Intron-spanning primers and genomic DNA

A primer pair that sits inside one exon will amplify genomic DNA as happily as it amplifies complementary DNA. The products are the same length. A quantification cycle cannot see the difference. An intron-spanning design is one way to make the genomic product larger, or to stop it forming in the extension time you use. This comparison is about when that design works, when the gene gives you no intron to span, and why a no-reverse-transcription well is still the evidence. The assay it protects is described in RT-qPCR for relative expression, on the path in from cells to a gene expression result.

DNA can wear an expression result

Genomic DNA is a frequent passenger in RNA preparations. Spin-column and bead methods reduce it. They do not sign an affidavit that it is gone. A DNase step reduces it further. The enzyme can fail, the incubation can be short, or the DNA can be protected in a clump the enzyme never saw. If your primers can copy what remains, the well reports DNA plus cDNA and the table calls the sum expression.

That inflation is not random. Samples that were harder to extract, or richer in DNA to begin with, pick up more false signal. A treatment that changes cell number or extraction yield can look like a transcriptional effect. The no-RT control is how you see it. Primer design is how you make the seeing easier.

A junction changes the product

Splicing removes introns from a primary transcript. Two exons that were far apart in the genome sit next to each other in the mature RNA. Place one primer in each exon, or place one primer so that it straddles the exon-exon junction, and the cDNA template is a short, continuous match.

On genomic DNA the same primers are separated by the intron. Two things can happen. If the intron is long relative to the extension time in the qPCR programme, the polymerase does not finish a genomic product and you see little or no signal from DNA. If the intron is short, a genomic product does form, and it is longer than the cDNA product. An endpoint gel can show both bands. A dye-based qPCR may still call a cycle from the long product. Probe assays and melt curves help only if you look at them. A cycle number alone will not tell you the amplicon was the short one.

A primer parked on the junction is pickier. It matches spliced RNA and mismatches genomic DNA at the intron interrupt. It is also picky about isoforms. An alternative splice that uses a different junction will not match, and you will report that isoform as absent. That can be what you wanted. It is a disaster if you thought you were measuring the gene.

Design against a named gene model. Ensembl shows exon boundaries for a chosen transcript. NCBI Primer-BLAST can test a pair against both the genome and the transcripts so you see the genomic amplicon size and the off-targets. Record the assembly. A primer that was perfect on an old transcript model can land inside a newly annotated exon arrangement.

Primers, DNase and the no-RT well

The reagent classes are ordinary oligonucleotides, a reverse transcriptase, a qPCR mix, and, if you use one, a DNase. Volumes and temperatures stay in the enzyme insert. Nothing here is a microlitre recipe. The molecular biology catalogue groups those classes. The extraction side, including silica columns that co-purify some DNA with RNA, is a different choice, covered in protecting RNA during extraction.

The no-RT control is the same RNA, the same primers, and no reverse transcriptase. Everything else follows the sample. A signal in that well means a DNA template was available. It does not, by itself, say whether the DNA is genomic or a previously amplified product. The no-template control separates those stories. A rising no-template well means contamination of the mix. A quiet no-template well and a rising no-RT well means the DNA came with the sample.

DNase is a reducer, not a control. Run the no-RT well after DNase if you want credit for the DNase. If the no-RT signal disappears after a proper digestion, the earlier signal was DNA the enzyme could reach. If it remains, the digestion failed or the template is being introduced after the digestion step.

Reporting practice for this control sits with the MIQE guidelines. Omit the well and a reader cannot tell expression from DNA, however elegant the junction.

Genes with introns, genes without

Compare three situations rather than declaring one design universal.

Both primers inside a single exon. cDNA and genomic DNA yield the same product. No gel and no melt will split them if the sequences match. This design is sometimes forced by a short transcript. It is never acceptable without a no-RT control. It is a poor choice when an intron exists and you simply did not look.

