Cell culture buffers

Most mammalian cells grow best between pH 7.2 and 7.4, and the tolerable band is narrower than people assume. Reported optima differ by cell type - transformed lines are often described as favouring the lower part of the range, around pH 7.0-7.4, and normal fibroblasts the upper part, around 7.4-7.7 - but in every case the working window is a few tenths of a pH unit wide.

Holding that window is not passive. Cells continuously acidify their medium: glycolysis produces lactate, and respiration produces carbon dioxide. In a dense culture the acid load is substantial, which is why an over-confluent flask turns yellow within a day. A buffer's job is to absorb that load without letting pH move outside the window before the next feed.

A buffer resists pH change most effectively within about one pH unit either side of its own pKa, so a buffer for cell culture needs a pKa near 7.4. That single requirement eliminates most laboratory buffers immediately. What remains has to satisfy several more conditions: it must not cross the plasma membrane and buffer the cytoplasm, must not chelate the calcium, magnesium, zinc and iron the formulation supplies, must not absorb light in ways that interfere with assays, must be stable in solution and non-toxic at working concentration, and must be affordable at the litre scale.

Tris illustrates what happens when these conditions are not met. It has a pKa of 8.06 at 25 degrees C, which is already too high, a temperature coefficient of -0.028 pH units per degree C, which is roughly twice that of HEPES, and enough lipid solubility to permeate membranes - which is precisely why it is toxic to many mammalian cells and is not used for cell culture despite being ubiquitous in biochemistry.

There is a second specification that is easy to forget: osmolality. A buffer that fixes pH while pushing total osmolality outside the 260-320 mOsmol/kg range that mammalian cells tolerate has not solved the problem. Buffer choice and osmolality are the same conversation.

The default buffer in classical media is sodium bicarbonate paired with gaseous CO2, and it is used for good reasons: it is the physiological buffer, it is cheap, and bicarbonate is also a nutrient that cells consume in carboxylation reactions.

The chemistry is a linked equilibrium. Carbon dioxide dissolves in the medium and hydrates to carbonic acid, which dissociates into bicarbonate and a proton. The apparent pKa of the carbonic acid system under physiological conditions is 6.1, and the resulting pH is described by the Henderson-Hasselbalch relationship:

pH = 6.1 + log( [HCO3-] / (0.03 x pCO2) )

where pCO2 is in mmHg and 0.03 is the solubility coefficient for CO2 in mmol/L per mmHg.

The important consequence is that pH is set by the ratio of bicarbonate to dissolved CO2, not by either alone. Neither the sodium bicarbonate in the formulation nor the CO2 in the incubator determines pH on its own. They determine it together, which is why they must be specified together.

This also explains the system's central weakness. CO2 is volatile. Room air contains about 0.04% CO2 against 5% or 10% in the incubator, so the moment a flask comes out, dissolved CO2 leaves the medium while bicarbonate stays behind. The ratio rises, and so does the pH. Under a laminar flow hood a bicarbonate-buffered medium can drift visibly alkaline within ten to twenty minutes.

A third consideration: bicarbonate solutions are chemically unstable in the open. Sodium bicarbonate is added late in medium preparation, and powdered media are commonly supplied without it precisely so that it can be added fresh.

This is the single most common configuration error in cell culture, and the arithmetic to get it right is worth doing once.

Take a 5% CO2 incubator at sea level. Correcting for water vapour at 37 degrees C, pCO2 is about 0.05 x (760 - 47) = 36 mmHg, so dissolved CO2 is about 0.03 x 36 = 1.1 mmol/L. For a target pH of 7.4:

log([HCO3-] / 1.1) = 7.4 - 6.1 = 1.3, so [HCO3-] = 1.1 x 20 = 22 mmol/L, which is 1.8 g/L sodium bicarbonate (MW 84.01).

Repeat at 10% CO2 and dissolved CO2 doubles to 2.1 mmol/L, requiring about 43 mmol/L bicarbonate - 3.6 g/L.

