Mycoplasma Testing Methods Compared: PCR, Culture, Indicator Cell Culture and Luminescence

Mycoplasma is the contamination that does not announce itself. Bacterial and fungal contamination turns the medium cloudy within a day or two; mycoplasma leaves a flask looking entirely normal while altering proliferation rate, metabolism, gene expression and drug response in the cells you are about to publish on. The only way to know is to test — and the four methods in routine use differ by roughly three orders of magnitude in sensitivity and by four orders of magnitude in turnaround, from twenty minutes to four weeks.

This guide compares direct agar and broth culture, indicator cell culture with a DNA stain (Hoechst 33258 or DAPI), nucleic acid amplification (PCR/qPCR) and enzymatic luminescence on the four axes that actually drive method selection — sensitivity, turnaround, cost and regulatory acceptability — then covers the decision most pages skip: what to do when a line comes back positive. Decontaminating an infected line is possible. It is usually the wrong choice, and the reasoning behind that default is set out in full below.

Why routine inspection cannot find mycoplasma

Mycoplasmas are bacteria of the class Mollicutes that lack a cell wall. Three consequences follow directly from that single structural fact, and together they explain the entire testing problem:

• They are invisible on inspection. No cell wall means no turbidity at the densities mycoplasmas reach, no rapid pH crash, and no distinct odour. Under a standard inverted light microscope at the magnifications used for routine culture checks, an infected flask looks like a healthy flask.
• They pass filters intended to sterilise. Mycoplasmas are among the smallest self-replicating organisms known, and a proportion of cells are small enough to pass through the 0.22 µm membranes routinely used to sterile-filter media and supplements. Filtration reduces the load; it does not guarantee exclusion.
• They are intrinsically resistant to whole antibiotic classes. Penicillins, cephalosporins and other beta-lactams act on cell wall synthesis. With no cell wall to target, they do nothing. A lab running routine penicillin/streptomycin is not protected against mycoplasma, and may be worse off — routine antibiotics mask the low-level bacterial contamination that would otherwise have revealed a lapse in aseptic technique.

A small number of species account for the large majority of cell culture contaminations — commonly Mycoplasma orale, M. hyorhinis, M. arginini, M. fermentans, M. salivarium and Acholeplasma laidlawii. The human oral flora species in that list point at the dominant transmission route: operator aerosol from talking or breathing over open vessels. The others point at contaminated bovine serum and at cross-contamination between lines sharing a cabinet session.

Published prevalence estimates vary widely and should be treated as indicative rather than precise. Figures commonly cited in the literature fall in the region of 5–35 % of continuous cell lines in active use, but the number depends heavily on the population sampled (academic labs test worse than industrial cell banks), the detection method used, and the era of the survey. Treat the range as evidence that the base rate is non-trivial, not as a figure to quote as fact.

The four methods compared

The table below summarises the practical differences. The figures that follow are drawn from compendial texts, peer-reviewed method comparisons and, where indicated, vendor kit documentation; sensitivity in particular varies materially between sources, and the section after the table explains why.

Why the sensitivity numbers disagree between sources

A recurring source of confusion is that method comparisons do not all measure the same quantity. Three distinct units circulate, and they are not interchangeable:

• CFU/mL counts colony-forming units — that is, viable organisms capable of growing on the culture medium used. It is the unit the compendial culture method works in.
• Genome copies (GC) per mL or per reaction counts DNA targets. It is the unit a PCR assay actually measures.
• Organisms per mL is sometimes used loosely for microscopy-based methods and rarely corresponds cleanly to either of the above.

The gap between GC and CFU is the crux. Mycoplasmas grow in chains and clumps and a proportion of genome-bearing cells are not culturable at any given moment, so one CFU generally corresponds to more than one genome copy. This is why a PCR method can be described as detecting “below 100 genome copies/mL” and “10 CFU/mL or better” in the same breath without contradiction — and why the ratio between the two matters enough that the European Pharmacopoeia now sets an acceptance criterion for it (below). When a vendor quotes a sensitivity figure, check which unit it is in and whether it refers to the reaction or the sample. A kit sensitive to 20 genome copies per PCR reaction is not thereby sensitive to 20 copies per mL of your supernatant; the dilution introduced by extraction and input volume sits between the two.

Where a specific figure is load-bearing for your validation, take it from the method’s own validation report against a characterised reference preparation, not from a comparison table — including this one.

What the compendia actually require

If your work is purely academic, the compendial chapters below are not binding on you, but they are still the best available statement of what a rigorous test looks like and are worth reading as a design reference. If you are producing a biological product, a viral vaccine or an advanced therapy, they are the requirement.

