasts. Heme is a far more ancient structure, and the replacement of the ferrous core with a magnesium ion was a later innovation which probably occurred around 2.45 billion years ago, as measured by mineral oxygen deposits. This innovation permanently altered the biosphere.

There are several reasons these porphyrin structures are so important for life, and the most significant can be explained by the nature of their electrons. Remember that covalent atomic bonds form by sharing electrons. Electrons are stored in the s-, p-, and d-orbitals we discussed earlier, so when we talk about sharing these electrons we’re really talking about sharing them from within these orbitals. This means that even though the various chemical bonds are all depicted as a line connecting the two atoms, there are actually different types of bonds.

The most common, found in essentially all single bonds, is what’s known as a “sigma” bond. This bond is named after the s-orbital, and is defined by each of the electron clouds pointing at the opposing atom. You could visualize it as something like this O-> <-O. Since these orbitals are extending out in various directions, sigma bonds give us the maximum degree of overlap between the two and creates a very stable bond.

The second most common, almost always found in double or triple bonds, is what’s known as a pi bond, named after the p-orbital. P-orbitals form a sort of dumbbell shape, with tear drop electron clouds extending out in opposite directions along an axis. Three p-orbitals can form along each of the three dimensions, and each can contain two electrons. When a sigma bond is already present, the space between the two atoms is already filled, so the next bond must use the p-orbitals in the two atoms instead.

The interesting thing about pi-bonds is that while sigma bonds are formed from overlapping electron clouds, pi-bonds instead use the parallel clouds perpendicular to the sigma bond. This means the electrons are “smeared” between each atom equally. This is a concept in chemistry known as electron delocalization, and it opens the door to all kinds of effects.

When multiple pi-bonds are present, they can link up to form something called a pi-conjugated system. This doesn’t require that the double bonds be directly next to each other, but they can be separated by no more than one single bond to fully conjugate. This means that when we have a molecule containing a significant amount of alternating double and single bonds in a row, things get interesting.

One of the best examples where this has an impact is in the chemistry of chromophores, molecules that absorb light, and pigments, molecules that absorb and reflect specific wavelengths creating color. But how does light play a role in the chemistry we’ve discussed so far? It depends on something known as the photoelectric effect.

Everyone has heard of Albert Einstein. He’s most famous for his equation E=mc^2, which led to the theory of general relativity and the understanding that space and time are linked. However, he also made several other contributions to chemistry which are arguably even more significant. He published all three papers within a fairly short period in the German physics journal Annalen der Physik or the “Annals of Physics.” Each of these three papers changed science massively.

One was his theory of general relativity, but the other two dealt with concepts linked more to chemistry, Brownian motion and the photoelectric effect. Brownian motion describes the tendency of small particles like pollen or dust to jiggle constantly in water, an effect which can be observed under a strong microscope. Einstein’s first paper cemented our understanding of the existence of molecules and atoms by explaining this motion using an equation that describes the particle's behavior based on random collisions with smaller particles (molecules).

His second paper was on the photoelectric effect, and proposed an instrumental explanation for the behavior of light. Einstein described light as only interacting with electrons, a behavior which had been observed previously in light’s ability to alter electric current/voltage. Einstein realized that light could be described as distinct packets of energy, or “photons” of light, and that if each of these packets could pass their energy to electrons it would create a relationship between the two where they can receive or exchange energy.

This also helps explain electrons' tendency to stay in their specific orbitals rather than moving closer to the positively charged nucleus. By gaining energy, electrons are able to pull away from the nucleus and jump to a higher orbital, and if they lose energy by releasing a photon, they can fall back to a lower orbital. My personal favorite equation in physics helps to define this principle, and describes the relationship between a photon’s wavelength and energy.

E = h𝜈 = hc/λ

Here “E” is the energy a photon contains, while “h” is Planck’s constant. Planck’s constant is one of the fundamental constants in physics, and describes a “quantum of action” as Planck famously put it. So the photon’s energy is equal to either its frequency “𝜈” times Planck’s constant, or Planck’s constant times the speed of light “c” divided by its wavelength “λ.” This means a photon’s energy increases in proportion to its frequency, and inversely with its wavelength. This is why longer wavelengths of light like infrared possess lower energy than shorter wavelengths like X-rays or UV.

