Puromycin for Mammalian Cell Selection

Puromycin is a natural product of Streptomyces alboniger. It is usually supplied as the dihydrochloride salt (CAS 58-58-2, molecular weight 544.43; the free base is 471.51), a pale yellow to white powder that is freely soluble in water.

Its mechanism is unusually clean, which is why it is a favourite selection agent. Puromycin is a structural analogue of the 3'-terminal aminoacyl-adenosine of tyrosyl-tRNA. Because it looks like the business end of a charged tRNA, it binds the ribosomal A site and the peptidyl transferase centre transfers the growing peptide chain onto it. At that point the analogy breaks down: in a real tRNA the amino acid is attached through an ester bond that can be hydrolysed and handed on, whereas in puromycin it is attached through a non-hydrolysable amide bond. The peptide is now covalently stuck to a molecule that cannot participate in the next round of elongation, so it is released from the ribosome as a truncated, puromycylated fragment.

Three consequences follow, and each of them matters practically:

• It acts on both prokaryotic and eukaryotic ribosomes, so it is effective in mammalian cells at low concentrations.
• It is fast. Protein synthesis is disrupted immediately rather than after a slow accumulation of damage. Non-resistant cells begin dying within 24-48 hours, and selection is usually complete within 2-5 days -- much faster than G418 or zeocin.
• It is used at low concentrations, typically 1-2 ug/mL, roughly two orders of magnitude below G418. That means less osmotic and chemical burden on the culture and a cheaper selection.

Outside selection work, the same chemistry is exploited in puromycylation-based methods for labelling nascent polypeptides, where a short puromycin pulse tags actively translating ribosomes.

Resistance is conferred by the pac gene, which encodes puromycin N-acetyltransferase (PAC), also from Streptomyces alboniger. PAC transfers an acetyl group from acetyl-CoA onto the free tyrosinyl amino group of puromycin. The acetylated product can no longer be accepted by the peptidyl transferase centre, so it is biologically inert.

Several practical points follow from this:

• Resistance is enzymatic and intracellular, not a pump or a target mutation. A cell must express enough PAC to detoxify the drug faster than it arrives. This makes selection stringency reasonably tunable by concentration -- raising the puromycin concentration genuinely selects for higher expression of the resistance cassette, and by extension for higher expression of a linked transgene.
• The gene is small, around 600 bp, which is a real advantage in lentiviral and AAV vectors where packaging capacity is tight. This is a major reason puromycin selection dominates lentiviral shRNA and CRISPR library work.
• Detoxification is local. Because PAC-expressing cells consume the drug, dense cultures of resistant cells can partially protect neighbouring non-resistant cells. This is the mechanism behind the most common selection failure, discussed below.
• In the literature the marker is written several ways -- pac, puroR, PuroR -- all referring to the same acetyltransferase.

Puromycin resistance is dominant and does not require a specific host genotype, unlike some metabolic selection systems (for example glutamine synthetase or DHFR selection, which depend on the host being deficient in the corresponding pathway).

The recommended working range for mammalian cells is 0.5-10 ug/mL, and most cell lines fall between 1 and 5 ug/mL. There is no universal number, and any concentration quoted for a cell line in a paper or catalogue should be treated as a starting point for your own kill curve, not as a value to adopt directly.

Commonly used starting points. Many adherent lines including HEK293 and its derivatives, HeLa and NIH/3T3 select well at around 1-2 ug/mL. Many CHO lines sit in a similar range, often 5-10 ug/mL for suspension-adapted lines in serum-free medium. Primary cells and some stem cell lines are considerably more sensitive and may select at 0.2-1 ug/mL. Some lines -- certain murine lines and cells with high endogenous drug-efflux activity -- need 10 ug/mL or more. Verify all of these by kill curve.

Typical starting concentrations by cell line. Every value below is a starting point for a kill curve, not a concentration to adopt directly. Reported optima vary between laboratories for the same named line, because density, serum content, medium and puromycin lot all shift the answer.