Those numbers are not coincidental. They are why standard DMEM is formulated at 3.7 g/L sodium bicarbonate and designed for a 10% CO2 incubator, and why formulations at 1.5-2.2 g/L are designed for 5% CO2. RPMI 1640 at 2.0 g/L is a 5% CO2 medium.

Medium / solution	NaHCO3	Approx. mM	Intended atmosphere
DMEM (standard)	3.7 g/L	44	10% CO2
DMEM (5% CO2 variants)	1.5-2.2 g/L	18-26	5% CO2
RPMI 1640	2.0 g/L	24	5% CO2
EBSS (Earle's)	2.2 g/L	26	5% CO2
HBSS (Hank's)	~0.35 g/L	~4	Ambient air
DPBS / PBS	none	0	Ambient air (phosphate-buffered)

Running a 3.7 g/L formulation at 5% CO2 leaves the medium persistently alkaline - typically drifting toward pH 7.6-7.8, which cells tolerate poorly and which shows up as a stubbornly pink-purple medium that never turns yellow even in a dense culture. The reverse, a low-bicarbonate medium at 10% CO2, gives persistent acidosis.

Altitude matters too, and is genuinely overlooked. The calculation above uses 760 mmHg. At 1,500 m, atmospheric pressure is roughly 630 mmHg, so the same 5% CO2 setting delivers about 17% less dissolved CO2 and the medium sits higher in pH. Laboratories at altitude commonly run a higher CO2 percentage to compensate.

The zwitterionic buffers described by Norman Good and colleagues were selected against exactly the criteria listed at the top of this page, and they are the answer to the volatility problem in the bicarbonate system.

HEPES is the one you will meet. pKa 7.48 at 25 degrees C, useful range about pH 6.8-8.2, working concentration 10-25 mM in culture. Because neither side of its acid/base pair is a gas, its buffering capacity is fixed by how much you added and is unaffected by the atmosphere. That makes it the standard addition for live-cell imaging without gas control, cell sorting, tissue dissociation on the bench, transport of live cultures, and any protocol with long handling steps.

Three cautions apply. HEPES reacts photochemically with riboflavin in the medium under visible light to produce hydrogen peroxide, so HEPES-containing media must be stored and handled in the dark. It is cytotoxic above the working range. And it is a buffer only - it contributes nothing nutritionally, so it supplements bicarbonate rather than replacing it.

MOPS has a pKa of 7.20 and buffers usefully from about pH 6.5 to 7.9. It appears in specialist and microbial formulations and in some electrophoresis and protein work, but has not displaced HEPES in mammalian culture.

Other members of the family - TES, BES, PIPES, TAPS - sit at various points across the physiological range and are chosen in biochemical work where a specific pKa is needed. They are rarely used in growth media.

The general rule for organic buffers in culture is that they are supplements, not replacements. Removing bicarbonate in favour of HEPES alone will slow or stop growth in many lines, because bicarbonate is a substrate as well as a buffer. The standard configuration is bicarbonate matched to the incubator, plus an organic buffer when the work demands it.

Phosphate is a genuine physiological buffer with a convenient pKa. The relevant dissociation, dihydrogen phosphate to hydrogen phosphate, has a pKa of 7.20, which puts the physiological range squarely inside its useful band.

Total phosphate content differs more than most people expect between the two bottles that sit side by side on every shelf. A standard PBS at pH 7.4 carries about 4.0 mM total phosphate, while Dulbecco's formulation (DPBS) carries about 9.5 mM - well over twice as much. DPBS therefore resists a pH shift considerably better when you add cells, an acidic reagent or a slug of medium to it, which is a practical reason to prefer it for anything beyond a quick rinse.