The photoelectric effect explained many properties of chemistry, including the spectroscopy of atoms, which absorb and release light only in specific color bands defined by their orbital structure. It also helped lay the foundation for quantum physics by showing us that photons, electrons, protons, etc, all behave as “quanta” or units/packets/particles and not just waves. This is why Einstein’s Nobel prize was won not for E=mc^2, but for the photoelectric effect.

So how does this relate to chlorophyll photosynthesis?

It turns out that pi-conjugated molecules are particularly good at interacting with light as a result of their large delocalized electron clouds, which catch photons very well. This is why chromophores and pigments include alternating double and single bonds in their structure. The wavelengths that each molecule absorbs also correlate with the diameter of their pi-conjugate motifs, and a few examples are shown in Fig. 1 of this paper, shown below.

If we look at larger examples of these structures, we find that many of them are used in photosynthesis, including chlorophyll! One of the most useful properties of the porphyrin base structure we’ve been discussing is that it acts as an ideal antenna for light, since it's a large planar molecule with a pi-conjugate band that surrounds the central porphyrin ring. You can see this band highlighted in red in the image below.

It’s important to understand that the properties of this structure extend throughout biology, and not just in plants. One area that’s been a focus of research has been the magnetic and electronic properties of the delocalized electrons in porphyrin molecules. Some have been found to act as excellent conductors, both as a result of their central metal ion and their surrounding electron clouds, and have applications in circuits and electronics.

Heme seems to display similar properties in the electron transport chain, where it’s present in cytochromes, including the key fourth step of the chain cytochrome C oxidase. Both the transition metal ion in its center, and the pi-conjugated porphyrin ring, play a role in electron transfer.

Hemoglobin may also benefit from these effects. In its de-oxygenated state, hemoglobin in red blood cells is paramagnetic as a result of its unpaired electrons, meaning it’s drawn toward magnetic fields and orients to them. Oxygenated hemoglobin is instead diamagnetic, being repelled by a magnetic field. The presence of magnetic fields has been shown to shift the pi-electrons of hemoglobin in a way that enhances their absorption of visible light. Effects like these may play a role in the biological impacts of magnetic and electric fields in a way that’s still not fully understood.

Now that we’ve covered some of the interactions of light, electrons, and magnetism within porphyrins, let’s return to chlorophyll and explore the role magnesium plays in photosynthesis.

To do this, let’s gradually zoom in to get a closer and closer look at the process. If we start with plant cells, there’s plenty of variability, but they tend to have more rigid cell membranes and larger internal volumes than many kinds of animal cells. The word “cell” itself was actually coined by the 17th century scientist Robert Hooke, who was observing cells in cork. It came from the latin root “cellula” or “little room,” and described the large pockets of empty space Hooke saw in cork. As we’ve discussed, cell’s are not just bags of water with mostly empty space, and are closer to being solid than liquid. However, if any cell fits this description it would be plant cells, which actually do contain a larger volume of water to increase their transparency.

This higher transparency allows more light to reach the non-transparent components in the center of the cell, green spherical structures known as chloroplasts. These are similar to mitochondria, but are unique to photosynthetic eukaryotes. It should be noted that plants do possess mitochondria as well, and use them to help metabolize the energy they store as glucose and produce extra ATP. Most of their energy producing organelles are chloroplasts.

If we look more closely at the structure of these chloroplasts, we find that they’re quite similar to mitochondria. Both are descended from the same symbiotic ancestor, and both contain two membranes, an inner and outer membrane, with a narrow space in between. However, while mitochondria and bacteria uses this intermembrane space to create a charge gradient, chloroplasts have a different approach.

Chloroplasts have essentially inverted mitochondria by creating a third membrane. Within the chloroplast we find structures that resemble stacks of pancakes, which are hollow inside and form a continuous space inside them. These structures are known as thylakoids, and the space inside them is known as the thylakoid lumen, while the fluid inside the chloroplast itself is known as the stroma.

