EMEM vs DMEM: What the Modification Actually Changed

A large share of the confusion behind this question is nomenclature rather than chemistry.

EMEM stands for Eagle's Minimum Essential Medium. MEM stands for Minimum Essential Medium. They name the same formulation - the one Harry Eagle published in 1959. ATCC labels its version EMEM; Gibco, Sigma and most other suppliers label theirs MEM. A protocol calling for EMEM is asking for MEM, and vice versa.

What is not identical is the bottle. Suppliers modify the base formulation and the differences are real:

• Gibco MEM (11095) contains Earle's salts with 2.2 g/L sodium bicarbonate, 1.0 g/L glucose and 2 mM L-glutamine, with no non-essential amino acids and no sodium pyruvate.
• ATCC EMEM (30-2003) contains Earle's salts with only 1500 mg/L sodium bicarbonate, and adds non-essential amino acids, 2 mM L-glutamine and 1 mM sodium pyruvate. Its specification is pH 7.0-7.4 and osmolality 260-320 mOsm/kg.

So an ATCC EMEM already contains the non-essential amino acids that a Gibco MEM user has to add from a 100X supplement, and it carries a third less bicarbonate. If you switch supplier and your cultures start behaving differently, this is the first place to look. Always read the modification line on the label rather than trusting the four letters.

Harry Eagle established the minimum set of nutrients that cultured mammalian cells cannot make for themselves. Basal Medium Eagle came first in 1955; Minimum Essential Medium followed in 1959 with a revised and slightly richer amino acid complement. The design philosophy is in the name - the minimum that works, not the maximum that helps.

Renato Dulbecco and Gordon Freeman modified Eagle's medium in 1959 for a specific purpose: growing mouse embryo fibroblasts to support plaque assays for polyoma virus. That application needed cells growing hard, so the modification enriched almost everything. The result was Dulbecco's Modified Eagle Medium.

This history explains the character of the two media. EMEM is lean and was designed as a defined minimum. DMEM is rich and was designed to push proliferation. Neither is better; they were built to different specifications, and the right one depends on which behaviour your cells need.

The common claim that DMEM has four times the amino acids of MEM is wrong, and getting it right matters if you are reasoning about nutrient limitation. Comparing the Gibco formulations component by component:

Vitamins: four times, almost exactly. Choline chloride, D-calcium pantothenate, folic acid, niacinamide and thiamine hydrochloride all go from 1 mg/L in MEM to 4 mg/L in DMEM. Riboflavin goes from 0.1 to 0.4 mg/L, and i-inositol from 2 to 7.2 mg/L. There is also a substitution rather than an increase: MEM uses pyridoxal hydrochloride, DMEM uses pyridoxine hydrochloride.

Amino acids: about twice, with exceptions. Leucine, isoleucine, lysine, methionine, phenylalanine, threonine, tyrosine, valine and cystine are all close to doubled - leucine goes from 52 to 105 mg/L, methionine from 15 to 30. But arginine goes down, from 126 mg/L in MEM to 84 in DMEM. Histidine is identical at 42 mg/L in both. Tryptophan rises only from 10 to 16 mg/L.

Glutamine: doubled, from 292 mg/L (2 mM) to 584 mg/L (4 mM).

Glycine and serine: added. MEM contains neither; DMEM contains both at 0.4 mM. These are the only two non-essential amino acids DMEM carries - it has none of the other five in the standard NEAA set.

Glucose: unchanged or quadrupled, depending on version. MEM contains 1.0 g/L. DMEM low glucose also contains 1.0 g/L; DMEM high glucose contains 4.5 g/L (25 mM).

Sodium bicarbonate: substantially raised, from 2.2 g/L (26.2 mM) in Gibco MEM to 3.7 g/L (44.0 mM) in DMEM. Sodium chloride is correspondingly lowered, from 6.8 to 6.4 g/L, to keep osmolality in range.

Ferric nitrate added at 0.1 mg/L, an iron source MEM lacks entirely.

Phenol red raised from 10 to 15 mg/L, which is why DMEM simply looks redder than MEM at the same pH.

The salt base is otherwise shared: both carry calcium chloride at 1.8 mM, magnesium sulfate at 0.81 mM and potassium chloride at 5.33 mM.

MEM is sold on two different salt bases, and this trips people more often than the amino acid differences do.