Cell line or type	Typical starting range	Notes
HEK293, HEK293T	1-2 ug/mL	Sensitive; selection often complete in 2-3 days
HeLa	1-2 ug/mL	Well-behaved, fast selection
A549, MCF-7 and many human epithelial lines	1-3 ug/mL
NIH/3T3 and many murine fibroblast lines	2-10 ug/mL	Murine lines commonly need more than human lines
CHO, adherent and suspension	5-10 ug/mL	Serum-free suspension often selects at the lower end
Jurkat and other suspension lymphoid lines	0.5-2 ug/mL	Frequently among the most sensitive
Primary cells	0.2-1 ug/mL	Slow growth reduces killing; extend the window rather than the dose
hESC and iPSC	0.2-1 ug/mL	Sensitive; start at the bottom of the range
Lines with high multidrug-efflux activity	10 ug/mL or above	Puromycin can be actively exported; consider another agent

Treat a value from this table, or from any published paper, as the centre of the dose series you test -- not as the answer.

What shifts the effective concentration:

• Cell density. This is the largest single variable and the most frequently overlooked. A dense culture consumes and dilutes the drug per cell, so the same nominal concentration is much less effective. Run your kill curve at the density you will actually use for selection.
• Serum content. Serum proteins bind puromycin to some extent, so a culture in 10% FBS generally needs more drug than the same cells in serum-free or reduced-serum medium. Moving a selection from serum-containing to serum-free medium usually requires re-optimisation downward.
• Doubling time. Puromycin kills cells that are translating. Slow-growing, confluent or quiescent cultures die more slowly and less completely, which is why selection should start on a subconfluent, actively dividing culture.
• Lot of puromycin. Potency varies between lots and degrades with storage. Re-run an abbreviated kill curve when opening a new lot.
• Medium pH. Puromycin is less stable at alkaline pH. A culture that has gone alkaline from poor CO2 control loses selection pressure faster than expected.

When to start selection. After transfection, allow 48-72 hours in normal growth medium before adding puromycin. The cells need time to express the resistance cassette; adding drug too early kills successfully transfected cells before PAC has accumulated. After lentiviral transduction, 24-48 hours is usually sufficient, since expression from an integrated provirus is established quickly.

A kill curve is a dose-response experiment that identifies the lowest concentration that kills 100% of untransfected cells within the intended selection window. Using more than this is not safer -- it increases the chance of killing genuinely resistant, lower-expressing clones and adds off-target stress.

Run this on parental, untransfected cells of the exact line, passage range and medium you will use for the real selection.

Day 0 -- plate

1. Harvest healthy, subconfluent parental cells in log-phase growth with viability above 90%.
2. Seed a 24-well plate at a density that will reach roughly 50-70% confluence the next day -- typically 2-5 x 10^4 cells per well for a fast-growing adherent line. Seed enough wells for 8 concentrations in duplicate or triplicate.
3. Match the seeding density to your planned selection density. A kill curve run on sparse cells will underestimate the concentration needed on a dense plate.
4. Incubate overnight in normal growth medium without puromycin so the cells attach and resume dividing.

Day 1 -- apply the dose series

5. Prepare a dilution series in complete growth medium. A standard series is 0 (untreated control), 0.25, 0.5, 1, 2, 5, 10 and 20 ug/mL. For a cell line expected to be sensitive, shift the series down (0 to 5 ug/mL); for a line expected to be resistant, extend it up.
6. Aspirate and replace the medium in each well with the corresponding puromycin concentration. The zero-drug control is essential -- it tells you whether any death you see is due to the drug or to the culture conditions.

Days 2-10 -- maintain and observe

7. Replace the selection medium every 2-3 days with freshly prepared puromycin at the same concentration. This is the step most often skipped and it invalidates the result: puromycin loses activity in medium at 37 C over a few days, and dying cells release debris and proteases that further reduce effective potency.
8. Examine daily by phase contrast. Record the proportion of rounded, detached or floating cells in each well. Photograph if you want a record.
9. Continue for up to 10 days, or until the untreated control reaches confluence and must be stopped.

Read the result

10. Identify the lowest concentration at which no viable cells remain, and note the day on which that was achieved. For puromycin this is usually day 2-5.
11. Confirm by trypan blue count or by a viability stain rather than by eye alone if the wells are ambiguous.
12. Use that concentration for selection. Many labs then maintain established stable lines at half the selection concentration to hold selective pressure without continuous heavy stress.

Sanity checks. If the untreated control is unhealthy, the whole curve is uninterpretable -- repeat it. If nothing dies even at 20 ug/mL, check that the drug was added, that the stock is not degraded, and that the cells are not already carrying a resistance marker from previous work. If everything dies including the control, look at the medium, the plate and the incubator before blaming the drug.

