Many people add the selective antibiotic prior to plasmid isolation. This sounds reasonable, but actually does not help too much. The normal high-copy-number plasmids are essentially never lost from cultures. You can see this immediately, when you try to cure them; this can mean a lot of effort.

What really is helpful is to start all experiments from a single colony, which has been tested for the relevant features, e.g. resistance. Subculturing over subsequent liquid cultures of course selects for those bacteria that grow faster. State of the art in genetics is always to start experiments from defined single colonies.

The plasmid copy number is also not too much affected by the addition of a selective antibiotic. Another very simple procedure, however, increases copy numbers of colE1-derived plasmids (this is true for the vast majority of vectors) tremendously: the addition of chloramphenicol (or spectinomycin) to an exponentially growing culture around OD=0.6 and subsequent aeration for 4 h (or overnight) prior to plasmid isolation. This gives plasmid replication (protein-synthesis independent) a huge advantage compared with chromosomal replication (protein-synthesis dependent initiation at the origin of replication).

Low genome but best function!

This is strategy of live cell. Studied in virology is clearer.

So when you do not provide the conditions for using the plasmid, either antibiotic resistance gene or your input gene, plasmid is not useful for bacteria.

Many people add the selective antibiotic prior to plasmid isolation. This sounds reasonable, but actually does not help too much. The normal high-copy-number plasmids are essentially never lost from cultures. You can see this immediately, when you try to cure them; this can mean a lot of effort.

What really is helpful is to start all experiments from a single colony, which has been tested for the relevant features, e.g. resistance. Subculturing over subsequent liquid cultures of course selects for those bacteria that grow faster. State of the art in genetics is always to start experiments from defined single colonies.

The plasmid copy number is also not too much affected by the addition of a selective antibiotic. Another very simple procedure, however, increases copy numbers of colE1-derived plasmids (this is true for the vast majority of vectors) tremendously: the addition of chloramphenicol (or spectinomycin) to an exponentially growing culture around OD=0.6 and subsequent aeration for 4 h (or overnight) prior to plasmid isolation. This gives plasmid replication (protein-synthesis independent) a huge advantage compared with chromosomal replication (protein-synthesis dependent initiation at the origin of replication).

Johannes is correct in that most of the time high copy plasmids are rarely lost. The use of antibiotics is not always necessary except in the original selection. However, there are exceptions. Sometimes plasmid clones, especially expression clones, will have a strong negative growth effect on the cells, and the non-plasmid carrying cells will rapidly out compete the plasmid carrying cells. In these cases antibiotic selection is important. Because one is never sure of the degree of selection against your clone, adding antibiotics is safer. 

Thank you Michael for adding this important point.

I will emphasize the points raised by as it happened to me. I lost a plasmid expressing the T7 polymerase (pGP1-2; p15A replicon and thus low copy number) within 5 hours of growth. The point is you don't necessarily know what will stress the cells so better add antibiotics then sorry.

Yoram

I found all this narratives extremely useful. Thumb up

Thank you Dr. Johannes and Dr. Michael for sharing important knowledge.