USP <63> Mycoplasma Tests

The United States Pharmacopeia’s general chapter <63> Mycoplasma Tests sets out two procedures: the agar and broth media procedure and the indicator cell culture procedure. The structural point to understand is that these are not alternatives to choose between — they are complementary, because the culture procedure alone misses non-cultivable species and the indicator procedure alone is less sensitive to those that do grow readily.

USP <63> permits a validated nucleic acid amplification technique or an enzymatic activity-based method in place of those procedures, on one condition: the alternative must be shown to be comparable to both compendial methods. That word “both” carries the entire regulatory burden of switching to PCR. A comparability study against culture alone does not satisfy it.

Ph. Eur. 2.6.7 Mycoplasmas — substantially revised, in force 1 April 2026

The European Pharmacopoeia’s general chapter 2.6.7 Mycoplasmas underwent a significant revision that is recent enough that a good deal of published guidance still describes the superseded text. The European Pharmacopoeia Commission adopted the revised chapter, along with 11 monographs referring to it, at its 181st session in March 2025. The texts were published in Ph. Eur. Issue 12.2 in October 2025 and entered into force on 1 April 2026. The changes that matter operationally:

• A risk-assessment-based, less prescriptive strategy. The revised chapter moves away from prescribing a fixed method towards requiring a justified approach. Eleven monographs were correspondingly revised to remove specific instructions on which method to apply or what volume to use.
• Culture and indicator cell culture used conjointly — or NAT as the alternative. The chapter specifies that both the culture method and the indicator cell culture method (or, alternatively, a NAT method) should be used together to ensure detection of both cultivable and non-cultivable mycoplasmas, unless a monograph prescribes otherwise or unless justified by a risk assessment and authorised by the competent authority.
• Sample composition is now explicit. Wherever possible, the sample should contain both cells and supernatant. This is the single most commonly botched step in in-house testing — see the sampling section below.
• A GC/CFU acceptance criterion. A limit of less than 10 is proposed as the acceptance criterion for the genome copies to colony-forming units ratio of reference preparations, unless otherwise justified. This is the formal answer to the units problem described above.
• Strain selection is risk-based. Suitable strains may be selected from the proposed list, and additional strains may be used based on a risk assessment that takes account of the product type and the manufacturing process.

The widely applied criterion for a NAT method replacing culture is that it demonstrates a limit of detection of 10 CFU/mL or below. Because the chapter has just been revised, verify the exact current wording against the in-force Ph. Eur. text rather than against a secondary summary — including this one — before writing it into a validation protocol.

FDA guidance on cell substrates

For viral vaccine work in the United States, the relevant FDA document is the February 2010 guidance Characterization and Qualification of Cell Substrates and Other Biological Materials Used in the Production of Viral Vaccines for Infectious Disease Indications, which addresses mycoplasma testing as part of cell substrate qualification. For the indicator cell culture procedure, Vero cells are the usual choice, with other validated substrates such as CV-1 or 3T6 permitted. After a typical 3–5 day incubation the indicator cells are stained with a DNA-binding fluorochrome — Hoechst 33258 is the classic reagent, with DAPI used similarly — and a positive is called on extranuclear fluorescence: punctate or filamentous signal over the cytoplasm and around the nuclear periphery, distinct from the bright, uniform nuclear stain of the indicator cells themselves.

Choosing a method: the decision rule

Method selection collapses to one question — what is the test for?

Routine in-house QC of working cultures

Use PCR, or luminescence backed by periodic PCR. The purpose here is frequent, cheap surveillance that catches an incursion before it spreads through the incubator and into a dataset. A four-week culture result is operationally useless for this: by the time it lands, the cells have been passaged five times, shared with two colleagues and frozen down.

Enzymatic luminescence is the cheapest and fastest option and is well suited to screening every line every month. Its weakness is that it is a vendor-defined qualitative screen without a compendially characterised limit of detection, so a negative is weaker evidence than a negative PCR. A workable compromise many labs adopt: luminescence monthly on everything, PCR quarterly and on anything entering or leaving the lab.

Qualifying a new line, or anything arriving from outside

Use PCR, and quarantine until it clears. Every incoming line — from a collaborator, a repository, or a returning postdoc’s cryovial — is handled in a separate cabinet session, ideally a separate incubator, until tested. This is the single highest-yield control point, because incoming lines are how most labs acquire the problem in the first place. Test at the same time as you handle authentication of key resources: mycoplasma testing and STR profiling are the two halves of cell-line quality control, and a line that is clean but misidentified is exactly as useless as one that is correctly identified but infected.