Timing. Add puromycin 48-72 hours after transfection or 24-48 hours after transduction. Non-resistant cells should be visibly dying within 24-48 hours and largely gone by day 3-5.

Refresh the drug. Change selection medium every 2-3 days throughout. Beyond maintaining potency, this removes the large amount of debris generated by mass cell death, which otherwise smothers surviving colonies.

Keep a kill control. Run untransfected parental cells in puromycin alongside every selection. When they are completely dead, selection pressure has been demonstrated for that specific experiment. Without this control, a plate of surviving cells is ambiguous -- they may be resistant, or the drug may simply have failed.

Density is the usual failure mode. Because PAC-expressing cells destroy puromycin, a dense population of resistant cells creates local zones where the drug concentration falls far enough for non-resistant cells to survive. Symptoms are a background lawn of untransfected cells that will not clear, and colonies that fail to expand cleanly. The fix is to keep the culture subconfluent during selection, split when it gets dense while maintaining the drug, and change medium often.

Expected duration. Selection is usually complete in 3-7 days. If you are isolating clones, continue selective pressure through colony expansion. For pooled populations, once the kill control is clear the pool can be expanded.

Maintenance. Established stable lines are often maintained at half the selection concentration, or cycled -- grown without drug and returned to selection periodically -- to reduce chronic stress. Cell lines can silence integrated cassettes over time, so verify expression periodically rather than assuming continued selection guarantees it.

A caution about interpreting survival. Puromycin selects for cells expressing PAC. It does not select for expression of whatever else is on your construct. Silencing of a downstream transgene while the resistance marker stays active is common, particularly with weak internal promoters or IRES-linked cassettes. Always confirm expression of the gene you care about by an independent assay.

Preparing stock. Puromycin dihydrochloride is freely soluble in water. The standard stock is 10 mg/mL in sterile water or in HEPES-buffered saline, sterile-filtered through 0.22 um. Ready-made 10 mg/mL solutions are widely supplied and avoid weighing a hazardous powder.

pH matters. Puromycin is most stable in slightly acidic to neutral solution, around pH 6.0-7.0, and degrades appreciably under alkaline conditions. Do not prepare stock in an alkaline buffer, and avoid storing it in medium that has gone alkaline from CO2 loss.

Storage.

• Powder: -20 C, desiccated and protected from light.
• Aqueous stock at 10 mg/mL: -20 C, protected from light, in single-use aliquots. Stable for roughly a year under these conditions.
• Short-term working stock: 2-8 C for up to about a month.
• Avoid repeated freeze-thaw. Each cycle costs potency, and a stock that has been thawed a dozen times is a common hidden cause of failed selection. Aliquot at 100-500 uL on first thaw.

In culture medium. Puromycin loses activity in complete medium at 37 C over a period of days. This is the reason for the every-2-3-days medium change during selection and kill curves. Do not prepare a large batch of selection medium and use it over two weeks.

Light. Protect stocks from prolonged light exposure. Wrapping tubes in foil is sufficient.

Handling. Puromycin is a protein synthesis inhibitor and is toxic. Handle powder in a fume hood or biosafety cabinet with gloves and eye protection to avoid inhaling dust, and dispose of solutions as chemical waste. Weighing pre-made solution is preferable to handling powder wherever possible.

Puromycin is the default for most stable line work because it is fast, cheap, effective at low concentration and carried by a small gene. It is not always the right choice.

Reach for puromycin when you want the shortest possible selection, you are working with a viral vector where insert size is constrained, or you are selecting a pooled library where speed reduces drift.

Reach for something else when:

• You need a second or third marker. Sequential or simultaneous selection for multiple constructs requires orthogonal agents -- puromycin plus blasticidin plus hygromycin is a common trio, since their resistance mechanisms do not overlap.
• Your cell line is intrinsically puromycin-resistant or requires such a high concentration that off-target toxicity becomes a concern.
• You need very stringent, slow selection for high-expressing clones. G418 selection is slower and some groups find it gives better enrichment of stable high expressers for that reason.
• You are working with cells that have high multidrug-efflux activity, where puromycin can be actively exported.

A note on zeocin, which behaves differently from the others: it is a DNA-cleaving glycopeptide rather than a translation inhibitor, it requires low-salt medium at pH 7.0-7.5 for full activity, and it is strongly light-sensitive. Failed zeocin selection is very often a medium-composition problem rather than a dose problem.

Whichever agent you use, the kill curve requirement is identical. Every agent, every cell line, every new lot.