The chloroplast instead uses the space inside the thylakoid stacks to create a charge gradient. It pumps protons into the lumen, creating a pH of around 5 within the thylakoids, and around 7.8 in the stroma. This means there are exponentially more protons within this space, and the chloroplast can harvest this dissipative flow of protons from the lumen into the stroma to create ATP using an electron transport chain.

If we zoom in still further, we can see the workings of this electron transport chain in more detail. To summarize before we move deeper, light provides energy at two protein complexes named photosystems I and II. Ironically, photosystem II comes first, and is responsible for one of the most important reactions in biology:

2 H2O -> O2 + 4 e- + 4 H+

By splitting the hydrogen in water into protons and electrons, the thylakoid is able to “burn” water for energy. The protons are released into the lumen, while the electrons are passed to a useful redox molecule known as plastoquinone, which can then be oxidized into plastoquinol by giving its protons and electrons to the next step, cytochrome b6f. Plastoquinone is analogous to CoQ10 or ubiquinone, which is another quinone used to transfer protons and electrons in the same way in mitochondria.

Cytochrome b6f uses plastoquinone to pump even more protons into the lumen, while passing its electrons to another electron acceptor known as plastocyanin. Plastocyanin is a blue (“cyan”) copper-containing protein which also acts as an electron acceptor/donor by passing electrons through the transition metal copper at its center. Plastocyanin is analogous to the protein cytochrome C in mitochondria, which passes electrons from a heme core on to the cytochrome C oxidase.

The final step is photosystem II, which instead is coupled to several ferredoxin reaction centers. Ferredoxin is another redox cofactor, which uses Fe-S or “iron-sulfur” clusters to pass electrons and protons on NADP+, where they can be recombined as hydrogen to form NADPH. NADPH is analogous to NADH, but it is used more in synthesis reactions, and in this case it will go on to provide hydrogen to the Calvin cycle to combine with carbon from CO2 to form glucose. These iron-sulfur clusters are extremely ancient, and I believe they’re a way of carrying a portable remnant of the iron-sulfur crystals which helped catalyze reactions for early life in hydrothermal vents. We also see the redox relationship between iron and copper playing out here as well.

So what happens if we go one level deeper? What role does magnesium play?

We can illustrate the complexity and beauty of this process by exploring photosystem II, the pivotal reaction in the entire process. The most important thing to understand about both photosystem proteins is that they are absolutely loaded with cofactors, particularly pigment molecules. While we most often think of chlorophyll as the cofactor for photosynthesis, there are actually dozens of chlorophyll molecules present, as well as chlorophyll analogs like pheophytin, which contains no ion in the center of its chlorin ring.

Carotenoids are also used in photosynthesis, and just like chlorophyll and pheophytin, they contain a large pi-conjugated system of electrons. In this case, the delocalized electrons are spread across their long carbon chain. Beta-carotene, xanthophylls, and more, are present and play an active role similar to chlorophyll.

As light enters the thylakoid, a photon is captured by one of the pi-conjugated electron clouds that surround chlorophyll. The energy of the photon is passed to a delocalized electron within the electron cloud, causing it to become excited and raised to a higher energy level. This forms a coupled state known as an “exciton.”

Normally the term exciton comes up most often in the context of semiconductor physics. It is a combination of two components, the electron that’s excited enough to have left its orbital, and the “hole” left behind which has a positive charge. There are several different types of excitons, which are characterized by just how far the electron can be from its hole. Frenkel excitons are localized so that the electron is linked closely with the orbital it left, while charge transfer or CT excitons are able to span the gap to an adjacent orbital, allowing the exciton to be split. Wannier-Mott excitons are even more delocalized, and can be spread across large conduction bands of delocalized electrons, such as in a large pi-conjugated structure like graphene.

All three forms of excitons are found in photosynthesis, and they help to explain its magic. The goal of photosystem II is to separate the electron from its hole across a greater distance, so that it can create a stable charge separation.