Master and working cell bank qualification

Use the full compendial battery. A cell bank is tested once and drawn on for years; there is no argument for economising on the test that qualifies it. This is where the culture-plus-indicator-cell pairing earns its cost, and where the four-week turnaround is affordable because it runs in parallel with everything else in bank qualification.

GMP release testing of a biological product

Use the compendial method, or a NAT validated for comparability to it. The choice is largely made for you by USP <63> and Ph. Eur. 2.6.7. The genuine decision is whether to invest in validating a rapid NAT method, and for most products that turns on shelf life.

Short-shelf-life products — ATMPs and cell therapies

NAT is not optional here; it is the only physically possible answer. An autologous cell therapy with a shelf life measured in hours or days cannot wait 28 days for a culture result. This is the principal driver behind the maturation of rapid mycoplasma NAT methods and behind the Ph. Eur. revision’s shift towards a risk-based, method-agnostic framing. The validation burden is real — comparability to both compendial methods, characterised reference preparations, and demonstrated absence of matrix inhibition in your specific product — but the alternative is releasing on an incomplete test or not releasing at all.

Sampling: where in-house testing usually goes wrong

A confident negative result from a badly taken sample is worse than no test, because it retires the question. Four rules cover most of the failure modes:

1. Sample both cells and supernatant. The revised Ph. Eur. 2.6.7 now says this explicitly. Many mycoplasmas adhere tightly to the cell membrane; testing clarified, spun-down supernatant alone can under-represent the load substantially.
2. Culture antibiotic-free for at least two passages before testing. Antibiotics active against mycoplasma — and several of the agents labs use routinely are — suppress the organism below detection without clearing it. Testing under antibiotic cover is a reliable way to generate a false negative on a genuinely infected line.
3. Test a near-confluent culture at least 48 hours post-feed. You want the highest achievable mycoplasma density, which means letting the medium become conditioned rather than sampling a freshly fed, sparse flask.
4. Run the controls, every time. A PCR run without a positive control cannot distinguish a true negative from a failed reaction, and one without an internal amplification control cannot distinguish a true negative from an inhibited sample — and cell lysates are inhibitory matrices. The same discipline that governs any quantitative qPCR or RT-qPCR run applies here.

You have a positive. Now what?

Step one: confirm before you act

Do not discard a line, and do not start antibiotics, on a single positive from a screening assay. A single PCR positive can be carryover contamination from a positive control or a previous high-titre sample — PCR’s sensitivity is exactly what makes it vulnerable to this. Confirm with a fresh sample, ideally taken by a different operator, and preferably by a second method with a different failure mode: a positive by both luminescence and PCR, or by PCR and indicator cell culture, is a confirmed positive.

At the same time, test everything else that shared a cabinet session, an incubator shelf, a bottle of medium or a serum lot with the positive line. Mycoplasma rarely stays in one flask. Assume the incursion is wider than the one positive result until testing says otherwise.

Step two: the decontaminate-versus-discard decision

This is the decision practitioners actually agonise over, and it has a clear default. The default is discard, and it is the right default in most cases. Treatment is the exception, justified only by irreplaceability.

The reasoning is not that eradication does not work — quinolone and macrolide/tetracycline-based regimens do clear mycoplasma from many lines. It is that a successfully treated line is not the same line you started with, and you generally cannot prove what changed. Specifically:

• The contamination has already altered the culture. Mycoplasma consumes arginine and other nutrients, shifts metabolism, and changes proliferation and gene expression. Eradicating the organism does not undo the selective pressure that has been running on the population for however many passages the infection went undetected.
• The treatment itself is selective and potentially mutagenic. Fluoroquinolones inhibit bacterial type II topoisomerases; at the concentrations and durations used for eradication they are not inert towards the eukaryotic cells either. Weeks of antibiotic exposure applies its own selection to the population.
• Suppression is easily mistaken for clearance. An apparent cure that is really suppression re-emerges weeks after treatment stops, typically after the line has already been returned to general use and shared.
• Provenance becomes unexplainable. If the line is used in published work, you now have a reagent with an infection history, an antibiotic exposure history, and an unknown amount of phenotypic drift — and reviewers, increasingly, ask. A treated line is a permanent footnote on every dataset generated from it.
• Replacement is usually cheap by comparison. Thawing an early, tested, uninfected vial from your own liquid nitrogen stock, or re-purchasing from a repository, costs a fraction of a multi-week eradication plus re-validation — and yields a line with clean provenance.