MEM with Earle's salts is the standard version. Its salt component is exactly Earle's Balanced Salt Solution - 2.2 g/L sodium bicarbonate, 6.8 g/L sodium chloride, 0.4 g/L potassium chloride, 140 mg/L monobasic sodium phosphate and 1.0 g/L glucose, plus calcium and magnesium. It is designed for a CO2 incubator, conventionally 5%.

MEM with Hanks' salts substitutes Hanks' Balanced Salt Solution, whose bicarbonate is only 0.35 g/L. That formulation is designed to hold pH in ambient air, in a sealed vessel, without a CO2 incubator.

DMEM is made only on an Earle's-derived base, and a modified one - more bicarbonate, slightly less sodium chloride and phosphate, plus the ferric nitrate. There is no Hanks'-salts DMEM, because 3.7 g/L bicarbonate and ambient air are incompatible by design.

The rule that connects all of this is that bicarbonate must be matched to the atmosphere. Suppliers give roughly: below 1.5 g/L for ambient air or 4% CO2, 1.5-2.2 g/L for 5%, 2.2-3.4 g/L for 7%, and above 3.5 g/L for 10%. Gibco DMEM at 3.7 g/L is formulated for 10% CO2. Most laboratories run it at 5% anyway and it works, but the medium is then above its design pH and this is why a flask of DMEM can look pink from the day it is opened. ATCC EMEM at 1.5 g/L is explicitly specified for 5% CO2 in air, and ATCC notes that more bicarbonate may be needed at higher CO2.

If you move a cell line between EMEM and DMEM without thinking about the incubator, the pH shift alone can account for the change in growth you observe.

Start with the cell bank's recommendation. ATCC, ECACC, DSMZ and JCRB each publish a specific medium for each deposited line, and those recommendations reflect what the line was characterised in. Deviating is legitimate, but it changes growth rate, morphology and often phenotype, and it makes your data harder to compare with published work on the same line.

EMEM/MEM suits slower-growing adherent cells with moderate metabolic demand: diploid fibroblast strains such as MRC-5 and WI-38, many primary cells, Vero cells, and a great deal of classical virology, where MEM has been the standard host-cell medium for decades. ATCC recommends EMEM for a large number of adherent lines including HeLa. Its leaner composition is an advantage where you want cells to grow steadily rather than as fast as possible, and where a defined minimum background matters to the assay.

DMEM suits rapidly proliferating, metabolically demanding lines: HEK293 and its derivatives, NIH/3T3, COS, many transformed and transfected lines, and most transient transfection and virus production work. The extra glutamine, vitamins and - in the high-glucose version - carbon are what those cells consume.

High versus low glucose DMEM is a separate decision. High glucose (4.5 g/L) supports high-density growth, transfection and virus production, and is standard for hybridomas. Low glucose (1.0 g/L) is preferred for primary hepatocytes, for cells whose metabolism you are studying, and anywhere lactate accumulation and rapid acidification are a problem. Note that low-glucose DMEM still carries 3.7 g/L bicarbonate - lowering the glucose does not change the CO2 requirement.

MEM alpha is worth knowing as the third option. It is a further modification of MEM carrying additional amino acids including the non-essential set, extra vitamins, sodium pyruvate and lipoic acid, and is sold with or without nucleosides and deoxyribonucleosides. It is the standard choice for mesenchymal stromal cells and much haematopoietic work, and is closer in richness to DMEM than to plain MEM.

Substituting DMEM for EMEM one-for-one is the common mistake, and the consequences are predictable rather than mysterious.

Going from EMEM to high-glucose DMEM raises glucose 4.5-fold, glutamine two-fold and vitamins four-fold at once. Cells respond by growing faster and producing much more lactate, so the medium acidifies sooner and needs feeding more often. Growth curves, doubling times, seeding densities and every schedule derived from them shift. If the incubator remains at 5% CO2, the higher bicarbonate also puts the starting pH above where it was.

Going the other way, from DMEM to EMEM, cells that were relying on the richer background may slow or stall - and if you are moving to a Gibco MEM rather than an ATCC EMEM, you have also removed the non-essential amino acids and the pyruvate without meaning to.

Three rules make the transition safe. Adapt gradually, over two or three passages, using intermediate blends rather than a single hard switch. Never change medium in the middle of an experiment or between an experiment and its control. And re-establish the growth curve after adapting, because every downstream number - seeding density, feed interval, harvest day - was derived from the old medium.

If you are moving from a supplier's MEM to a formulation that lacks non-essential amino acids, add them back at 1X from a 100X supplement, which gives 0.1 mM of each of the seven. That single addition accounts for a large part of the apparent difference between an ATCC EMEM and a plain MEM.