Calcium and magnesium concentrations, for the versions that contain them:

Solution	Calcium (Ca2+)	Magnesium (Mg2+)	Total phosphate
PBS pH 7.4	none	none	~4.0 mM
DPBS, no Ca/Mg	none	none	~9.5 mM
DPBS, with Ca/Mg	0.90 mM (CaCl2 100 mg/L)	0.49 mM (MgCl2.6H2O 100 mg/L)	~9.5 mM
HBSS, with Ca/Mg	1.26 mM (CaCl2 140 mg/L)	0.90 mM (MgCl2.6H2O 100 + MgSO4.7H2O 100 mg/L)	~0.78 mM

What the divalent cations are actually for. They are not there to buffer anything - they hold cells attached. Integrins require a divalent cation at their metal-ion-dependent adhesion site to bind matrix, and cadherins require calcium to hold the rigid extracellular conformation that lets them bind each other. A calcium-containing wash therefore keeps a monolayer stuck down and keeps cell-cell junctions intact; a calcium-free wash begins loosening both the moment you add it. Calcium and magnesium are also required by the metalloprotease dissociation enzymes, dispase and collagenase, which is why those are diluted in a Ca/Mg-containing solution.

One consequence worth knowing: calcium and magnesium phosphates have limited solubility, so calcium-containing phosphate salines are prone to precipitation in 10X concentrates and on autoclaving. Sterile filtration at 0.22 micron is the safer route for any phosphate saline containing calcium.

Phosphate-buffered solutions are the workhorses of handling: washing monolayers before trypsinisation, diluting reagents, rinsing, brief transport, and resuspending cells for counting. They are simple, stable, inexpensive, and require no incubator gas.

They are not growth media, and the reasons are worth being explicit about.

Phosphate has limited buffering capacity against a real metabolic acid load. It supplies no carbon, no amino acids, no vitamins and, in most formulations, no glucose. Phosphate at higher concentrations precipitates with calcium as calcium phosphate, which is why the calcium- and magnesium-containing formulations exist as separate products from the Ca/Mg-free ones and why you should not casually increase phosphate content. And bicarbonate-free solutions deprive cells of a substrate they need.

The calcium/magnesium distinction is the one that trips people up. Ca/Mg-free DPBS is the correct wash before trypsinisation, because divalent cations are exactly what the EDTA in trypsin-EDTA is there to remove - washing with a calcium-containing formulation immediately beforehand works against the dissociation reagent. Ca/Mg-containing DPBS is the correct diluent when you need adhesion and cation-dependent enzyme activity preserved, which includes diluting dispase and collagenase, both of which are calcium-dependent metalloproteases.

For handling steps that need a little more support than PBS - short-term holding of tissue, brief incubations, dissociation buffers - a balanced salt solution such as HBSS or EBSS is the better choice, because it supplies glucose and a low bicarbonate content alongside the salts.

Phenol red is not a buffer. It is a pH indicator included in most media at approximately 15 mg/L so that the state of a culture can be read at a glance.

Its colour transitions are:

• Yellow below about pH 6.8 - the medium is acidic. In a culture this normally means high cell density with lactate accumulation, or bacterial contamination, both of which need attention now rather than at the next scheduled feed.
• Orange to red through pH 7.0-7.4 - the working range. A healthy culture sits red and moves toward orange as it becomes dense.
• Pink to purple above about pH 8.2 - the medium is alkaline. The usual causes are a bicarbonate/CO2 mismatch, an incubator that has lost its CO2 supply, a door left open, or a flask that has simply been on the bench too long.

Two limitations matter. First, phenol red is a coarse instrument: it will show you a problem of several tenths of a pH unit, but it will not distinguish pH 7.2 from 7.4, and judging colour by eye is subjective under different lighting. If pH is an experimental variable, measure it.

Second, phenol red is not biologically inert. It has weak oestrogenic activity and can bind oestrogen receptors, which is a real confounder in hormone-responsive systems - breast cancer lines in particular. It also fluoresces and absorbs in ranges that interfere with fluorescence-based assays and absorbance readings. This is why phenol-red-free versions of essentially every medium and reagent exist: they are for hormone work, imaging and plate-reader assays, and choosing one is a deliberate decision to give up the visual pH readout in exchange for a cleaner background.