Work through these questions in order:

1. Do you have an earlier, tested-clean frozen stock? If yes, discard the infected culture, thaw the clean vial, quarantine and test it before general use. Stop here. This is why disciplined cryopreservation with a properly stratified master and working bank is a contamination control, not just a storage convenience.
2. Can the line be re-obtained from a repository or the originating lab? If yes, discard and re-obtain. A repository vial arrives with authentication and testing documentation attached, which your treated line never will.
3. Is the line genuinely irreplaceable? A patient-derived primary line, a unique engineered clone with no surviving clean stock, an isolate from a closed cohort. Only this answer justifies treatment.
4. If you treat: is it worth the containment burden? An infected line under treatment must be handled as contaminated for the entire multi-week period — dedicated incubator or physically segregated shelf, dedicated media and consumables, last in the cabinet session, cabinet decontaminated after. If you cannot commit to that, treatment will simply spread the contamination while it runs.

Step three: if you do treat, define clearance before you start

Eradication regimens in common use combine agents that mycoplasmas are susceptible to and that eukaryotic cells tolerate — typically a fluoroquinolone such as ciprofloxacin or enrofloxacin, or an alternating macrolide/tetracycline regimen. Follow the specific product’s protocol for concentration and duration rather than improvising; these are the parameters the supplier has actually characterised.

What matters more than the choice of agent is the clearance criterion, set before treatment begins:

• Complete the full treatment course — do not stop early because the cells look unhappy, which they will.
• Return the line to antibiotic-free medium and passage it at least three times, over a minimum of two to three weeks, before testing. Testing immediately post-treatment measures suppression, not cure.
• Test at least twice, on separate occasions, by a method with a characterised limit of detection — PCR rather than luminescence alone.
• Re-authenticate by STR profiling after clearance. You have just put the population through weeks of selection; confirming it is still the line you think it is is a small additional cost on top of everything already spent.
• Re-bank immediately once clean, and record the infection and treatment history in the line’s documentation so it travels with the cells.

If two clearance tests do not both come back negative, discard. Continuing to treat a line that has failed one eradication attempt has a poor record and consumes months.

Preventing the next one

Detection is a control of last resort; the controls that actually reduce incidence sit upstream of it.

• Quarantine everything incoming. Separate cabinet session and, where possible, separate incubator until a clean test is in hand. No exceptions for lines from trusted collaborators — that is precisely where the false confidence lives.
• One line open at a time. Never two open vessels of different lines in the same cabinet simultaneously. This controls cross-contamination and mycoplasma spread with the same measure.
• Dedicated media bottles per line. Shared bottles turn a single infected flask into a lab-wide event. Aliquot rather than repeatedly re-entering a shared bottle.
• Do not run prophylactic antibiotics routinely. They do not prevent mycoplasma, and they mask the bacterial contamination that would otherwise tell you your technique has slipped. Reserve them for genuinely justified situations such as primary isolation from tissue.
• Qualify your serum and media supply. Bovine serum has historically been a documented source; supplier certificates of analysis and consistent lot practice matter. This is one of the practical arguments for chemically defined and serum-free formulations where the biology allows.
• Test on a schedule, and record it. Monthly for lines in active use, on receipt, before banking, before any experiment that will be published, and after any suspected breach. A testing log that shows the line was clean at the time the data were generated is the record that answers a reviewer’s question.
• Maintain the cabinet. Certification, correct working practice and airflow integrity in a Class II biosafety cabinet are what make single-line working effective rather than nominal.

Reporting mycoplasma status in publications

Journal and funder expectations have converged on cell line quality control being reported, not assumed. Mycoplasma testing status and cell line authentication both appear in the reagent sections of structured reporting checklists — see the Nature Reporting Summary and MDAR checklist for how this is operationalised at submission. Report, at minimum: the method used, the date relative to the experiments, and the result. “Cells tested negative for mycoplasma” with no method and no date is not a meaningful statement, since a negative luminescence screen taken under antibiotic cover eighteen months earlier would satisfy it. Where a line is identified by a persistent identifier, an RRID makes the specific reagent traceable in a way a line name alone does not.

Frequently asked questions

Which mycoplasma test is the most sensitive?

