I've been learning about PCR and plasmids. I understand that the reason for having both a forward and a reverse primer is to extract and amplify the specific piece of DNA between these two primers.

What I'm struggling to understand is that why after the first replication, a fully circular DNA strand isn't formed and hence, why isn't a fully circular DNA strand formed for each replication.

For the following diagram , let the strands be labelled strand 1, strand 2, strand 3, strand 4 from top to bottom. (Diagram from Lehninger Principle of Biochemistry, 5th Ed).

2nd cycle of PCR

For example, in the diagram here. When cloning a plasmid I assume strand 1 is fully circular, so strand 2 (which was transcribed from strand 1) should also be fully circular and in the diagrams it should show DNA to the left of the site of interest.

Does the transcription stop somewhere along the line, preventing it from copying the whole strand? Or does the strand never join up into a circle, so that the DNA that would appear to the left of strand 2 has been copied, but doesn't join up to the site of interest so wasn't included in the diagram.

Thanks for your help.

  • $\begingroup$ The diagram clearly implies that strand 2 and 3 are products from the first round of replication. You're therefore correct that it would be more accurate if the diagram showed the grey parts on the left and right side respectively, but then having a small gap between the end of the grey part and the start of the blue part (to indicate the linearity of the product). However, that would make the diagram less illustrative and necessitate more explanation, so I think this is an acceptable inaccuracy. $\endgroup$
    – Armatus
    Apr 3, 2018 at 10:16

1 Answer 1


PCR produces linear DNA, not circular. You can replicate whole plasmids using PCR, but the DNA will still be linear, as if you cut the plasmid with a blunt-end restriction enzyme.

So why doesn't the whole plasmid get copied when using PCR? Let's assume the primers are designed to bind 1000 bases apart from each other. The first round of PCR will probably produce longer pieces of DNA, because the oligos can only bind to the plasmid and the Polymerase can keep going. But the polymerase will fall off at some point, either randomly or when the temperature is pushed up for the next round of PCR. But in round 2, there should be equal numbers of circular plasmid and linear copy. The proportion of linear copy will increase every round. As the linear copies increase, the oligos are more and more likely to bind to linear copies than to circular plasmid.

Binding to linear copies is why the PCR product size is controlled. Let's assume that a linear copy is 1500 bases, because the polymerase ran an extra 500 bases past the other oligo site on the plasmid. But when the reverse oligo binds that site on the linear copy, the polymerase will run 1000 bases and fall off the end of the linear copy. This new copy will have the forward and reverse oligo sites at the ends, so it's the right size. Any copies produced from this strand will also be the right size. As the reaction proceeds, the proportion of correct-sized strands increases, and create more correct-sized strands.

  • $\begingroup$ This makes good sense. Thank you. Is there any way to know how long before the polymerase falls off? $\endgroup$
    – meldob
    Apr 3, 2018 at 4:45
  • 1
    $\begingroup$ @meldob Most polymerases used in PCR are pretty good at staying on. I used to PCR entire 10,000 base plasmids on a regular basis. When setting up the PCR you choose the extension time based on the length you want to copy and the speed of your polymerase. So if your enzyme runs at 1000 bases per minute and you need to copy 2000 bases, use 2 minutes. Going longer probably won't hurt anything, but it wastes time. Going shorter will cut off the reaction before you get the entire sequence copied. $\endgroup$
    – user137
    Apr 3, 2018 at 5:06
  • $\begingroup$ We have different polymerases optimised for different purposes nowadays. Some focus on high fidelity (i.e. low error rate during copying) but sacrifice speed and range, others can copy 1kbp in less than 10s but make more errors and fail if the desired sequence is too long. $\endgroup$
    – Armatus
    Apr 3, 2018 at 10:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .