Written and maintained by CASRAI Editorial Board
Last updated
Transfection is the deliberate introduction of nucleic acid into eukaryotic cells (see cell culture basics for the underlying workflow). It sounds like one technique; it is really a family of quite different methods with different costs, different failure modes, and — a point most protocol pages omit entirely — different regulatory obligations before you are allowed to start.
This guide covers how to choose a method, how to optimise it, the controls that make an efficiency number meaningful, and the institutional biosafety approval that applies to most of this work.
First, three words that get confused
| Term | Meaning |
|---|---|
| Transformation | Uptake of nucleic acid by bacterial or plant cells. Using it for mammalian cells is incorrect — in mammalian biology “transformation” means becoming cancerous, which is a genuinely confusing collision. |
| Transfection | Introduction of nucleic acid into eukaryotic cells by non-viral means — chemical or physical. |
| Transduction | Introduction of nucleic acid using a viral vector. |
People routinely say “viral transfection”, which conflates the last two. The distinction is not pedantry: it determines which biosafety framework applies to you.
Transient versus stable
Transient expression peaks within roughly a day or two and is lost as the episomal DNA is diluted by division and degraded. Fine for reporter assays, short knockdowns, and protein production runs.
Stable expression requires the construct to integrate into the genome (or be maintained episomally), followed by selection — typically antibiotic selection over one to several weeks — to isolate cells that kept it. Necessary for long-term studies and cell-line generation, and far more work.
Decide which you need before choosing a method, because it constrains the choice.
Choosing a delivery method
Chemical
- Lipid-based (lipofection) — cationic lipids complex with nucleic acid and fuse with the membrane. The default for most adherent cell lines: easy, effective, no special equipment. Reagent cost adds up, and it is often poor on primary cells.
- Calcium phosphate — very cheap, works well on robust lines like HEK293, and is notoriously sensitive to pH and precipitate formation. Reproducibility between operators is its weakness.
- Cationic polymers (PEI) — cheap and scalable, which is why it dominates large-volume protein production.
Physical
- Electroporation / nucleofection — a pulsed field transiently permeabilises the membrane. This is the workhorse for the cells lipofection cannot reach: primary cells, suspension cells, stem cells, immune cells. Costs are higher (instrument, proprietary cuvettes and buffers) and cell viability takes a real hit, so you are trading survival for delivery.
- Microinjection — one cell at a time. Unbeatable control, unusable throughput.
- Biolistic (“gene gun”) — niche, mostly plant and tissue work.
Viral (transduction)
- Lentivirus — infects dividing and non-dividing cells, integrates, gives stable expression. The standard answer for hard-to-transfect and primary cells.
- AAV — low immunogenicity, largely non-integrating, strong in vivo track record; limited cargo capacity.
- Adenovirus — very high transient expression, non-integrating, more immunogenic.
- Retrovirus — integrates, but only in dividing cells.
Viral delivery buys efficiency in exactly the cells where nothing else works. It costs you a substantially heavier biosafety and approval burden, covered below.
The decision, in practice
Four questions settle it:
- What cell type? An easy adherent line (HEK293, HeLa, CHO) — lipofection. Primary, suspension, stem or immune cells — electroporation or viral.
- Transient or stable? Stable pushes you toward integrating viral vectors or a selection workflow.
- How big is the cargo? AAV in particular has a hard capacity limit.
- What can you actually get approved and contained? A method requiring BSL-2+ containment your lab does not have is not an option, however well it works on paper.
Optimisation: the variables that actually matter
- Cell confluency at transfection. The most under-controlled variable in the whole procedure. Too sparse and you get poor uptake and toxicity; too dense and efficiency collapses. Fix a confluency and stick to it — this alone explains a great deal of week-to-week variation.
- Nucleic acid to reagent ratio. Titrate it for your cell line rather than inheriting a number from a paper. More DNA is not better; past an optimum, efficiency falls and toxicity climbs.
- DNA quality. Endotoxin-free preparations matter for sensitive and primary cells. An A260/280 ratio around 1.8 is the usual target.
- No antibiotics during transfection. This is the classic avoidable killer: penicillin/streptomycin plus a transiently permeabilised membrane means antibiotic entering the cytoplasm, and the resulting death gets misread as reagent toxicity. Transfect in antibiotic-free medium.
- Serum — some reagents tolerate it during complex formation, many do not. Follow the specific reagent’s guidance rather than a generic protocol.
- Passage number. High-passage cells transfect worse and drift phenotypically. Record it.