A buffer is only as good as the bottle it came out of, and concentrated stocks are where most avoidable variability enters a laboratory. A few rules apply across the shelf.

Aliquot anything you will use repeatedly. Every return to a stock bottle is an opportunity to contaminate it, and for anything enzymatic every freeze-thaw cycle costs activity. Split on receipt into volumes sized for how you actually work.

Filter, do not autoclave, anything heat-sensitive. This covers more reagents than people assume: enzymes are destroyed outright, glucose and other reducing sugars brown in the presence of amino groups, calcium-containing phosphate salines precipitate, and HEPES solutions discolour. Sterile filtration at 0.22 micron is the default for cell culture stocks.

Protect light-sensitive stocks from light. HEPES is the important case in a buffer context, because it generates hydrogen peroxide photochemically with riboflavin, but the instruction appears on many reagent labels and is worth following rather than rationalising.

Record the date the bottle was opened or thawed, not just the expiry printed on the label. Most reagent problems are age-in-use problems.

IPTG is a useful worked example, because it is stored differently from almost everything else on this page and the difference matters. Isopropyl beta-D-1-thiogalactopyranoside has a molecular weight of 238.30, so a 1 M stock is 238.3 g/L - in practice, 2.38 g dissolved in water and made up to 10 mL. It dissolves quickly in water and needs no heating.

1. Dissolve in cell-culture-grade or deionised water and make up to volume.
2. Sterilise by filtration through a 0.22 micron membrane. Never autoclave IPTG - heat degrades it.
3. Aliquot into small single-use volumes, typically 0.5 or 1.0 mL.
4. Store at -20 degrees C, where the stock keeps for up to a year. Avoid repeated freeze-thaw; that is what the small aliquots are for.

Working concentrations are typically 0.1-1 mM. Adding 1 mL of a 1 M stock to a 1 L culture gives a final concentration of 1 mM. For blue-white screening a 100 mM stock is the more convenient strength. The general point generalises beyond IPTG: for any small-molecule stock, check whether heat, light or freeze-thaw is the thing that degrades it, and store accordingly rather than defaulting everything to the fridge.

Osmolality. Mammalian cells tolerate roughly 260-320 mOsmol/kg, and buffer decisions move that number. Adding HEPES to a complete medium raises osmolality, because a 1 M stock is far more concentrated than the medium receiving it; media formulated with HEPES compensate by reducing sodium chloride, and media you supplement yourself do not. Repeated feeding with concentrated supplements over a long fed-batch run accumulates osmolality steadily. Base addition to counter lactate acidification adds sodium. None of these is individually large; together, across a two-week run, they are. If viability declines late in a culture that looks nutritionally adequate, measure osmolality before looking for anything more exotic.

A short diagnostic sequence for pH problems:

1. Medium stays purple or pink and never turns yellow, even in a dense culture. Bicarbonate/CO2 mismatch - most often a 3.7 g/L formulation running in a 5% CO2 incubator. Check the medium's specification against the incubator setting. Also check that the incubator's CO2 supply has not run out and that its sensor is calibrated.
2. Medium turns yellow within a day of feeding. Either the culture is too dense and should be passaged, or it is contaminated. Bacterial contamination acidifies fast and is usually accompanied by turbidity.
3. Medium turns yellow only in the outer wells of a plate. Evaporation concentrating the medium. Use a humidified incubator and consider a plate design that sacrifices the perimeter wells.
4. pH drifts alkaline during a benchtop procedure. Normal behaviour for a bicarbonate-only medium. Add 10-25 mM HEPES for that step, or use a HEPES-containing formulation.
5. pH reads correctly at room temperature but not at 37 degrees C. Buffer pKa is temperature-dependent; HEPES shifts by about -0.014 pH units per degree C. Titrate and verify at working temperature.
6. pH is correct but cells look stressed. Measure osmolality. pH and osmolality are separate specifications and a medium can be correct on one and wrong on the other.