The heart of photosystem II is a pair of chlorophyll molecules lined up like two parallel planes. This is known as the “special pair” and acts as an entry point for photons. When a photon reacts with the special pair, it creates an exciton, from which the energy of excitation is transferred. The energy is transferred using what’s known as Forster resonance energy transfer or FRET. The excitation of one chlorophyll molecule is able to pass its excitation energy to another chromophore via a coupling of the positive and negative shifts in their electron clouds known as dipole coupling.

You can think of this as the electrons swinging back or forth slightly at different points in the structure, so that some points are slightly more positively charged while others are more negative. This creates a point of slight charge separation that’s known as a dipole moment, and by synchronization of these dipoles so that the positive and negative points line up, they can form a dipole couple. These dipole couples allow FRET to occur via a “non-radiative” mechanism.

The reason it’s referred to as non-radiative is that even though chlorophyll and other chromophores are fluorescent, and would normally release a photon, no actual photon of light is released here. Instead, the coupling allows a transfer of what’s known as a “virtual photon” which passes the excited state from one molecule to another. These virtual photons are used in physics to describe electromagnetic interaction between molecules where no actual photon can be observed.

Let’s set up a scenario to compare this to. If you told someone you’re going to send them a letter, say an invoice for some money they owe you, you’d expect to have to have a postman deliver the letter to them. However, if they’re already aware of how much they owe you because the two of you are in communication (dipole coupling), they could instead write up an identical invoice themself, eliminating the need to have a postman deliver it.

This energy transfer is determined by the spectrum of light that can be absorbed by each chromophore. The two must overlap for the transfer to occur, despite the fact that no actual photon is emitted. The transfer ability is also proportional to distance, rapidly falling off over longer separation as dipole coupling can no longer occur. This is why dozens of chlorophyll, pheophytin, and carotenoid molecules must be packed close together at distances of ~1 nm or less, and why a variety of different chromophores are used. It minimizes energy loss by ensuring that a wide range of wavelengths are absorbed.

This also brings us into a fascinating new field known as quantum biology. In recent years there’s been an increased interest in how quantum effects might help explain the highly efficient nature of energy transfer in photosystems. Normally we’d expect some randomness to cloud the process, since there are numerous chromophores the exciton could move through to reach its destination, and some may be further away than others creating dead-ends or less efficient routes.

To solve this, the exciton performs what’s known as a “quantum walk,” where instead of just picking a random path of reaction centers to move through, it is delocalized. In some sense, it’s as if the exciton takes every path at once, only materializing when it reaches a concrete destination. This allows the exciton transfer to reach an astonishing efficiency of 99%! This relies on a concept known as quantum coherence, which is outside the scope of this article.

So what is the exciton’s destination? It moves through this system of chromophores until it reaches a stable redox cofactor, like one of the quinones in photosystem II. From here these electrons can be transferred out of the photosystem to plastoquinone and on down the electron transport chain. This leaves the exciton’s “hole” in the special pair of two chlorophyll molecules.

Thankfully, help is nearby. Another electron can be moved to fill the hole in the special pair, and this electron comes from the reaction center known as the oxygen-evolving complex or OEC. This complex is linked closely with the special pair, and a tyrosine molecule acts as a bridge to transfer the electron from the OEC to the special pair.

The OEC itself is another example of a particular element serving a fine-tuned purpose. While magnesium’s charge/radius is used because it allows magnesium to hold closely to its hydration shell, the OEC uses another element for a similar purpose, manganese. Manganese has a similar atomic radius to magnesium (161pm vs. 145pm), and a similar charge ranging from +2 all the way up to +7. In the OEC we find a cluster of manganese oxide and calcium, Mn4CaO5.

The manganese atoms exist in this cluster in the Mn3+ state, and form a tight hydration shell around it. When the cluster transfers an electron to the special pair, it oxidizes one of the manganese ions into a +4 state. Once this has occurred four times, with four excitons being created and transferred, the special pair being replenished four times, etc, then OEC is ripe to be able to split two molecules of water. Technically this proceeds in four phases, known as S1-S4, with hydrogen-splitting reactions in water occurring after S2 and S4.

Obviously this process is incredibly complex and finely tuned. We’ve discussed many effects and the roles of many other ions in this process, but what about magnesium? We know it plays a role in the main chromophore chlorophyll, but how?