Controls and measuring efficiency honestly
| Control | Purpose |
|---|---|
| Untransfected cells | Baseline viability (see cell counting and viability) and autofluorescence. |
| Mock (reagent, no nucleic acid) | Separates reagent toxicity from construct effect. Frequently skipped, and it is the one that tells you whether your phenotype is biology or poisoning. |
| Reporter construct (e.g. GFP) | Measures delivery efficiency independently of whether your construct works. |
| Empty vector | Controls for the vector backbone itself. |
On measurement: Flow cytometry gives a real per-cell efficiency; eyeballing a fluorescence image does not. A field of view chosen because it looks good is not a measurement. Report how you measured it, and report viability alongside efficiency — 90% delivery into cells that are 40% dead is not a good result, and quoting efficiency without viability hides that.
Efficiency also varies enormously by cell type. A number from HEK293 tells you nothing about what to expect in primary neurons, so do not carry expectations across cell types.
The compliance layer most protocols omit
This is where a research-administration perspective adds something a vendor protocol will not, and where researchers most often discover a problem after starting.
- Recombinant DNA work generally requires Institutional Biosafety Committee registration. Under the NIH Guidelines for Research Involving Recombinant or Synthetic Nucleic Acid Molecules, institutions receiving NIH funding must review this work through an Institutional Biosafety Committee, and the obligation typically extends across the institution rather than only to the NIH-funded project. Approval must be in place before work begins — it is not a retrospective formality, and starting without it is a reportable non-compliance.
- Viral vectors raise the containment bar. Lentiviral work is commonly handled at BSL-2 or BSL-2 with enhanced practices, depending on the vector generation, the insert, and whether the construct carries oncogenic or toxin sequences. That risk assessment is the IBC’s to make, not the bench scientist’s to assume.
- The insert matters as much as the vector. A construct expressing an oncogene, a toxin, or a gene that could alter host range changes the risk assessment even in a familiar vector.
- Occupational health. Lentiviral work usually triggers exposure-response planning and may involve health surveillance; sharps injuries during viral work are treated as exposures.
- Training and documentation. Expect documented biosafety training, a written protocol, and inventory records for vectors.
Practical scheduling point for anyone planning a grant timeline: IBC review runs on a committee meeting cycle, so build weeks, not days, into the plan — and start the registration while you are still optimising, not after.
Troubleshooting by symptom
Low or no efficiency
Check confluency first, then DNA quality and quantity, then the reagent ratio. Confirm the construct actually expresses — a promoter inactive in your cell type produces a perfect-looking transfection with no signal. Verify the plasmid by restriction digest or sequencing rather than trusting the tube label. If the cell type is intrinsically hard, no amount of lipofection optimisation will rescue it; change method.
High toxicity
Reduce reagent, shorten exposure and change medium afterwards, remove antibiotics, and check endotoxin. Compare against the mock control to establish whether the toxicity is the reagent or the construct.
Efficiency varies week to week
Almost always confluency, passage number, or a new reagent lot. Fix and record all three. This is also where mycoplasma contamination shows up — it depresses transfection efficiency and is invisible without testing.
Good transient expression, no stable clones
Check that the selection antibiotic concentration was established by a kill curve on your specific line, that the resistance cassette matches the drug, and that you selected long enough. Integration is rare — most transiently expressing cells were never going to become stable clones.
Frequently asked questions
What is the difference between transfection and transduction?
Transfection introduces nucleic acid by chemical or physical means; transduction uses a viral vector. The distinction determines which biosafety framework and containment level apply.
Why should I not use antibiotics during transfection?
Transfection transiently permeabilises the membrane, letting antibiotics such as penicillin/streptomycin into the cytoplasm where they are toxic. The resulting cell death is commonly misattributed to the transfection reagent.
What transfection efficiency should I expect?
It depends entirely on cell type and method, and any single number is misleading out of context. Measure it in your own system by flow cytometry with a reporter construct, and always report viability alongside it.
Do I need IBC approval to transfect cells?
For recombinant or synthetic nucleic acid work, generally yes — and approval must be in place before you begin. Check with your institutional biosafety office; requirements and exempt categories vary, and viral vectors almost always require review.
Which method works for primary cells?
Lipofection usually performs poorly on primary cells. Electroporation/nucleofection or viral transduction are the realistic options, with viral generally giving the highest efficiency at the cost of a heavier approval and containment burden.
Why did my transfection efficiency drop suddenly?
Check confluency, passage number and reagent lot first. Then test for mycoplasma — it reduces transfection efficiency and cannot be seen without testing (see aseptic technique).