Direct culture on Mollicutes-selective media remains the reference standard for cultivable species and is credited with detection down to around 1 CFU/mL, but it misses non-cultivable species entirely — which is why compendial testing pairs it with an indicator cell culture procedure rather than relying on it alone. Validated NAT methods are cited at 1–10 CFU/mL or below 100 genome copies/mL, close enough to culture that Ph. Eur. permits substitution at a demonstrated limit of detection of 10 CFU/mL or better. In practice the most sensitive method for your sample is the one that has been validated against your matrix, because inhibition in a cell lysate can cost a PCR assay an order of magnitude of real-world sensitivity.

How often should cell lines be tested for mycoplasma?

There is no single mandated frequency for research-use cultures. A defensible routine is: on receipt of any incoming line, before and after banking, monthly for lines in continuous active use, before generating data intended for publication, and immediately after any suspected contamination event elsewhere in the lab. Regulated cell banks and product release testing follow the schedule set by the applicable monograph and by your validated control strategy, not by a general rule of thumb.

Can PCR replace the compendial culture method?

Yes, subject to validation. USP <63> permits a validated nucleic acid amplification technique or enzymatic activity-based method provided it is shown comparable to both the agar and broth procedure and the indicator cell culture procedure. Ph. Eur. 2.6.7 similarly permits NAT as an alternative, with the commonly applied criterion being a demonstrated limit of detection of 10 CFU/mL or below. The validation, not the technique, is what makes it acceptable.

Does DAPI or Hoechst staining alone detect mycoplasma reliably?

Direct DNA staining of the test culture is the least sensitive of the routine options and is best regarded as a confirmatory or visual-evidence method rather than a screening one. Its sensitivity improves substantially when the sample is first amplified on an indicator cell monolayer — which is what the compendial indicator cell culture procedure does, using Vero or another validated substrate stained with Hoechst 33258 and read for extranuclear fluorescence. Reading the stain reliably requires experience; cellular debris and apoptotic fragments are the common false-positive traps.

Should I treat a mycoplasma-positive cell line or discard it?

Discard, unless the line is genuinely irreplaceable. Eradication regimens work often enough to be tempting, but a treated line carries an unresolvable provenance problem: the infection has already applied selection to the population, the treatment applies more, and suppression is readily mistaken for cure. If you hold an earlier tested-clean frozen stock, or the line can be re-obtained from a repository, discarding is both cheaper and scientifically cleaner. Treat only when neither is true, and define your clearance criteria — three antibiotic-free passages and two separate negative PCR tests, plus re-authentication — before you start.

Why do penicillin and streptomycin not prevent mycoplasma contamination?

Penicillin acts on bacterial cell wall synthesis, and mycoplasmas have no cell wall. Streptomycin has limited activity against Mollicutes at the concentrations used in culture. Routine pen/strep therefore offers no meaningful protection against mycoplasma while actively masking the bacterial and fungal contamination that would otherwise have warned you about a technique or environment problem.

Are luminescence kit sensitivity figures independently validated?

Generally not in the sense that matters. Performance figures published in kit documentation are vendor data, generated by the manufacturer against its own reference material, and should be attributed as such rather than treated as independent validation. Peer-reviewed head-to-head comparisons of commercial mycoplasma assays exist and are the better source when a performance figure is load-bearing — but even those are typically run on a defined panel of strains in a defined matrix, which may not be your matrix.

Related reading

• Cell culture basics for new lab members — the wider context this test sits inside
• Aseptic technique in the lab — the practices that prevent the contamination in the first place
• qPCR and RT-qPCR — controls, inhibition and interpretation for the assay most mycoplasma NAT methods are built on
• Cryopreservation of cells — the clean frozen stock that makes “discard” an easy decision
• Types of cell culture media and FBS heat inactivation — supply-side contamination risk
• Transfection methods and controls — where an undetected infection most visibly distorts results
• Lab equipment cluster hub — instrumentation and method-selection guides across the analytical bench

Sources and verification note

Compendial requirements above are stated from the European Directorate for the Quality of Medicines & HealthCare’s published announcement of the revised Ph. Eur. general chapter 2.6.7 (adopted at the 181st EPC session, March 2025; published Ph. Eur. 12.2, October 2025; in force 1 April 2026) and from published descriptions of USP <63> Mycoplasma Tests. The FDA document referenced is the February 2010 guidance on characterization and qualification of cell substrates for viral vaccine production. The full text of both pharmacopoeial chapters is behind a subscription and was not read directly for this page — where a specific incubation period, volume or acceptance criterion will be written into a validation protocol or a regulatory submission, read the in-force compendial text itself rather than relying on any secondary summary, this one included. Sensitivity figures attributed to commercial kits are vendor-reported and are identified as such throughout; they are not independent validation.