Primers in neighbouring exons, product spanning the intron. cDNA is short. Genomic DNA is long or absent. Prefer an intron long enough that genomic extension fails under the programme you actually run. Check that length on the gene model. A 70-base intron is not a moat. Confirm with a no-RT well anyway, because a failed assumption about length should be visible in the data.

One primer on the junction. Strong rejection of unspliced DNA, and strong dependence on one splice form. Use it when the isoform is the question. Do not use it as a casual way to represent "the gene" if several junctions exist.

When the gene lacks introns, the junction options disappear. Say so in the method. The no-RT control becomes the only honest check. Some references people like to use as normalisers are awkward in exactly this way, or have processed pseudogenes that look like spliced sequence in the genome. A pseudogene can give a short "spliced-looking" genomic product. Primer-BLAST against the genome is how you notice. A housekeeping symbol copied from a rodent paper is not a guarantee that the human locus has a convenient intron.

Bacterial expression studies do not get eukaryotic splicing for free. Treat those primers as the intronless case unless you have a different biological interrupt you have validated.

Three designs, three things a Cq can mean

The last row is the decision, not a defeat. Many real assays live there. They stay interpretable if the no-RT well is quiet relative to the sample, under a gap your protocol states, and if the no-template well is quiet.

A signal in the no-RT well

Read the pattern, then decide.

No-template flat, no-RT flat, sample rising. The design and the prep are consistent with a cDNA signal. You still have not proved isoform identity if you never checked size or melt.

No-template flat, no-RT rising close to the sample. Do not report expression. The well is measuring DNA. Digest again, redesign across a real intron, or accept that this assay cannot separate the two on this RNA.

No-template flat, no-RT rising but many cycles later than the sample. DNA is present and is not the bulk of the signal. Whether that side contribution is acceptable is a threshold in the protocol for this assay. Write the threshold down. Do not invent it after seeing the result you hoped for.

No-template rising. Stop. The plate is contaminated. A junction design does not explain a no-template band. Contamination control is the subject of PCR controls and contamination control.

A Cq in the no-RT well means DNA. It does not mean the reverse transcriptase was slightly leaky in a way you can ignore. Reverse transcriptase is absent from that well on purpose. Polymerase copying DNA is the activity you are looking at.

RNA-seq is not immune to DNA, but a gene count and a junction primer will not be fooled in the same way. Disagreement between a qPCR that lacks a no-RT control and a stranded, poly(A) count is often this artefact. Fix the control before you average the methods.

Not a clinical DNA test

This comparison does not approve a diagnostic RNA assay or a residual-DNA specification for a medical product. It is a research control strategy. Biosafety follows the sample. DNase buffers and qPCR mixes follow the safety notes on those reagents. An institutional biosafety decision is not hidden inside a primer.

Warm benches and leftover DNA

DNase and reverse transcriptase do not enjoy a long wait on a warm bench. In a hot laboratory, set up from a cold block and follow the temperature on the enzyme label rather than a room-temperature line written for a milder climate. A humid day also favours contamination: wet gloves, open plates, and aerosols from a previous amplicon. The no-template well is how that shows up. Keep completed PCR products away from the bench where RNA is reverse transcribed. Addgene's protocol collection is a reminder that separation of steps is part of the written method.

A power cut that warms a freezer of RNA does not create DNA, but a degraded RNA sample makes the DNA contribution relatively larger because the cDNA signal falls. After a known thaw, rerun the no-RT comparison. Do not trust an old gap.

What to say if primer design is part of the ask

State the gene model, whether the locus has introns, whether pseudogenes exist, and whether you need a gene-level or an isoform-level amplicon. Say that no-RT wells are part of the assay. If the question is really transcriptome-wide, a library discussion belongs on the mRNA sequencing enquiry reference, and a count comparison on the differential expression analysis enquiry reference. Those pages are prompts. A primer question should say so, rather than arriving disguised as a sequencing order.

The assay sits on the nucleic acid analysis pathway. Put the locus and the control plan in the quote request.