Chlorophyll and magnesium

There are a few reasons why magnesium is ideally suited for its role in photosynthesis. First, it’s a divalent cation, which allows it to fit neatly into the chlorin ring structure. If it were a transition metal like iron or copper, it would have multiple electron states which would mean it could be oxidized, altering its charge and potentially ruining its ability to bind to chlorophyll.

It also has a small atomic radius, which fine-tunes the width of the pi-conjugated band around chlorophyll. This means the absorption spectrum of chlorophyll, which is based on the width of its pi-electron band, is also determined by magnesium. A larger or smaller atom would shift the width of the molecule, altering the absorption spectrum.

Chlorophyll has absorption peaks in the red and blue spectrum, and this helps it to maximize its ability to harvest energy from light. Compared to earlier bacteriorhodopsin photosynthesis in bacteria, which instead reflected blue/red light, magnesium/chlorophyll photosynthesis is extremely efficient. The high degree of overlap in absorption spectra between chlorophyll molecules, which is necessary for the FRET mentioned earlier, also depends on magnesium.

Furthermore, the ability of chlorophyll to hold an exciton and transfer excitons between chlorophyll molecules is dependent on magnesium. If we were to replace the magnesium with an iron atom, the exciton would be quenched due to the molecule becoming paramagnetic similar to deoxygenated heme. This occurs via an effect known as spin-orbit coupling, and would completely inhibit the hyper-efficient energy transfer between chromophores.

A similar effect would occur if we replaced magnesium with a larger divalent cation like zinc. In this case, the quenching occurs instead as a result of the heavy atom effect. This would mean that any excitons created would quickly return to the ground state, and dipole coupling would be canceled out limiting FRET even if an exciton were maintained.

In short, magnesium is fine-tuned to enable every aspect of photosynthesis. It is one of only a handful of ions with the right radius to fit the chlorin ring of chlorophyll and form the correct number of bonds. Its absorption spectra provide maximum absorption of some of the wavelengths that provide the highest energy, and its small size and lack of magnetic effects allow chlorophyll to maintain excited electron states.

As one paper found: “Mg(2+) increased absorbance for visible light, improving energy transfer among amino acids within the PSII protein complex and accelerated energy transport from tyrosine residue to chlorophyll-a. The photochemical activity and oxygen evolving rate of PSII were also enhanced by Mg(2+).”

Final thoughts

This has been my longest piece of writing to date, and for those that made it to the end, thank you for reading. This has been a sweeping overview of a wide variety of subjects, from the origin of life in deep hydrothermal vents, to the interplay between stimulation and relaxation, and finally to quantum biology and the ability of plants to harvest energy from light. In each of these domains we find magnesium, the most widely used metal cofactor in enzyme biology.

As with any subject, there’s always more I could have covered. For example, it’s been suggested that magnesium could play a role in lipid membrane stability by forming cross-links between the charged phosphate heads of phospholipids. I also only briefly touched on the use of magnesium as a signaling ion similar to calcium, and there’s a growing body of research on its use for this purpose.

I hope that this has given you a glimpse into how crucial magnesium is for biology, stabilizing and activating ATP, coiling and repairing DNA, regulating calcium across many systems, and even protecting the brain. This makes it all the more concerning that the majority of people are estimated to be magnesium deficient. So to finish things off, I’m going to include a brief section on how I’d personally go about increasing magnesium intake.

How to Supplement Magnesium

With any nutrient, it’s important to take balance into account. For example, many people are mistakenly led to believe that their anemia is a result of iron deficiency, when it can just as often occur as a result of copper deficiency, or lack of B12 or other B-vitamins. Even the suggestion of reducing sodium to combat high blood pressure rarely takes other factors like potassium intake and insulin resistance into account.

When it comes to magnesium, things are relatively simple. We don’t have to worry about it depleting or competing with other nutrients, as it has its own designated uptake channels in the gut (TRPM channels). In a general sense, you do want to make sure you promote plenty of calcium and potassium intake to complement it. I’d shoot for 3.5-4g of potassium and 1.5-2g of calcium per day personally, though individual requirements may vary.

As for magnesium intake, we want to treat the RDI as the minimum required, and assume that during times of high stress, physical exertion, poor sleep, etc, magnesium demand will be increased to the point that the RDI is insufficient. For this reason the best approach to magnesium intake seems to be to increase it to the point of surplus, while titrating up gradually to avoid any discomfort.

The RDI for magnesium is 400mg per day for adults, so this means we should shoot for at least this much. When supplementing, it’s important to remember that each supplement will contain magnesium plus something else, usually in the form of an ionic salt like magnesium chloride, or an amino acid chelate like magnesium glycinate. There are also other forms like magnesium citrate, malate, or oxide. Personally I suggest magnesium chloride, glycinate, or taurate as the best forms.

When calculating intake it’s important to remember to look for the percentage of elemental magnesium, not just the weight of any given form. For example, magnesium glycinate is 14.1% magnesium, magnesium chloride is 11.9% magnesium, and magnesium taurate is 8.9% magnesium. A good starting point would be in the range of 4-5mg/lb bodyweight per day, but start slowly and titrate up.

One of the biggest side effects people get with magnesium supplementation is a laxative effect or other impacts on digestion. This is particularly common with forms like magnesium oxide, citrate, and hydroxide. It stems mostly from magnesium’s limited absorption as a result of TRPM channel availability. Historically magnesium has only been present in smaller amounts from food, so we’re only able to get limited absorption in a given time period.

If you experience a laxative effect from magnesium, you took more than you could absorb. Since magnesium is a very “hygroscopic” or water-loving substance, it will tend to hold water in the bowel, leading to this side effect. The solution is to spread magnesium intake throughout the day so that small amounts can be absorbed at a time.

For example, you can use the magnesium chloride “salt” form of magnesium dissolved in water. One teaspoon is equal to about 3g or 750 mg elemental magnesium, and by dissolving a small amount in water (something like 1/4 tsp) you can add both magnesium and chloride as electrolytes to water, then drink it gradually over the next few hours. This is an easy way to get sustained magnesium intake throughout the day.

Another option is to use single doses, such as in capsules. The amino acid chelate magnesium glycinate has much better absorption, and can be taken in capsules with meals, for example morning/noon/night. Something like 500mg per capsule is a good starting point, and you can always increase from there or combine it with magnesium chloride in water. You can also swap other forms like magnesium taurate for an additional calming effect from taurine.

Lastly, topical magnesium baths or sprays can be used. These have the advantage of not being limited in the same way by gut absorption. Using an epsom salt bath, you can essentially maximize magnesium intake via topical absorption. The most extreme example of this is sensory deprivation tanks, which commonly use 1,000 to 1,500 lbs of epsom salt per float tank! So despite the high absorption, unless you have kidney issues your body should be able to regulate topical magnesium distribution just fine, and excess will be passed through the urine.

You may experience muscle relaxation or a very mild sedating effect as a result of topical magnesium acting as a calcium channel blocker, but otherwise, side effects are minimal. If you’re doing a bath at home, you can easily use several pounds of either epsom salt (magnesium sulfate) per bath, or use smaller amounts to soak your feet/hands. Magnesium chloride can also be substituted for epsom salt, though both should have similar absorption.

If you want to make a homemade magnesium spray, simply dissolve magnesium chloride in water as a factor of volume. For example, ¼ cup magnesium chloride per cup of water for a 25% spray, which you can put in an empty spray bottle. Apply particularly around the back, chest, and neck, as these have the best absorption into the brain and cardiovascular system.

As far as food sources of magnesium go, a few of my favorites are: pumpkin, chia, flax, and hemp seeds, seafood like salmon, halibut, shrimp, and various shellfish, cruciferous vegetables like broccoli (especially sprouts) dark leafy greens like spinach or kale (when well cooked, avoid if sensitive to oxalates), sprouted brown rice, chocolate (source carefully to avoid heavy metals), and oats, rye, or buckwheat (assuming you tolerate grains, source well). Grains, nuts, and seeds should ideally be sprouted to significantly reduce their antinutrient content, which limits magnesium absorption. Soaking is a good alternative if sprouting isn’t an option.